ACCELERATING CHILD SURVIVAL AND DEVELOPMENT (ACSD)

Final Report The Retrospective Evaluation of ACSD: Ghana

Submitted to UNICEF on 7 October 2008

Institute for International Programs Johns Hopkins Bloomberg School of Public Health Baltimore, MD

Disclaimer:

This report was prepared by IIP-JHU under contract with UNICEF. All photos were taken by members of the IIP-JHU evaluation team after requesting permission from those who were photographed. All text, data, photos and graphs should be cited with permission from the authors and UNICEF.

ii IIP-JHU | Retrospective evaluation of ACSD in Ghana

Summary

Introduction UNICEF implemented the Accelerated Child Survival and Development (ACSD) project between 2001 and 2005 in 11 countries in Africa with support from Canadian CIDA. The aim of ACSD was to reduce mortality among children less than five years of age by working with governments and other partners to increase coverage with a set of proven interventions. In the “high-impact” countries of Benin, Ghana, Mali and Senegal, a total of 16 districts worked to deliver the full set of interventions grouped into three packages: “EPI+” including vaccinations, vitamin A supplementation and the use of insecticide-treated nets (ITNs) for the prevention of malaria; “IMCI+” including promotion of exclusive breastfeeding for six months, timely complementary feeding, use of iodized salt and improved and integrated management at the health facility and community levels of children suffering from pneumonia, malaria and diarrhea, including home-based ORS use, treatment of malaria, and treatment of pneumonia with antibiotics; and “ANC+” including intermittent preventive treatment of malaria with SP (Fansidar) for pregnant women (IPTp), tetanus immunization during pregnancy to prevent maternal and neonatal tetanus and supplementation with iron/folic acid during pregnancy and with vitamin A post-partum. An internal evaluation by UNICEF estimated through modeling that the levels of coverage achieved through ACSD were associated with about a 20 percent reduction in all-cause under-five mortality relative to comparison districts in participating districts in four “high-impact” countries. This retrospective evaluation was commissioned by UNICEF to confirm these findings and provide additional information that could be used in planning effective programs to reduce child mortality and achieve the 4th Millennium Development Goal (MDG-4) in poor countries in Africa. The IIP evaluation team worked with ACSD managers at international and national levels to develop a generic ACSD framework that defined the pathways through which ACSD activities were expected to lead to reductions in child mortality and improvements in child nutritional status. The generic framework served as the “backbone” of the evaluation design. The country-specific evaluations also addressed equity across socioeconomic and ethnic groups, for urban-rural residence and for girl and boy children. At the request of UNICEF, the evaluation does not include an economic evaluation or a full assessment of the effects of ACSD on national policy.

Aim of the independent retrospective evaluation in Ghana The aim of the evaluation was to provide valid and timely evidence to child health planners and policy makers about the effectiveness of ACSD in reducing child mortality and improving child nutritional status in Ghana, as a part of the larger retrospective evaluation designed to inform future programs intended to reduce child mortality and accelerate progress toward MDG-4. Equity was also assessed. Two questions served as a guide to the analysis and reporting of the evaluation findings: a) Was ACSD implementation associated with improvements in coverage, nutrition and mortality over time? b) If so, was progress in the ACSD districts faster than that observed for the national comparison area?

ACSD implementation in Ghana UNICEF-Ghana received approximately $3.8 million from Canadian CIDA to support ACSD activities in sixi “high-impact” districts (HIDs) with a combined population of about one million located in the Upper East region, and two expansion regions (Upper West and Northern regions) between 2001 and 2004. ACSD was implemented at the regional, district and sub-district levels in partnership with the Ghana

i These six districts subdivided into eight districts in 2005 during redistricting.

IIP-JHU | Retrospective evaluation of ACSD in Ghana iii

Health Service (GHS) and other development partners. The GHS supported EPI+ and ANC+ activities after 2004 by incorporating them into routine health services. After a hiatus of about one year, other ACSD activities received continued support from UNICEF funds, DANIDA and the Government of the Netherlands. ACSD inputs and activities in the Ghana HIDs, comprised of the entirety of the Upper East region, focused on:

1) Providing essential drugs, supplies, equipment and other support for outreach and campaign activities. ACSD-Ghana: a) provided an estimated 814 bicycles, 18 motorcycles and one vehicle to the HIDs over the course of the project for outreach and supervision activities; b) equipped health facilities with 553 refrigerator units for cold chain; c) supported local and national campaigns for vaccination and vitamin A supplementation, as well as routine health-facility outreach activities; and d) supplied commodities including vitamin A supplements, antihelminths, ORS, antimalarials and ITNs and retreatment chemicals for the prevention of malaria. 2) Supporting distribution and retreatment of ITNs at various levels. Over 200,000 ITNs were distributed in the HIDs between 2002 and 2005 through health centers, community outreach and distribution systems and campaigns. ACSD supported retreatment efforts at the community and facility levels, as well as through campaigns starting in 2004. All health workers and volunteers involved in ITN distribution and retreatment received training. 3) Training and supervising of facility-based workers. Forty-eight clinicians and three regional staff received standard 11-day IMCI training in 2005. 4) Training, equipping and supervising community health workers. ACSD-Ghana provided support for the training and supervision of over 1900 community-based agents (CBAs) in 600 communities to deliver messages to promote infant feeding, careseeking and treatment of childhood illness and ITNs, and immunization. The CBAs received health kits containing chloroquine, ORS, and handwashing and educational materials. ASCD also provided training and educational materials to community-based mothers’ groups for the promotion of infant feeding practices. 5) Supporting facility and outreach activities for pregnant women. The ANC+ package of ACSD included support for tetanus toxoid supplemental immunization activities, as well as facility and community distribution of postnatal vitamin A. IPTp was introduced in 2004 and ACSD supported its regional scale-up. Important barriers to full implementation of the ACSD implementation plan, as reported by program staff and reflected in project documentation, included: a) commodity insecurity, particularly stockouts of ITNs from late 2005 to late 2006; b) changes in the first-line antimalarial policy and the delayed authorization to distribute these drugs at the community level; and c) inadequate incentives and support and supervision systems for community-based workers.

Evaluation design and methods The IIP evaluation team worked with UNICEF-Ghana, the Government of Ghana and other partners to adapt the generic ACSD evaluation design to ACSD as implemented in Ghana. The intervention area was defined as the six HIDs located in the Upper East region. The comparison area was the remainder of Ghana excluding the urban areas of Greater Accra and Ashanti regions (Accra and Kumasi).

The primary data sources for estimates of intervention coverage were DHS surveys conducted in 1998-99 and 2003 at baseline, and a national MICS survey carried out in 2006 supplemented by a special extension of the MICS in the HIDs carried out in 2007. Information was collected and summarized in order to document ACSD intervention activities and contextual factors through key informant interviews, document reviews and field visits carried out as part of a mapping exercise by investigators at Kwame Nkrumah University of Science and Technology (KNUST). All results and interpretations were reviewed with representatives of the Government of Ghana and UNICEF-Ghana.

iv IIP-JHU | Retrospective evaluation of ACSD in Ghana

Results In Ghana, coverage for most of the ACSD interventions improved over time in the HIDs and reached the target coverage levels set by ACSD. Indicators showing positive trends over time in the HIDs included vaccinations, vitamin A, ITNs, antibiotics for suspected pneumonia, timely initiation of breastfeeding, exclusive breastfeeding, antenatal care, IPTp and the presence of a skilled attendant at delivery. Indicators that were observed to stagnate or decline included case management of common childhood illnesses, tetanus toxoid vaccination and postnatal vitamin A. Utilization of ITNs, antibiotics for pneumonia, breastfeeding initiation, skilled delivery and IPTp for pregnant women increased significantly more in the HIDs than in the comparison area. Appropriate management of childhood fever and diarrhea decreased in the HIDs, while stagnating in the comparison area; the difference in trends was statistically significant.

For coverage, the answers to the two primary evaluation questions are as follows: (a) Coverage indicators related to vaccination, vitamin A, ITNs, feeding behaviors, antenatal care and skilled delivery improved over time in the HIDs and most reached the target coverage levels set by ACSD. Indicators of correct management of childhood illness declined over time. (b) Comparison with the rest of the country showed mixed results. Coverage increased rapidly for a greater number of interventions in the HIDs than in the comparison area. On the other hand, coverage declined significantly more for interventions related to the case management of childhood illness in the HIDs than in the comparison area. For nutritional status: (a) The HIDs showed a reduction between 1998-9 and 2007 in the prevalence of stunting and underweight, but not in wasting. The largest decline in stunting occurred between 1998-9 and 2003, before sufficient time had elapsed for interventions supported by ACSD to have had an impact on nutrition (b) Relative to the national comparison area, stunting declined faster in the HIDs in the period from 1998-9 to 2006-7. Most of this drop occurred before 2003, before ACSD inputs and activities could have contributed, but the decline was maintained and extended during the ACSD project period from 2003 onwards. Wasting declined significantly in the comparison area while remaining stable in the HIDs.

For mortality:

(a) There was a reduction of 20% in under-five mortality in the HIDs from before to after ACSD implementation, close to the ACSD goal of 25%. This trend was ascertained through the full birth history technique, and the reduction was close to reaching statistical significance (p=0.10).

(b) Data on under-five mortality trends in the comparison area were available from a different source than those for the intervention area, with data points available through 2003. Other analyses suggest that mortality levels remained stable at around 115 deaths per thousand live births. Although these results must be interpreted with caution, they do suggest that the drop in under- five mortality was greater in the HIDs than in the national comparison area.

The assessment of equity in coverage was limited to the period after ACSD implementation, because of limited sample sizes available from earlier periods. There were no inequalities in coverage based on the sex of the child, and few differences between urban and rural households. Results by socioeconomic level were mixed, with few inequalities for interventions delivered through campaign approaches (e.g., vaccinations, vitamin A supplementation and ITNs), moderate levels of inequality for diarrhea management and antenatal care visits, and large differences favoring wealthier households for the presence of a skilled attendant at delivery. Children in the poorest households were somewhat more likely to be stunted and to die before the age of five years than children in the least poor households. Ethnic diversity within and between the HIDs and comparison area precluded examination of inequities by

IIP-JHU | Retrospective evaluation of ACSD in Ghana v

ethnic group membership. When HIDs were compared to the rest of the country, there was no evidence of differences in patterns of health inequalities.

Discussion and interpretation ACSD in Ghana focused available resources on filling gaps in EPI, distributing ITNs, expanding C-IMCI through community health workers and promoting antenatal care interventions. The highest coverage levels in the endline surveys in the HIDs were achieved for vaccinations, vitamin A supplementation to children, antenatal interventions (including IPTp and TT) and ITNs, and most of these interventions progressed faster in the HIDs than in the national comparison area. Exclusive breastfeeding also showed rapid increases in both the HIDs and the comparison area during the ACSD project period. Coverage levels for the correct case management of malaria and diarrhea were low and decreased in the HIDs from before to after ACSD. Taken together, the interventions showing large gains in coverage are likely to have had only limited impact on the main causes of death in Ghana (malaria, neonatal conditions, pneumonia, diarrhea and undernutrition) and hence are consistent with the 20 percent reduction in under- five mortality observed in the HIDs.

Interpretation of these findings jointly by the IIP evaluation and Ghana team focused on the missed opportunities for saving further child lives through ACSD, including the need for: 1) greater emphasis on interventions to address child undernutrition; 2) more intensive efforts to change behaviors related to the management of childhood illnesses, skilled delivery and child feeding; 3) greater support and training for the community-based workers that were a key part of intervention delivery; and 4) increased commodity security to ensure adequate and continuous supply of essential commodities. The team also believed that stronger supervision and monitoring systems would have increased ACSD effectiveness.

These results must be considered in light of the many international, bilateral and Ghanaian agencies that were active in the HIDs before and concurrent with the ACSD project. Special advantages and contributions of the ACSD project in this complex environment were defined by the implementation team as: 1) the program’s ability to concentrate on a package of effective interventions; 2) the provision of additional resources for commodities, equipment and human resources; 3) clearly stated targets; 4) the establishment of productive partnerships and synergies across institutions; and 5) achievement of strong commitment from the Government of Ghana and other donors. An important methodological issue for this and future evaluations is that the presence of other partners throughout Ghana makes it impossible to attribute observed changes to ACSD alone, and limits the validity of results based on comparisons between the HIDs and broader geographic areas.

In summary, the ACSD HIDs accelerated gains in coverage of several key interventions relative to gains in the rest of the country, despite the fact that the HIDs were among the poorest in Ghana and geographically remote. However, several key interventions for reducing the main causes of death in Ghana, showed little change and even some decreases in coverage. While stunting prevalence declined during the ACSD period, there was a similar decline in the remainder of the country from 2003 to 2006. In total, the changes in intervention coverage are consistent with the 20 percent reduction in under-five mortality observed in the HIDs, and compares with what appears to be little or no reduction in the rest of the Ghana.

vi IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table of Contents

1. The external retrospective evaluation of ACSD in four countries ...... 1

2. Evaluation methods ...... 5

3. Characteristics of the “high-impact” districts and comparison area ...... 11

4. ACSD as implemented in Ghana ...... 21

5. Coverage and family practices ...... 29

6. Nutrition ...... 49

7. Mortality ...... 55

8. Equity of coverage, nutrition and mortality ...... 61

9. Conclusions ...... 67

References ...... 71

Appendices

A. Description of Ghana and “high-impact” districts B. Methodology for documentation of implementation activities and contextual factors C. Documentation of implementation D. Definition of key indicators E. Survey Questions F. Methodologies of surveys in Ghana 1998-2007 G. Tables presenting priority coverage indicators over time for ACSD “high-impact” districts H. Tables presenting comparisons of priority coverage indicators over time in ACSD “high-impact” districts and the comparison area I. Tables presenting 2007 MICS results for key coverage indicators in the ACSD “high-impact” districts by socio-demographic characteristics of the population J. Additional tables for nutrition K. Methodological challenges L. References for the appendices M. Mapping of partners’ activities in ACSD “high-impact” districts (Upper East region) and nationally

IIP-JHU | Retrospective evaluation of ACSD in Ghana vii

Acknowledgements

This evaluation could not have been conducted without full participation of the representatives from the Ministry of Health, the Ghana Health Services, the Ghana Statistical Service, UNICEF-Ghana and other development partners. We thank them for their commitment to child survival, as reflected in their willingness to share their time, as well as information and their personal opinions about the contributions and limitations of the ACSD project. We specifically would like to thank Eddie Addai and George Amofah from the Ministry of Health and Ghana Health Service who were strong supporters of the evaluation. Vida Abaseka provided valuable information on ACSD implementation; we are grateful to her and the Ghana Health Service team in Upper East region for their dedication and openness to our questions. The Ghana Statistical Service carried out surveys integral to this evaluation; we especially thank Faustina Ainguah and Rochester Appiah for their on-going efforts. Easmon Otupiri at KNUST Department of Community Medicine carried out the program mapping activities, essential to understanding the context in the Upper East region, as well as contributing to the data interpretation.

UNICEF-Ghana staff were responsible for working with governments and partners to implement the ACSD project and collaborate in activities related to the independent retrospective evaluation and we thank them for their commitment to child survival and to the evaluation process as a means of improving program effectiveness. We would also like to express our appreciation to Dorothy Rozga, Yasmin Haque, Mark Young, Tamar Schrofer, Victor Ankrah, Bo Pedersen, Elias Massesa, George Fom Ameh, Augustine Botwe, Felicia Mahata and Joanne Greenfield. UNICEF-Ghana also provided financial support for the supplemental survey and advanced technical assistance from Macro, International. This support was essential, as without it there would have been few data to analyze. We would also like to thank UNICEF staff at regional and global levels for their efforts to provide us with documentation about ACSD and the values and conceptual frameworks that guided its implementation.

Additionally, we would like to thank the members of the IIP-JHU for their insights and help throughout the evaluation, as well as Macro International and Trevor Croft for technical assistance. Lanie Morgan provided valuable assistance in the documentation of ACSD implementation and contextual factors. Finally, we thank the leadership of UNICEF and CIDA, for their continuing commitment to the importance of independent evaluations and their efforts to see that this work was completed.

viii IIP-JHU | Retrospective evaluation of ACSD in Ghana

Acronyms

ACSD Accelerated Child Survival & Development Project

ACT Artemisinin combination therapy for use in treating fever/malaria.

ANC Antenatal care

ANC+ One of the ACSD intervention packages, consisting of antenatal care and the intermittent prevention of malaria during pregnancy (IPTp)

BASICS Basic Support for Institutionalizing Child Survival, a project supported by the United States Agency for International Development.

BFHI Baby Friendly Health Initiative

CBA Community Based Agent

CDC US Centers for Disease Control and Prevention

CHW Community health worker

CHO Community Health Officer

CHPS Community-based Health Planning and Services

CIDA Canadian International Development Agency

C-IMCI Community component of Integrated Management of Childhood Illness

DANIDA Danish International Development Agency

DFID Department for International Development, government of the United Kingdom

DHS Demographic and Health Surveys (DHS), supported by USAID.

DPT Diphtheria, Pertussis, Tetanus immunization

EPI Expanded Programme on Immunization

EPI+ One of the ACSD intervention packages, consisting of the full EPI schedule as well as the provision of vitamin A and deworming twice each year for children aged six to 59 months, and the provision of insecticide-treated nets for the prevention of malaria.

F-IMCI Facility component of Integrated Management of Childhood Illness, which includes improving the skills of facility-based health workers as well as strengthening aspects of the health system needed to provide appropriate care for children less than five years of age.

GAVI Global Alliance for Vaccines and Immunizations

GHS Ghana Health Service

GoG Government of Ghana

IIP-JHU | Retrospective evaluation of ACSD in Ghana ix

GoN Government of the Netherlands

GRCS Ghana Red Cross Society

GSS Ghana Statistical Service

Hib Haemophilus influenzae type b immunization

HIDs “High-impact” districts for ACSD implementation, defined as Bawku East, Bawku West, Bolgatanga, Bongo, Builsa, and Kasena-Nankana in the Upper East region in Ghana

IEC Information, Education and Communication

IHNS Integrated Health and Nutrition Survey in Northern, Upper East, and Upper West Regions of Ghana, 2002

IIP The Institute for International Programs at JHU

IMCI Integrated Management of Childhood Illness

IPTi Intermittent preventative treatment for malaria in infancy

IPTp Intermittent preventative treatment for malaria in pregnancy

ITN Insecticide-treated net

JHSPH The Johns Hopkins University Bloomberg School of Public Health

JICA Japan International Cooperation Agency

KNUST Kwame Nkrumah University of Science and Technology

LLITN Long-lasting insecticide-treated net

MBB Managing Budgets for Bottlenecks, a tool developed by UNICEF and the World Bank to support results-based planning for maternal, newborn and child survival in developing countries.

MDG Millennium Development Goal

MDG-4 The fourth millennium development goal, which aims to reduce mortality among children less than five years of age by two-thirds from levels in 1990.

MICS Multiple Indicator Cluster Survey designed by UNICEF

MOH Ministry of Health

NGO Non-governmental organization

NHIS National Health Insurance Scheme

NIDs National Immunization Days

ORS Oral Rehydration Salts, usually pre-packaged in a sachet

x IIP-JHU | Retrospective evaluation of ACSD in Ghana

ORT Oral Rehydration Therapy, can be either pre-packaged in a sachet or prepared in the home

pp Percentage points

PMTCT Prevention of mother-to-child transmission of HIV

RHMT Regional Health management team

SP A combination of two drugs, sulfadoxine and pyrimethamine. This drug combination is commonly known as Fansidar.

SIA Supplementary Immunization Activity

SWAp Sector-Wide Approach: World Bank

TBAs Traditional Birth Attendants

TT2 Two doses of Tetanus toxoid vaccine during pregnancy

UER Upper East Region

UNICEF United Nations Children’s Fund

USAID United States Agency for International Development

VCT Voluntary Counseling and Testing

WHO World Health Organization

IIP-JHU | Retrospective evaluation of ACSD in Ghana xi

xii IIP-JHU | Retrospective evaluation of ACSD in Ghana

1. The external retrospective evaluation of ACSD in four countries

UNICEF implemented the Accelerated Child Survival and Development (ACSD) project between 2002 and 2005 in 11 countries in Africa with support from Canadian CIDA and other partners. The main objective was to use results-based planning techniques to increase coverage with three packages of high-impact interventions known to reduce child mortality (see Box 1). In Benin, Ghana, Mali and Senegal, 16 “high-impact” districts worked to deliver all three packages; in the remaining Box 1: countries, the focus was on the “EPI+” package ACSD High-Impact that included vaccination, Vitamin A and Implementation Packages* insecticide-treated nets (ITNs) for the prevention of malaria. Internal UNICEF Immunization plus (EPI+) evaluations in 2003 and 2004 showed . Routine immunization and periodic measles increases in coverage for the EPI+ package in catch-up and mop-up all countries; UNICEF modeled the associated . Vitamin A supplementation bi-annually reductions in mortality using the "Marginal Budgeting for Bottlenecks " (MBB) tool and . Distribution and promotion of Insecticide Treated estimated an overall mortality reduction of 20 Nets for all children who are fully vaccinated as percent in the “high-impact” districts in the four well as pregnant women, and re-dipping of countries, relative to comparison districts.1 bednets every six months

UNICEF and the evaluation team recognized Improved management of pneumonia, malaria the limitations of a retrospective evaluation, and diarrhea (IMCI+) including the difficulties associated with . Promotion of exclusive breastfeeding for six reconstructing project assumptions and months, timely complementary feeding activities on a post hoc basis, and making the . Improved and integrated management (at the best possible use of available data and health facility, community and family levels) of information despite their shortcomings. children suffering from ARI, malaria and Readers are reminded to treat the results with diarrhea, including home-based ORS use, caution.2 treatment of malaria with anti-malarial blisters, and treatment of ARI with antibiotic blisters The aim of the evaluation is to provide valid . Promotion of household consumption of iodized and timely evidence to child health planners salt and policy makers about the effectiveness of ACSD Phase I in reducing child mortality and Antenatal Care (ANC+) improving child nutritional status. The specific . Intermittent preventive treatment (IPT) of objectives are: malaria with SP (Fansidar) for pregnant women . Tetanus immunization during pregnancy to 1. To evaluate the impact of ACSD on prevent maternal & neonatal tetanus mortality and nutritional status among children . under five. Supplementation with iron/folic acid during pregnancy and with vitamin A post-partum. 2. To document the process and ______intermediate outcomes of ACSD and results- * UNICEF grouped these interventions into based planning as a basis for improved paragraphs in different ways at various points planning and implementation of child health during the project; we have adopted the grouping programs. used in the final report from UNICEF to CIDA for the ACSD project in 2005.1 3. To document the contextual factors necessary for effective implementation of efforts to reduce child mortality in order to be able to extrapolate evaluation findings to other settings. 4. To assess the process, outcomes and impact of ACSD and results-based planning on socio- economic, ethnic and gender inequities. Achievement of these objectives should also expand regional and global capacity for large-scale effectiveness evaluations of strategies, programs and interventions designed to improve child health in low-income countries.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 1

1.1 Evaluation design

Geographic focus: The global retrospective evaluation covers the four countries within which UNICEF defined districts as “high impact” for the ACSD project. Within each country, we focus on these “high- impact” districts (HIDs).

Development of a generic impact model for ACSD: The first step in any evaluation is to define what those implementing the project expect to happen because of project activities. We developed an impact model that specifies the pathways through which UNICEF and implementing countries expected ACSD activities to result in reductions in child mortality.3 Figure 1 presents the generic ACSD impact model in two parts. Figure 1A shows the “top” of the framework describing expected ACSD inputs and processes from the point of introduction at national level in a country through the definition of the three packages of interventions recommended for accelerated implementation (see Box 1 for a description of the three packages). We derived the “top” of the framework from ACSD documentse.g.4 and discussions with ACSD implementers at all levels. Figure 1B shows the “bottom” of the framework, defining the pathways through which each of the three packages was expected to result in reductions in under-five mortality and improvements in the nutritional status of infants and young children. ACSD documents did not describe the pathways in the “bottom” of the model in detail, but made reference to other sources where the effects of the interventions are defined and quantified.5,6 For the internal evaluation,1 UNICEF utilized the estimates of effectiveness published in these sources and changes in intervention coverage as the basis for modeling the impact of ACSD on child mortality.

A central tenet of the evaluation is that the coverage, family practices and impact reflected in the “bottom” of the framework cannot be attributed to ACSD alone. UNICEF and country partners designed ACSD to reinforce existing activities in child survival by the government of each country and its partners. Therefore, increases or decreases in coverage and mortality must be understood as the result of a combined implementation effort, tempered by contextual factors. A key challenge for the current evaluation is to arrive at a qualitative assessment of ACSD’s role as a part of this overall effort; quantified attribution of the results to ACSD alone is not warranted given the implementation approach.

Definition of priority indicators for coverage and family practices. Priority coverage indicators address the prevalence of key family practices and intervention coverage for each of the elements defined in the “bottom” of the framework. Although some of these indicators reflect behaviors—such as exclusive breastfeeding and complementary feeding—rather than intervention coverage, these will be referred to as coverage indicators throughout the text. Appendix D defines the priority indicators of coverage utilized in the evaluation. Whenever possible, the ACSD priority coverage indicators are consistent with those supported by a consensus of United Nations (UN) agencies and multi- and bi- lateral partners for tracking progress toward MDG-4.7,8 Where no international consensus indicator exists, we contacted technical experts in the topical area to obtain advice on selection of a valid coverage indicator that could be calculated using the data available in Ghana.

Definition of priority indicators of impact (nutrition and mortality). The main objective of the ACSD project was to reduce mortality among children less than five years of age. The primary impact indicator in the evaluation is the under-five mortality rate, expressed as the probability of dying between birth and exact age five years. Additional priority indicators include infant and child mortality. Some ACSD project documents described expected improvements in child nutritional status, reflecting the synergy between undernutrition and infectious disease.9 In Ghana, the regional management team in the HIDs specified ACSD targets to reduce undernutrition by 15 percent in three years and by 25 percent in five years,10 although specific indicators of undernutrition were not defined. Priority impact indicators include prevalence of stunting, wasting and underweight. Appendix D presents the detailed definitions of the priority indicators for mortality and nutritional status.

2 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure 1B ACSD impact model: “Bottom” model showing interventions to impact “Top” model showing inputs and processes Antenatal Immunization + care +

DPT, Hib, measles Vitamin A Insecticide treated Iron/folic Tetanus Post-partum IPT for malaria vaccines supplementation nets acid toxoid Vitamin A High attendance at facilities/outreach sessions; deployment at community level High attendance at facilities/outreach sessions

Increased coverage Increased coverage

Pneumonia Measles Meningitis / sepsis Diarrhea Malaria Preterm delivery Neural tube defects Neonatal tetanus ?????

Spillover effect (co-morbidity)

IMCI + Reduced mortality Improved nutrition ? Improved nutrition? Reduced mortality

Pneumonia Breastfeeding Malaria treatment ORT treatment promotion

Deployment of interventions at community level

Increased coverage

Malaria Diarrhea Pneumonia Other infections

Spillover effect (co-morbidity)

Reduced mortality Improved nutrition

IIP-JHU | Retrospective evaluation of ACSD in Ghana 3

Equity. As part of the evaluation, we examine inequity in coverage and impact indicators, including socio-economic status, sex of the child, place of residence (urban or rural) and ethnic groups.

Documenting contextual factors. Contextual factors are defined as variables that can confound the association between the delivery of interventions and their health impact, or modify the effects of the approach.11 We documented contextual indicators in the HIDs and comparison area, including: (1) indicators of implementation-related contextual factors such as characteristics of the health system (e.g., utilization rates), child health policy, drug policy, and availability of drugs; and (2) indicators of impact-related contextual factors including baseline levels and patterns of child morbidity and mortality that can affect the potential magnitude of program impact.11

Economic evaluation. At the request of UNICEF, the evaluation does not include an economic component.

4 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2. Evaluation Methods

2.1 Evaluation design

Overall design.

The overall evaluation design was retrospective, drawing on existing population-based surveys with over-sampling of the MICS in the Upper East region, commissioned for the purpose of this evaluation. We re-analyzed pre-existing data sets whenever possible to ensure that the indicator definitions were correct and consistent. Preliminary results were reviewed in meetings of the evaluation team with representatives of the Ghana Health Service (GHS), the Ghana Ministry of Health (MOH), the Ghana statistical service (GSS) and the UNICEF country office in Accra, Ghana in July 2008.

Coverage and family practice indicators.

We reanalyzed existing household survey data to calculate the ACSD priority coverage and family practice indicators. As described above, these indicators are consistent with those used internationally for monitoring progress toward the Millennium Development Goals7,8 and are presented in appendix D. Appendix E provides the specific survey questions used for the indicator calculations.

Nutrition and mortality indicators.

We reanalyzed existing household surveys to calculate the priority nutrition indicators using the 2006 WHO Growth Standards.12 Appendix J and section 6 present more details on these methods. For calculation of priority mortality indicators, the evaluation team analyzed mortality retrospectively, using direct under-five mortality estimates based on full birth histories collected in the 2007 MICS supplemental survey in the HIDs. Estimates of under-five mortality in the comparison area were based on available direct and indirect estimates.

Intervention area.

The intervention area included the Upper East region (UER), selected for ACSD “high-impact” implementation. When ACSD was first implemented, the Upper East region comprised six districts: Bawku East, Bawku West, Bolgatanga, Bongo, Builsa, and Kasena-Nankana. In 2005, new health districts and boundaries were defined, these eight districts in Upper East region are: Bawku Municipality, Bolgatanga Municipality, Bongo, Builsa, Bawku West, Kasena-Nankana, Guru Tempane and Talensi Nabdam (Figure 3). Throughout the body of this report and appendices, we refer to the six “high-impact” districts (HIDs) defined at the inception of ACSD, unless otherwise noted.

Comparison area.

The main comparison area is the remainder of Ghana excluding the urban areas of Greater Accra and Ashanti regions (Accra and Kumasi). We have excluded Accra and Kumasi because access to services and living conditions in these areas differ considerably from the predominantly rural HIDs.

Intervention activities.

We documented the timing and scale of intervention activities using information collected from field visits to the HIDs, key informant interviews and document review, such as administrative and supervision reports and monitoring data.

Equity.

To examine inequities, we performed analyses of selected intervention coverage and impact measures stratified by sub-groups of the population, including household assets (expressed in quintiles), sex of the child, place of residence (urban/rural) and ethnic group.

Contextual factors.

We collected standard information on contextual factors, defined above, in order to assist in interpretation of the results and the potential contributions of ACSD. Certain elements, such as economic status, ethnicity and access to clean water were re-analyzed for HIDs and comparison areas using existing household survey data. Field visits to the HIDs, key informant interviews and

IIP-JHU | Retrospective evaluation of ACSD in Ghana 5

document review provided contextual information not available in existing surveys. A program mapping exercise, carried out by investigators at Kwame Nkrumah University of Science and Technology (KNUST), documented health and development activities in the HIDs and nationally between 1999 and 2007. Appendix B and M provide further details on the methods used to collect contextual factors.

2.2 Data sources and methods

Tables 1a and 1b summarize the different types of information used in the evaluation. The 1998-9 and 2003 Demographic Health Surveys (DHS) and the 2006 Multiple Indicator Cluster Survey (MICS) with a supplemental survey in the Upper East region (HIDs) conducted in 2007 served as the primary data sources for estimates of intervention coverage and nutrition in the HIDs and comparison area. For estimation of the endline coverage and nutrition results in the HIDs, we utilized the Supplemental MICS 2007; the 2006 national MICS was utilized to provide endline estimates in the comparison area, excluding the HIDs and urban areas of Greater Accra and Ashanti regions. We did not merge the MICS 2006 data for the HIDs with the supplemental MICS 2007 data due to incompatible sampling strategies and the small sample size of the data in the HIDs (Upper East region) in the 2006 MICS.

The 2007 supplemental MICS included a full-birth history module used to estimate child mortality both before and after ACSD implementation. The full-birth history method allows the calculation of period estimates of mortality ranging from the previous 12 months to 10 or more years in the past. No comparable data was available for the comparison area. Estimation of under-five mortality in the comparison area was based on indirect child mortality estimation as measured in the DHS 2003 and MICS 2006, and direct estimates from DHS 2003. Section 7 describes the mortality analysis methods in more detail.

Other survey data were available, but given lesser prominence in the analyses because they did not fully meet the quality criteria established for the evaluation. These criteria were: 1) full data sets and documentation, including sampling weights, available to the evaluation team so that the data could be reanalyzed using the standard definitions for priority indicators; and 2) no more than 5 percent missing values on key socio-demographic variables (e.g., child age) or the variables needed to construct the priority indicators. We did not use data from the Integrated Health and Nutrition Survey (IHNS) 2002 and the CDC-ACSD 2003 survey in the primary analyses because they did not fulfill these criteria. However, we use these data to explore time trends between 1998-9 and 2006-7. Descriptions of the methodology and conduct of surveys used in the evaluation are presented in appendix F and full documentation of 2003 ACSD-CDC survey data quality issues is available upon request from IIP-JHU evaluation team.

Table 1b presents sources of information used in the documentation of intervention activities and contextual factors. We collected information through: 1) review of documents, including administrative and monitoring reports; 2) key informant interviews; and 3) searches and review of published and grey literature. Technical staff at UNICEF-Ghana provided input and revisions throughout the process of documentation.

Due to the retrospective nature of the evaluation, it was difficult to collect complete and standardized information on ACSD implementation activities and other health activities in the Upper East region. The collaborative nature of ACSD makes it difficult to distinguish which activities were: 1) carried out as part of the ACSD program, 2) carried out with only partial technical and/or financial support from the ACSD program, or 3) carried out by ACSD partners, but independent of the ACSD program. In some cases, the information presented in administrative reports was inconsistent; for example, annual reporting of the number of bednets treated varied slightly. Appendix C notes observed discrepancies in implementation reports.

6 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table 1a: Data sources for independent retrospective evaluation of ACSD in Ghana, population- based surveys.

TYPE OF DATA DESCRIPTION USE IN EVALUATION Population- DHS 1998/1999: Nationally representative Used to establish baseline based surveys household survey conducted from November levels of priority coverage that met to February 1999. and nutrition indicators in inclusion HIDs and comparison area. criteria DHS 2003: Nationally representative Used to estimate interim household survey conducted from July to coverage and nutrition October 2003. indicators in HIDs and comparison area.

MICS 2006: Nationally representative Used to estimate priority household survey conducted from August to coverage and nutrition November 2006. indicators in HIDs and comparison area.

MICS supplemental 2007: Household survey Used to estimate endline in Northern, Upper West and Upper East coverage and nutrition region conducted from September to indicators in HIDs. Used for December 2007 with additional EA’s collected retrospective estimation of in February 2008. mortality in HIDs.

Other IHNS 2002: The Integrated Health and Reported in appendices, but population- Nutrition Survey in Northern, Upper East, and given limited weight in based surveys Upper West regions conducted from February analysis due to availability of to March 2002. a usable datafile.

CDC-ACSD 2003: Household survey of 2341 Reported, but given limited households in the Upper East region carried weight in analysis due to out from July to September 2003. concerns about data quality.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 7

Table 1b: Data sources for independent retrospective evaluation of ACSD in Ghana, routine data, administrative reports and key informant interviews.

TYPE OF DATA DESCRIPTION OF KEY DOCUMENTS USE IN EVALUATION Routine health Routine data collected through health facilities information pertaining to intervention coverage, compiled at Documentation of MOH and ACSD activities. system data the local, regional and national levels.

Administrative Annual UNICEF reports: Three administrative reports reports from 2003 – 2005 detailing implementation and inputs;

ACSD annual reports and presentations – Upper East region: Eight reports/presentations on ACSD progress 2004 - 2006, one EPI+ report 2004; Documentation of Ghana Health Service: Upper East region Health ACSD and partners’ Sector Annual Reviews: 2000 – 2006; activities.

Bawku West annual reports: five health sector update reports: 2004 – 2006;

IMCI training/monitoring reports: Nine documents prepared by the regional offices and KNUST.

Job aids and Job aids and tools, such as visual aids and tools register books, used in the implementation of Documentation of ACSD were collected and reviewed where ACSD and partners’ possible. activities.

Summary UNICEF Assessment of ACSD, 2004. Documentation of report ACSD activities. Program KNUST contracted to perform sub-study on mapping of partner activities in UER and nationally; development (Appendix M presents the full methods and Documentation of activities in sources list). contextual factors. UER and nationally

Key informant Approximately 24 interviews at the national, Documentation of interviews regional and district level: see appendix B for ACSD activities and summary. contextual factors. Working Field visit and discussions: November 2006; Discussion and discussions Review of preliminary results: July 2008. documentation of ACSD activities and contextual factors.

8 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2.3 Analysis We employed the Habicht et al framework13 for real-life evaluations. Starting with an adequacy evaluation, we assessed whether trends in coverage indicators were moving in the expected direction, and whether goals were met. Next, we carried out a plausibility evaluation, defined as a controlled, non-randomized study that assesses whether observed impact can be attributed to program implementation. ACSD in Ghana was a combination of separate interventions – vaccines, mosquito nets, vitamin A supplementation, etc – that are highly efficacious if delivered at optimal coverage. The evaluation did not assess the efficacy of these interventions, but instead focused on their impact when delivered under routine conditions. We carried out the analysis of coverage and nutrition in four steps, explained below. Section 7 describes the analysis of under-five mortality.

Step 1: Generating indicator levels for each survey in the analysis

Objective: To describe levels of priority indicators for coverage and nutrition in all surveys included in the analysis, overall and for specific subsets of children defined by age, sex, geographic location of the household, mothers’ education and socioeconomic status, where sample sizes permit. We applied standard indicator definitions to the reanalysis of all datasets to ensure the comparability of indicators over different surveys. For each indicator, only data for women and children with known responses for that indicator were included in the analyses; cases with missing or unknown data were excluded. The point estimates of indicators presented here may therefore differ slightly from those calculated using standard DHS and MICS tabulation programs, which do not exclude missing records from the analyses.

Step 2: Comparing rates of change over time within each ACSD district (“time trends”).

Objective: To determine whether there are statistically significant differences in indicator levels within HIDs from before ACSD was implemented to after ACSD was implemented in ACSD areas, with a mid-point during the process of implementation where adequate data are available, overall and for specific subsets of children. This step refers to the adequacy evaluation.

Step 3: Comparing rates of change between ACSD and non-ACSD districts within each country (“time trend with comparison”).

Objective: To determine whether there are statistically significant differences in the rates of change for indicator levels between the HIDs and comparison area where ACSD was not implemented (the comparison area is comprised of the rest of Ghana, excluding Accra, Kumasi and the HIDs), overall and for specific subsets of children.

Step 4: Attributing improvements to ACSD and related child survival activities at country level.

Objective: To determine whether any statistically significant changes in indicator levels can be attributed to ACSD activities, including activities implemented by others in collaboration with ACSD and the national child survival plan, overall and for specific subsets of children. Steps 3 and 4 refer to the plausibility evaluation, assessing whether progress was greater in the ACSD than in the comparison area, and whether or not external factors can account for these differences.

For all comparisons across time and geography, we initially calculated a simple chi-square statistic of difference. The simple chi-square statistic does not take into account the design effect of the survey, thus it under-estimates the variance. If no statistical differences were observed using the simple chi- square statistic, we assumed that none would be observed after the design effect was taken into consideration (adding to the variance) and that the groups were therefore not statistically different from one another. For comparisons with a significant chi-square, we calculated standard errors and 95 percent confidence intervals that take into account the survey design effect, using the Taylor Linearized Variance method. We used a “difference-in-differences” approach to compare whether the change in each indicator over time differed significantly between the HIDs and comparison area.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 9

10 IIP-JHU | Retrospective evaluation of ACSD in Ghana

3. Characteristics of the “high-impact” districts and comparison area Figure 2: Map of Ghana and This section presents pertinent characteristics of Ghana as a whole its neighbors and the HIDs and the comparison area. We emphasize differences between the HIDs and comparison area, as well as factors that have changed over the evaluation period to help guide the interpretation of evaluation results. Some of the quantitative results (table 2) presented here are based on our reanalyses of available survey data, because these provide the most recent information disaggregated by the HID and comparison area. Appendices A and M present additional information on the geographic, socio-demographic, economic, health and health service factors in Ghana and the UER.

3.1 The Ghana context

Ghana, located in West Africa, maintains three international boarders and a coast off the Gulf of Guinea (Figure 2). Togo is situated to the East, Cote d’Ivoire to the West, and Burkina Faso to the North and Northwest. Great Britain established a colony in 1874 known as the Gold Coast, and Ghana declared independence in 1957. The first president of Ghana, Dr. Kwame Nkrumah was overthrown in a military coup in 1966.17 A cession of military leaders ruled Ghana until Jerry Rawlings seized power in 1981 and gradually restored civilian rule, with the first free elections in 1992.17 The current president John Agyekum Kufor holds office in his second and final term ending in December 2008.

Of Ghana’s estimated 23 million population, 38 percent are younger than 15 years old.18 The estimated growth rate is currently 1.9 percent with a total fertility rate of 3.8 19 Box 2: births per woman. In 2000, 41 percent of the population was urban.19 Well- Overview of child health in Ghana endowed with natural resources,

Ghana’s per capita output is twice that Causes of under-five deaths in Ghana* of neighboring West African countries.

Pneumonia Malaria Despite prosperity relative to its 15% 27% 33% neighbors, Ghana maintains a 5.7 billion Injuries (US$) debt, 26 percent of the Gross 3% National Income. According to a new Ghana Living Standards Survey reported by the World Bank, poverty

HIV/AIDS levels have dropped from 52 to 29 6% percent between 1992 and 2005.

Measles 3.2 Child health in Ghana 3% Diarrhea 12% Ghana had an estimated population of 3.2 million children under age five in Neonatal 29% 2006. The under-five mortality rate has stagnated at 120 per 1000 live births 1990 2006 between 1990 and 2006, falling short of Mortality rates (per 1000 live births)** progress needed to achieve the two- Under-five 120 120 thirds reduction from 1990 levels defined Infant 76 by the fourth Milliennium Goal (40 per Prevalence of undernutrition*** 1000 live births). Box 2 shows the major Stunting (% mod + severe) 28 causes of under-five deaths in Ghana in 14 Underweight (% mod + severe) 13 2003 as reported by WHO Almost one-third of all under-five deaths occur Sources: *WHO, 200614; **SOWC15; ***MICS 200616 in the neonatal period. Among these deaths, infections account for

IIP-JHU | Retrospective evaluation of ACSD in Ghana 11

approximately one-third (32%) with the remainder attributed to preterm births (26%), asphyxia (23%), congenital (6%) tetanus (4%) and other causes (9%). Child undernutrition is also a problem in Ghana. In 2006, estimates using the new WHO growth standards indicated that 28 percent of children under five years of age were either moderately or severely stunted, 6 percent were wasted, and 13 percent were underweight.20 Appendix A includes the full profile of maternal, newborn and child health from the Countdown to 2015 2008 report.20

3.3 Selection of the ACSD “high-impact” districts in Ghana

UNICEF and the Government of Ghana Figure 3: Map of eight health districts, as (GoG) selected the six districts in the of 2005, Upper East region, Ghana Upper East region (UER) for “high- Bawku Mun. impact” implementation of ACSD. UER is one of the poorest regions in Ghana Kassena/Nankana Bongo and had high levels of under-five Bawku mortality, contributing to its choice for Bolgatanga M. West Garu- implementation of the ACSD approach. Tempane UNICEF had been supporting activities Talensi-Nabdam in the Bawku West and Builsa districts Builsa in the UER since 1995; the ACSD “high-impact” districts (HIDs) included these two districts as well as the remaining four districts in the UER. Redistricting occurred in 2005 and the UER is now comprised eight districts (Figure 3). Key informants reported that other factors considered in the choice of the UER included: 1) political stability; 2) a strong regional health team; and 3) a passable road network, ensuring high accessibility to the entire region.

3.4 Socio-economic and demographic factors

Figure 4 shows the incidence of poverty in the HIDs and the geographic comparison area as measured in the 2000 Housing and Population Census and the Ghana Living Standards Survey 4.21 The poorest districts are located in the northern areas of the country, with the six HIDs, noted in the call-out box, among the poorest in the country. Table 2 presents socio-demographic variables as measured in household surveys in 1998-9, 2003 and 2006-7 in the HIDs and comparison area. Based on our re-analyses of the DHS 1998-9, DHS 2003 and MICS 2006-7, households in the HIDs remained significantly poorer relative to the comparison area throughout the evaluation period (p<0.001), based on measures of household assets. The proportion of women with primary and higher education was significantly greater in the comparison area than the HIDs across all time points (p<0.001). In 1999 and 2003, almost two-thirds of women in the HIDs reported no schooling, decreasing to 58 percent of women without schooling in 2006-7. Similarly, female literacy was two times greater in the comparison area as compared to the HIDs in 1999 and 2003. The difference in literacy between the HIDs and comparison area narrowed in 2006-7, but was still statistically significant (p<0.001). The majority of households in the HIDs are of the Mole-Dagbani, Gruma and Grussi ethnic groups, while the comparison has much larger proportion of Akan, Ewe and Ga households (p<0.001). The apparent decline in the Gruma households in the 2003 DHS is thought to be due to different classifications of responses between surveys or the sampling error.

3.5 Environmental characteristics

The HIDs fall into Ghana’s savannah zone, with the forest and coastal zones in the central and southern areas of the country. The HIDs experience much less rainfall than the central and southern areas of Ghana, with particular drought hazards between January and March in the HIDs.22 Accordingly, malaria transmission is seasonal in the HIDs, with highest transmission between June and October.23 Models predict that the length of annual malaria transmission is longer in the southern areas of the country, becoming more seasonal in the northern zones,24 although other models predict

12 IIP-JHU | Retrospective evaluation of ACSD in Ghana

similar transmission intensities and prevalence of parasitemia among children less than five years of age in the northern, central and southern areas of the country.25 Before and during ACSD implementation, resistance of malaria parasites to chloroquine grew.26-28 Evidence suggests that levels of chloroquine resistance differed by geographic region, with the highest chloroquine resistance in the south and the lowest resistance levels in the north of the country.29

The HIDs are significantly more rural than the comparison area (p<0.001), which excludes the major metropolitan areas of urban Greater Accra and Ashanti regions (table 2). The apparent decrease in rural residences between 1999 and 2003 is likely due to previously rural localities reclassified as “urban” (population greater than 5,000) after the 2000 census.21 The proportion of households with access to an improved water source was greater in the HIDs than the comparison in 1998-9 and 2006-7 (<0.001). In both areas, access to improved water sources significantly increased, with greater increases over time in the comparison area. Less than five percent of households in the HIDs reported access to improved sanitation facilities in 2003 and 2006-7. Access to improved sanitation was significantly greater in the comparison area (p<0.001), but still less than 15 percent in 2003 and 2006-7.

Our investigations did not reveal any natural disasters in the HIDs over the primary evaluation period. However, severe flooding took place in the northern regions of Upper East, Northern and Upper West in August to November 2007, prompting the government to declare a state of disaster. Data collection for the Supplemental MICS 2007 endline survey was on going at this time; we discuss the implications of the flooding on intervention coverage (section 5) under methodological challenges.

Figure 4: Incidence of poverty in Ghana districts as measured by the 2000 Housing and Population Census and 1999 Ghana Living Standards Survey

Source: Coulombe, 200521

IIP-JHU | Retrospective evaluation of ACSD in Ghana 13

Table 2: Selected characteristics of the “high-impact” districts and comparison area, as measured in the DHS 1998-9 and 2003, and MICS 2006-7, Ghana. 1998/99 DHS 2003 DHS 2007 S. MICS 2006 MICS HIGH IMPACT GEOGRAPHIC HIGH IMPACT GEOGRAPHIC HIGH IMPACT GEOGRAPHIC DISTRICTS COMPARISONS¥ DISTRICTS COMPARISONS¥ DISTRICTS COMPARISONS¥ INDICATORS n* % n* % p n* % n* % p n* % n* % p Wealth quintiles

Poorest 43% 17% 57% 13% 33% 7% Poorer 271 37% 4615 22% <0.001 279 15% 4497 20% <0.001 3324 28% 4344 8% <0.001 Poor 3% 16% 5% 21% 19% 19% Less poor 4% 25% 14% 24% 10% 31% Least poor 14% 20% 8% 22% 10% 35% Education among women None 288 74% 3588 30% <0.001 310 72% 3870 31% <0.001 3288 58% 4167 30% <0.001 Primary school 10% 20% 12% 22% 21% 22% Secondary school+ 16% 50% 16% 46% 21% 48% Literacy among women 287 20% 3588 53% <0.001 309 14% 3853 37% <0.001 3257 27% 4141 41% <0.001 Ethnicity Akan(Asante, Akwapim, Fante and Other Akan) 1% 58% 0.3% 52% 0.6% 47% Ewe 0.9% 18% 0.2% 15% 0.2% 15% Gruma 29% 5% 1% 3% 28% 2% 288 3588 <0.001 310 3868 <0.001 3288 4167 <0.001 Mole Dagbani 43% 5% 41% 15% 45% 12% Grussi 21% 2% 17% 1% 12% 1% Ga/Adangbe, Guan, Hausa, Mande 2% 8% 2% 10% 8% 11% Other 3% 5% 38% 4% 7% 12% Rural residence 271 89% 4615 79% <0.001 279 80% 4497 70% 0.11 3324 78% 4344 71% <0.001 Hygiene§ Improved water source 271 70% 4613 53% <0.001 277 66% 4490 61% 0.43 3316 83% 4297 72% <0.001 Improved sanitation n/a n/a 279 4% 4492 12% <0.001 3316 4% 4339 13% ¥Excluding urban Great Accra and Ashanti region and High Impact districts *Weighted § MDG definitions

14 IIP-JHU | Retrospective evaluation of ACSD in Ghana

3.6 Baseline health conditions

Section 3.2 presents a profile of child health in Ghana as a whole, including the cause of death profile. Cause of death information is not available disaggregated by HIDs and comparison area. However, a vitamin A trial conducted in the Kassena-Nankana district (one of the six HIDs) in 1989 to 1991 found that children aged six to 59 months died from diarrhea (26%), malaria (23%), measles (19%), pneumonia (13%), malnutrition (8%).30 The proportionate causes of child mortality found in the study were slightly different than 2008 estimates for Ghana, likely due to decreases in measles deaths since 199131 and more HIV/AIDS deaths nationally and in later years. However, these findings suggest that the primary causes of death in the HIDs are similar to Ghana as a whole. We present and consider baseline levels of undernutrition and under-five mortality in sections six (nutrition) and seven (mortality).

3.7 Health service characteristics

Availability of health services.

The health services of Ghana have been decentralized, with regions and districts having more autonomy than in the past. Since Alma-Ata in 1978, there has been a focus on development of primary health care at the sub-district level, mostly through the training of health providers and installation of health facilities.32 Table 3 presents the approximate coverage of all health facilities in the HIDs and comparison area in 2002 and 2007. According to the Ghana Health Service (GHS) annual reports,33 the HIDs had approximately one health facility per 15,500 population in 2002, with coverage increasing to approximately one facility per 10,000 population in 2007. In the comparison area, coverage was estimated at one facility per 12,000 population in 2002, and increased to one facility per 10,000 population in 2007. These differences and increases are difficult to interpret, as available measures included all public and private health facilities, maternity centers, as well as nutritional rehabilitation centers. Community-based Health Planning and Services (CHPS) compounds, discussed below, were not included in these estimates.

The Community-based Health Planning and Services (CHPS), is an expansion of the primary health care concept, through community engagement and placement of community health officers (CHOs) to make primary health services more accessible. It began in the Kassena-Nankana district (one of the HIDs) as a research project in 1994,34-36 and has since been expanded to other communities in the HIDs. In the HIDs, the GHS in UER reported seven functioning compounds in 2002 and 82 functioning compounds in 2006. The catchment area for community health officer or CHPS compound is to be comprised of approximately 3000 individuals.37 The MOH planned to deploy 1570 community health officers (community health nurses) to various communities nation-wide by 2006; however, implementation is far behind schedule and only 258 CHPS compounds were reported to be functioning in the comparison area in 2006. Coverage of CHPS compounds in 2007 in the HIDs was much greater (approx 11,220 population per CHPS facility) than in the comparison area (58,000 population per CHPS compound).

Table 3: Coverage of health facilities and CHPS compounds in 2002 and 2006, Ghana. GHANA, EXCLUDING YEAR INDICATOR HIDs (UER) ACCRA & HIDs* 2002 Population** 920,089 15,077,264 Hospitals 6 80 Total health facilities*** 59 1244 CHPS compounds 7 32

2007 Total health facilities*** 92 1425 CHPS compounds 82 258 *estimates exclude Greater Accra, but not urban Ashanti region **estimates from 2000 Housing and Population census ***Includes hospitals, clinics, health centers and maternity homes, as well as private facilities; 2002 estimates taken from the GHS 2002 annual report; 2007 estimates from the GHS annual report and website33

IIP-JHU | Retrospective evaluation of ACSD in Ghana 15

Changes in health policies.

A number changes in national policies influencing child health took place between 2000 and 2007. In 2002, Hib vaccination was introduced into national policy and included in routine EPI vaccination schedules. Due to growing chloroquine resistance, national policy changed to recommend ACTs as the first-line antimalarial drug in April 2004. ACTs became available in health facilities & CHPS compounds in late 2005; the community-based distribution of ACTs was not authorized until late 2007.

The GoG introduced the National Health Insurance Scheme (NHIS) in 1998, but inadequate and slow reimbursement limited its effectiveness. The GoG passed a law concerning the NHIS in 2003 to support districts to set up mutual health insurance schemes and to initiate activities to recruit and register clients. The NHIS automatically covered children less than 18 years of age if parents have paid at least the minimum contribution. No estimates of NHIS coverage were available at the writing of this report, although the 2007 Supplemental MICS survey will provide estimates of coverage for the Northern, Upper East, Upper West and Central regions.

3.8 Other projects that may impact child health

Child health partners and activities in the HIDs (UER).

As Belch states in his background document on Upper East region: “If signboards are held to constitute development, then Northern Ghana has no further need of it.38” A multitude of international and local development partners and NGOs implemented child survival, health and other development activities in the HIDs both before and during ACSD implementation. Table 4 provides a summary of the activities and approximate coverage of selected health projects in the HIDs from 1998 to 2007; appendix M provides further details. We documented major child health and nutrition activities in the HIDs during this period, given available data; this list should not be considered comprehensive of all child health activities in the HIDs.

Many activities supported by partners in the HIDs focused efforts on child nutrition. The LINKAGES project, funded by USAID and carried out by Academy for Education Development (AED), provided support from 2000 to 2003 for activities to improve infant feeding practices in the northern regions, including the HIDs. This project trained a variety of actors and provided technical support to NGOs, the GHS, and UNICEF to implement packages to promote appropriate infant feeding, including early initiation of breastfeeding, exclusive breastfeeding and complementary feeding. There are approximately 30 supplementary feeding centers in the HIDs (Upper East region) which provide supplementary feeding for children and mothers through support from the World Food Program and impart educational messages about child nutrition and survival. Fifteen nutritional rehabilitation centers have also been established with support from churches, Catholic Relief Services (CRS) and World Vision International.

With support from the American Red Cross, the Ghana Red Cross Society (GRCS) established over 60 mothers’ clubs in the Bawku East, Bawku West and Bolgatanga districts to promote child health and infant feeding, and provide home based management of fever and diarrhea in 1999. The mother’s clubs and health and nutrition promotion activities expanded over time, covering 200 communities in 2002, partially in collaboration with the LINKAGES project described above. Between 1999 and 2000, GRCS served an estimated 16,500 children under-five annually in the HIDs through these activities. Starting in 2003, GRCS collaborated with the UNICEF ACSD project to train and equip over 1800 community-based agents (CBAs) to carry out health promotion and community management of common illnesses. Section 4, “ACSD as implemented in Ghana” describes these activities in further detail.

16 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table 4: Summary of selected child health and nutrition projects and activities outside of routine services in the HIDs from 1998-2007, Ghana BEST ESTIMATE OF GEOGRAPHIC COVERAGE OR PROGAM TYPE OF ACIVITY COVERAGE INTENSITY TIMING Trainings & community meetings for improved nutrition; CRS Bongo district over 200 trained 2000-2002 on-going support to feeding centers 1 hospital & 7 health centers; n/a w ith outreach points Avg. 235 deliveries Diocesan Reproductive health / skilled annually in 1998-2002; Health and deliveries Six communities in Avg. 1014 deliveries 1998 to development Bongo, Builsa & annually in 2004-2006 present Kassena- Avg. of 7422 ANC visit programmes, in A NC coordination Nankana annually 1999-2006 with CRS Avg. 21,045 children Child w elfare clinics served annually in 2001- 2005 C-IMCI training 20 staff trained 2005 47 communities ~16,500 children served (99) to 200 annually through mothers communities (01) clubs; training of mothers in 1999-2002 Ghana Red in Baw ku West, mothers' clubs in CS Cross, UER Baw ku East, interventions & infant HBM of malaria & diarrhea; ARI Bolgatanga feeding (with support recognition & referral; infant Training and support from EU, feeding; immunization promotion (bicycles & ,medicines) of 2003-2005 AMCROSS & 1820 w omen through 900 communities UNICEF) ACSD project in UER Training and support of 9750 w omen 2006 CHWs/volunteers Workshops w ith ~100 Message & materials participants from radio development to promote BF, 2000 stations, GHS, CRS, GRCS, EBF, complementary feeding UNICEF Training of partners & over 150 participants 2000-2003 LINKAGES providers in nutritional BCC (health staff, NGOs, radio) Training of trainers for mother- Project (USAID UER (& UWR & to-mother support groups for Training of over 50 trainers 2001-2003 NR) supported, AED nutrition BCC as implementers) Over 400 Mother-to-mother Formation of mother-to-mother support groups formed & support groups for improved 2001-2003 supported by CRS, GRCS, infant feeding & nutrition GHS, UNICEF, A CDEP

Radio broadcasts 500 in UER, UWR & NR 2000-2003 Early trials on Navrongo effectiveness of ITNs, Health Child and maternal health entire Kasena- vitamin A; CHPS delivery 1993-present Research research Nankana district strategy; malaria treatment Center and prevention via antimalarials 4 communities in constructed in Nutrition rehabilitation ceters 4 centers constructed Bongo district 1996-9 World Vision Scales, vaccination, vitamin Logistic/equipment support to A, dew orming; See International Bongo district 1998 - 2007 district Appendix L for details on Bongo Area inputs Development Supplemental feeding to school Approx. 2000 children Bongo district 1999-2003 Program children served annually Training on Approx. 500 mothers Bongo district 1999-2007 nutrition/breastfeeding trained annually

IIP-JHU | Retrospective evaluation of ACSD in Ghana 17

The Diocesan Health Services was active in the HIDs of Bongo, Builsa and Kassena-Nankana, with one hospital and seven health centers with outreach points. Through these services, approximately 7000 ANC visits took place annually and a skilled provider attended 1000 deliveries annually. Between 2001 and 2005, these services served approximately 21,000 children annually through child welfare clinics providing preventative services, such as vaccinations, and curative care. Also located in Kassena-Nankana district, the Navrongo Research Center conducted effectiveness trials on supplementation of children with vitamin A,30 use of insecticide treated bednets for children35 and pregnant women39 and the CHPS strategy for primary health care34-36 over the last decade. The Navrongo Center has continued to implement and conduct research around these and other interventions in the entirety of the Kassena-Nankana district.40

Other NGOs, among them Rural Help Integrated, Action Aid, and World Vision International, also conducted water and sanitation, reproductive and sexual health and community development projects throughout the period of 1999 to 2007. Additionally, CIDA, the World Bank and the GoG supported water and sanitation projects in over 1000 communities in the HIDs.

Child health partners and activities in the rest of Ghana.

We provide here a brief overview of the external investments in health activities taking place in the rest of Ghana (our comparison area) both before and after ACSD, appendix M provides detailed descriptions of these activities. Donor support for health and HIV/AIDS in Ghana totaled approximately US$150 million annually between 2003 and 2007, the majority in grants rather than through credit.

USAID supported promotion and marketing of ITNs, initiatives to improve the quality of care, and community-based health & planning services (CHPS), as well and family planning and HIV/AIDS activities in the comparison area before and during ACSD implementation. These activities focused in 30 target districts in the southern and central regions of Western, Central, Volta, Greater Accra, Eastern, Ashanti & Brong-Ahafo and were carried out by partners including: Ghana sustainable change project; Population council; JHPIEGO; Quality Health Partners; Engender Health; Abt Associates; AED; CRS, Futures Group, DELIVER; Opportunities for Industrialization Centers International; and Netmark. USAID supported technical assistance and partner programming with annual budgets of approximately US$9-12 million. Approximately 50 percent of these funds supported child survival and 50 percent supported HIV/AIDS activities through 2002; in 2003 to present, approximately 30 percent of funds were targeted to child survival projects and 70 percent for HIV/AIDS.

Japan International Cooperation Agency (JICA) provided support to EPI programs, GHS static and outreach services, and HIV/AIDS logistic support in the south of the country and nationally. The Global Fund to Fight Aids, Malaria and Tuberculosis (Global Fund) awarded the national MOH almost US$9 million for malaria control in 2003; another US$38 million grant to control malaria followed in 2005. Additionally, the Global Fund has granted approximately US$51 million for HIV/AIDS programming and US$24 million for tuberculosis programming to date, starting in 2003. The Global Alliance for Vaccines and Immunizations (GAVI) supported national vaccination initiatives with approximately US$5 million annually.

The World Bank provided support of approximately US$35 million through 1998-2002 through the Sector-Wide Approach (SWAp) pooled health funding mechanism and provided over US$100 million for health programming in 2003 to 2007. The Danish International Development Agency (DANIDA) provided approximately US$10 million annually through the SWAp for health systems strengthening between 2003 and 2007. During this same period, the British government’s Department for International Development (DFID) provided over US$20 million annually for health at the national level. The World Health Organization (WHO) also contributed to national pooled funding, as well as supporting health system strengthening and child health clinics (preventative and curative) in eight districts in the south and central Ghana.

The Government of the Netherlands (GoN) provided approximately US$10 million annually in 2003 to 2005 for health nationally. In 2006 and 2007, the GoN increased their support to more than US$28 million annually. Part of this investment was in support of the GHS child health strategy, High Impact Rapid Delivery (HIRD), which is based closely on the ACSD approach. In addition to the GoN,

18 IIP-JHU | Retrospective evaluation of ACSD in Ghana

DANIDA supported the roll out of HIRD in the Northern, Upper West, Upper East and Central regions, with US$0.80 to US$1.4 million for HIRD provided to each of the four regions in 2006. In 2006, DFID donated US$11 million to UNICEF for the purchase and national distribution of almost two million long-lasting ITNs to children under-two through a national campaign at the end of 2006.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 19

20 IIP-JHU | Retrospective evaluation of ACSD in Ghana

4. ACSD as implemented in Ghana

This section provides an overview of the ACSD activities in the HIDs. We consider adaptation of the generic ACSD package, funding, results-based planning and the timeline of activities; the inputs and activities for each ACSD component are then briefly described. Appendix C includes further textual description of the intervention implementation and detailed timelines of ACSD activities.

4.1 Introduction and adaptation of the generic ACSD intervention package

UNICEF introduced the ACSD approach to the Government of Ghana (GoG) in late 2001, followed by planning meetings with regional officials in the Upper East and Northern regions. According to key informants, UNICEF presented the generic ACSD framework to regional and district officials, assessed current levels of intervention coverage, set coverage targets and planned how to achieve the ACSD targets. As described above, various development partners—including UNICEF, the Ghana Red Cross Society, and Navrongo Health Research Center—in collaboration with the Ghana Health Services (GHS) had been supporting child survival activities in selected districts and communities of the UER for over a decade. Interventions such as immunization, vitamin A distribution, iron and folic acid supplementation for pregnant women, and the promotion of exclusive breastfeeding, complementary feeding, and iodized salt were well established before ACSD was introduced, albeit at less than ideal coverage levels. Before ACSD, some districts had started to implement the promotion and distribution of insecticide treated nets (ITNs), as well as community case management of diarrhea and malaria.

ACSD drew on the experiences of these programs to package the ACSD strategy of interventions for region-wide scale-up. Additionally, the ACSD strategy supported the development training materials and scale up of C-IMCI volunteers and introduced IPTp, PMTCT, deworming and post-natal supplementation with vitamin A. The ACSD strategy was implemented at the regional, district and sub-district levels in partnership with the Ghana Health Service (GHS) and other development partners.

4.2 Funding

UNICEF-Ghana received support of US$3.8 million through Canadian CIDA for implementation of ACSD, equivalent to approximately US$25 per child under-five years of age in the HIDs, as well as US0.7 million in resources from other donors from 2002 to 2004.ii The last transfer of CIDA funds was in 2003, and by the end of 2004, expenditure was 97 percent of the initial funds. At the end of the CIDA funds, the government continued in EPI+ and ANC+ through routine services. There was a lag in external support for activities for over one year, until the Government of the Netherlands and DANIDA provided significant funding for the northern regions to continue ACSD activities.

4.3 Results-based planning

ACSD implementers chose the package of interventions to be implemented in the four “high-impact” countries based on evidence and cost-effectiveness. The ACSD strategy set specific targets for each package and UNICEF monitored results actively at the regional, district, sub-district and community level, in coordination with GHS implementers.10 Key informants noted that GHS implementers presented a report bi-annually to UNICEF to justify funds used for activities. We did not find evidence that ACSD in Ghana included performance contracts or other innovations linking results to specific incentives.

ii Assuming that all funds were spent in the HIDs. We were not able to disaggregate funding by implementation area, i.e. to identify support in the HIDs versus expansion areas. A review of ACSD conducted in 2004 estimated a US$1.9 million price-tag for ACSD implementation outside of routine GHS expenditures from 2001-2003,10 with an approximate expenditure of US$12.35 per child under-five over this period. UNICEF’s final report to CIDA estimated a per capita annual cost US$0.34 through all funding sources and per capita annual costs of US$0.29 through CIDA funding for children in the HIDs and expansion regions.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 21

4.4 Time line of ACSD activities in the “high-impact” districts

UNICEF introduced the ACSD strategy to the GoG in late 2001 and logistical support for EPI+ started in 2002, while ITN distribution and treatment programs kicked-off in early 2003. Figure 5 presents summary timeline for the start of selected interventions within the ACSD approach, as well as for household surveys. The C-IMCI program was scaled-up in all districts in late 2003 and ANC+ by early 2004. Table 5 and Appendix C provide additional information about the timing of specific activities.

Figure 5: Time line of the accelerated implementation of selected ACSD interventions and surveys conducted to evaluate intervention coverage, 2001 - 2007, Ghana

Figure Key: • Grey bars represent implementation before ACSD, colored bars implementation supported by ACSD • Spotted area represents ITN stock-outs • Bars represent districts in the following order: Builsa, Bawku East, Kasena-Nankana, Bolgatanga, Bawju West, Bongo

22 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table 5: Start times for accelerated implementation of selected ACSD interventions in the ACSD “high-impact” districts of Ghana. INTERVENTION PACKAGE APPROXIMATE START TIME IN HIDS (UPPER EAST REGION) Routine EPI on-going prior to ACSD; scale-up of EPI+ logistical EPI+ support began in early 2002

ITNs On-going before ACSD in Kassena-Nankana and Builsa districts; Started region-wide in second half of 2003 2004: first Tetanus Toxoid - Supplementary Immunization Activity (SIA) campaigns; ACSD funded IPTp begins; postnatal vitamin A ANC+ 2004: de-worming for pregnant women

2004/2005: training of facility staff F-IMCI

1st quarter 2003: Planning and budget meetings

2nd-4th quarter 2003: Training of trainers and CBA training C-IMCI sessions begin

2004: Sensitization workshops and full scale delivery

2004 – 2005: Intermittent stock-out of retreatment chemicals (KO ITN stock-outs tablets)

Late 2005 – 2006: Stock-out of ITN nets

Child Health Promotion Week May 2004 (CHPW)

4.5 Description of ACSD activities in the “high-impact” districts

EPI+.

The EPI+ strategy in Ghana included: 1) routine immunization and periodic measles catch-up; 2) twice yearly vitamin A supplementation of children six to 59 months of age; and 3) twice yearly de- worming of children with anti-helminthic drugs. Distribution, promotion and retreatment of ITNs for children under-five was promoted as part of the IMCI+ package in Ghana; however, we present this intervention as part of EPI+ for consistency with overall ACSD documents.

Vaccination, vitamin A and de-worming Delivery of routine vaccination occurs at health centers and through outreach activities, as well as through National Immunization Days (NIDs), which occurred every quarter during the ACSD period. The ACSD strategy in the HIDs started with EPI+ in early 2002 with US$0.5million in USAID support.

ACSD focused on developing and supporting strategies to improve defaulter tracing using community- based surveillance systems. CBA volunteers used a register to trace children due for vaccinations. Mop-up campaigns occurred after National Immunization Days (NIDs) to vaccinate “zero dose” children identified by polio vaccinators. Vitamin A supplementation of children six to 59 months of age started nationally in 1996. In 2004, NIDs incorporated de-worming, vitamin A supplementation and tetanus toxoid (TT) supplemental immunization activities (SIA) as part of the ACSD program.

Many development partners contributed to vaccination activities and it is difficult to identify ACSD- specific contributions. UNICEF provided vehicles, motorcycles, bicycles and fuel for outreach and supervision activities linked to routine vaccination activities. In addition, UNICEF provided logistical

IIP-JHU | Retrospective evaluation of ACSD in Ghana 23

assistance to the health sector when required. For instance, UNICEF purchased polio vaccines to assist GHS in achieving the polio eradication goals. Table 6 summarizes available information, extracted from administrative and summary reports, about ACSD inputs intended to reinforce EPI+ activities. To provide rough guidance on the potential coverage of these activities, we present several of the indicators as ratios per 1,000 children under- one year of age or under-five years of age. Appendix C presents further description of EPI+ activities and timing, as well as quantitative monitoring data.

Table 6: Description of inputs related to the accelerated implementation of the EPI+ intervention package in the HIDs, Ghana. INTENSITY OF COVERAGE DESCRIPTION OF ACTIVITY TIMING ACTIVITY ESTIMATE Estimated as 814 bicycles, 18 * Provision of bicycles, motorcycles 2002 – 2004 motorcycles and 1 truck for the district and vehicles for outreach activities 553 refrigerator units distributed to 2002-2007 health facilities to support cold chain Training of health agents in EPI 1522 CBA- 40 per 1000 provision, monitoring and 2002 – 2004 surveillance children 0-11 surveillance volunteers trained months of age** 356 supplements 54,803 vitamin A 2004 per 1000 children supplements Vitamin A supplementation under-five** integrated in NIDs and CHPW 2005 Quantitative data not available

170,736 1110 doses per 2004 antihelminthic 1000 children Integration of child de-worming doses*** under-five** into NID activities 177,553 1154 doses per 2005 antihelminthic 1000 children doses*** under-five**

Monitoring data available in Appendix C; EPI+ annual totals 2003 - 2006 Table C2

*Unknown for which ACSD program transportation was purchased; cumulative total uncertain; see Table C5 in Appendices for further information **Estimated population from 2004 projections: 153,799 under five children, 36,223 pregnant women & 38,450 children 0-11m ***Delivered through two campaigns in March and September

Insecticide-treated nets (ITNs). ACSD-supported distribution of ITNs began in the second half of 2003, the exact start of implementation varying by district. ITNs were distributed to the district offices, then to the volunteers and then to the communities. The ACSD strategy employed multiple approaches for bednet delivery and treatment:

. Sale of bed nets to target groups at health centers;

. Sale of bednets by commissioned volunteer sales agents accompany nurses on health outreach sessions;

. Distribution and treatment by community-based agents (CBAs) trained in C-IMCI activities;

. Campaigns for distribution and retreatment.

All volunteers, CBAs and nurses involved in the ITN program received training on bed net distribution and treatment.

ITNs were sold at a reduced price to families with children under five and pregnant women through a chit (voucher) system. Initial insecticide treatment was provided with the net and included in the price.

24 IIP-JHU | Retrospective evaluation of ACSD in Ghana

However, as demand for ITNs increased, the subsidized nets were sold using the 20/80 rule. Target groups could purchase 80 percent of the ITNs for 5000 cedis (~US$0.50) and non-target groups could purchase the remaining 20 percent for 23,000 cedis (~US$2.30). Volunteers were responsible for advising customers at the time of purchase to retreat the net every six months, through the health centre or a volunteer. Retreatment cost 2000 cedis (~US$0.20) per net and the ITN volunteer agents received 1000 cedis (~US$0.10) per net sold or retreated. In May 2004, nets were retreated free of charge during national child health week. Reports indicated that although each ITN purchased came with an initial insecticide treatment kits, supply and stock-outs of insecticide for retreatment of existing nets was an on-going issue.

A large number of ITNs were distributed through 2005 (table 7), although key informants reported, and monitoring data reflect, stock-outs of ITNs starting in late 2005. The GHS in the HIDs reported procuring ITNs through the Global Fund in 2006. In late 2006, two million long-lasting ITNs were distributed countrywide to children under-two during the national immunization and vitamin A campaign, with support from DFIF (~$US11 million) and UNICEF. The HIDs distributed almost 90,000 long-lasting ITNs during this campaign in late 2006.

Table 7 summarizes available information, extracted from administrative and summary reports, about ACSD inputs intended to reinforce ITN activities. Once again, to consider the potential coverage of these activities, we present several of the indicators as ratios per 1,000 children under-five years, even though it is recognized that the ITNs may have been used by non-targeted members of the population. As a result, the coverage estimates below are likely overestimated. Appendix C presents further description of ITN activities and details on exact timing.

Table 7: Description of inputs related to the distribution, promotion and treatment of ITNs in the ACSD “high-impact” districts of Ghana. DESCRIPTION OF INTENSITY OF COVERAGE ACTIVITY TIMING ACTIVITY ESTIMATE 109,579 ITNs distributed for 712 ITNs per 1000 children under-five children under-five* 2002 –2004 1000 ITNs per 1000 36,223 ITNs distributed to estimated pregnant pregnant women women* Distribution of ITN 132,270 ITNs distributed for 860 ITNs per 1000 children under-five children under-five* 2005 1120 ITNs per 1000 40,576 ITNs distributed to estimated pregnant pregnant women women* 2006 93,832 long-lasting ITNs 610 ITNs per 1000 distributed** children under-five* Approx. 82 nets retreated June 2003 – Approximately 14,000 nets per 1000 nets estimated June 2004 retreated in the community Re-treatment 103 nets retreated per campaigns 2005 25,034 nets retreated 1000 nets estimated in the community 28 nets retreated per mid 2006 6,829 nets retreated 1000 nets estimated in the community Cumulative total of 287,850 ITNs distributed from 2002 to mid-2006 (not Reported Mosquito nets provided inclusive of national campaign LLITNs in late 2006) 2006 NB: Global Fund provided 80,000 ITNs in the HIDs, included by UNICEF in numbers above *Estimated population from 2004 projections: 153,799 under five children, 36,223 pregnant women **includes long-lasting ITNs distributed through DFID-supported national campaign in late 2006

IIP-JHU | Retrospective evaluation of ACSD in Ghana 25

IMCI+.

The IMCI+ strategy in Ghana included: 1) integrated management of childhood malaria, diarrhea, and pneumonia at facility and community/household levels; 2) promotion of infant feeding practices, including exclusive breastfeeding and complementary feeding; and 3) promotion of improved hygiene and consumption of iodized salt. The IMCI+ strategy in Ghana also included distribution and promotion of ITNs, which are discussed above.

Community IMCI The UER has a strong history of community-based health volunteers and before ACSD, Ghana Red Cross, Catholic Relief Services, and others supported community volunteers and community integrated management of childhood illness (C-IMCI) on a small scale. ACSD built on and harmonized these experiences, implementing C-IMCI at a greater scale than before. Through a memorandum of understanding (MOU) UNICEF, Kwame Nkrumeh University of Science and Technology (KNUST), Ghana Red Cross and Ghana Health Services (GHS) collaboratively developed and supported a CBA training program, with almost 2000 community based agents (CBA) trained starting in the second half of 2003 to deliver services and educational messages in 600 communities. The volunteer CBAs received training to carry out following activities:

. Promotion of appropriate infant feeding practices;

. Provision of health education to mothers, including recognition and referral of childhood pneumonia;

. Treatment of fever with pre-packed chloroquine, management of diarrhea with ORS;

. Promotion of immunization and iodized salt;

. Mobilization of communities for participation in de-worming, NIDs and other programs.

The CBA volunteers were equipped with bicycles, educational materials and health kits containing chloroquine doses for children and infants, ORS sachets and handwashing materials, although key informants and document review revealed that not all CBAs received bicycles, educational materials or health kits. CBAs earn a small percentage of medicine sales, for example a CBA earns 100 cedis (US$~0.01) for every ORS sachet sold.

Monitoring reports documented that CBAs focused on illness management and health education activities were sporadic or absent in many cases. CBAs treated more than 80,000 children with fever and more than 60,000 children with diarrhea between 2003 and 2006; CBAs referred less than 1000 children with pneumonia annually. In 2004, due to growing levels of antimalarial resistance, Ghana national policy changed the first-line antimalarial to artemisinin combination therapies (ACTs). CBAs retained chloroquine in their kits until the end of 2006, and received authorization and training to use ACTs only at the end of 2007. Appendix C presents detailed information about childhood illnesses treated and health education activities conducted at the community level as reported through routine monitoring systems.

The KNUST team and the regional UNICEF office carried out monitoring and supervision of CBA activities, at times integrated into routine supervision by the Regional Health Management Team (RHMT). Key informants noted that integrated supervision was problematic at times: for example, sub-district supervisors are reluctant to carry out CBA supervision without additional funds for fuel. They also noted stronger supervision and monitoring of CBAs in CHPS zones where these activities were incorporated in the responsibilities of the community health officers (CHOs).

In 2006, supervision teams found that out of 1366 CBAs visited, almost one-half had no bicycles, more than half (~60%) had no health kits and one-third did not have reporting forms. We did not find comprehensive information about retention of CBAs initially trained through ACSD; however, key informants in Bongo district noted that 60 percent of CBAs remained active in 2006.

Facility-based IMCI Facility-based IMCI started after C-IMCI in the HIDs. In 2005, ACSD supported the standard IMCI training of 48 clinicians and three regional staff. The training-of-trainers at the regional level included sessions pertained to CBA supervision. In 2006, the IMCI monitoring team evaluated IMCI-trained prescribers and found high non-compliance with the IMCI objectives regarding assessment, diagnosis and supervision. Standard IMCI training for facility-based providers is on-going in the HIDs.

26 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Infant feeding practices. In addition to infant feeding promotion included in IMCI activities, UNICEF supported facilities to implement the Baby Friendly Health Initiative (BFHI) to promote appropriate infant feeding practices. In partnership with the Ghana Red Cross Society, ACSD also supported promotion of appropriate infant feeding practices by developing guidelines on exclusive breastfeeding, training mother-to- mother support groups and providing training materials. ACSD explicitly did not focus on nutritional rehabilitation.

Table 8 summarizes available information about ACSD inputs intended to reinforce IMCI activities. Again, to estimate rough coverage, we present selected inputs and activities as ratios per 1,000 children under-five years of age or ratios of CBAs supplied. Appendix C provides more description of IMCI+ activities and timing.

Table 8: Description of inputs related to the implementation of the IMCI+ intervention package in the ACSD “high-impact” districts of Ghana.

DESCRIPTION OF INTENSITY OF COVERAGE ACTIVITY TIMING ACTIVITY ESTIMATE 72 prescribers & 10 1 IMCI-trained provider Facility-based agent 2004 - 2006 clinicians trained in per 1000 children under- (prescriber) training HIDs five 1982 CBAs trained 13 trained CBAs per 1000 CBA training in C-IMCI 2003 - 2006 in 600 communities children under-five*

Provision of 1400 706 kits per 1000 trained health/medicine kits CBAs Provision of 2022 1020 booklets per 1000 2003-2006 reporting booklets trained CBAs C-IMCI CBA supplies (cumulative total) Provision of 4851 doses chloroquine 746,100 doses of per 1000 children under- chloroquine five* Provision of 4200 ORS sachets per 645,900 ORS 1000 children under-five* sachets Supervision ongoing by regional teams, facility- Supervision 2004 - 2006 based health staff and KNUST**

Monitoring of the number of cases with diarrhea, 2004 - 2006 Monitoring fever and ARI seen and referred**

*Estimated population from 2004 projections: 153,799 under five children, 36,223 pregnant women & 38,450 children 0-11m **See further details in Appendix C, Table C2

ANC+.

ANC+ in Ghana was implemented with the aim of preventing maternal and neonatal tetanus and low birth weight due to malaria and severe anemia in pregnancy. The ANC+ strategy in Ghana included: 1) distribution and promotion of ITNs for pregnant women; 2) IPT for malaria with sulphadoxine pyremethamine; 3) tetanus toxoid immunization during pregnancy; 3) supplementation with iron and folic acid during pregnancy and vitamin A in the post-natal period; and 4) voluntary counseling and testing (VCT) for HIV/AIDs and prevention of mother to child transmission (PMTCT). The ANC+ strategy did not focus explicitly on delivery and neonatal care.10

The GHS provided ANC and delivery services, including supplementation with iron and folic acid and TT immunization, before ACSD. ACSD and other partners supplied support for the distribution and

IIP-JHU | Retrospective evaluation of ACSD in Ghana 27

promotion of ITNs for pregnant women; these activities are discussed above under the EPI+ component. ANC+ included support for TT immunization, with many women immunized during NIDs. The first round of TT supplementary immunization activities (SIA) took place in early 2004, with an estimated 46 percent coverage of the targeted population. Subsequent rounds of TT SIA achieved higher rates of coverage (table 9).

IPTp was coupled with routine ANC services at facilities and promoted through radio spots. Bongo and Bawku East districts received Global Fund support for IPTp activities starting May 2004. In mid- 2004, ACSD extended IPTp to the remaining four HIDs, with a training of trainers and district level training. Approximately 25,000 doses of SP were administered annually, although monitoring data show high drop out rates after first and second dose. Deworming for pregnant women and postnatal vitamin A supplementation began in mid-2004 through facilities. In 2005, TBAs and CBAs received training to distribute vitamin A in the postpartum period.

Table 9 summarizes available information, extracted from administrative and summary reports, about ACSD inputs intended to reinforce ANC+ activities. Appendix C presents further description and exact timing of ANC+ activities.

Table 9: Description of inputs related to the implementation of the ANC+ intervention package in the ACSD “high-impact” districts of Ghana.

DESCRIPTION OF COVERAGE ACTIVITY TIMING INTENSITY OF ACTIVITY ESTIMATE 143,954 / 230,700 women 2004 62%** TT SIA targeted

2005 No data 2006 National EPI+ report for UER: TT2 – 78%**

IPT training 2004 May: Training of trainers June: District level training IPT1:50%* 35,257 doses of SP May –Dec 2004 IPT2:31%*

IPT3:16%* IPT1:30%* 23,260 doses of SP 2004 - 2005 IPT2:21%* IPT dosing IPT3:13%* IPT1:30%* 24,046 doses of SP 2006 IPT2:22%*

IPT3:15%* Postnatal Vitamin A 5,973 vitamin A supplements 16%** July – Dec 2004 supplementation

2,217 vitamin A supplements 2005 6%**

10,351 vitamin A 2006 supplements 29%**

* as reported in ACSD monitoring reports, **coverage estimated based on target population from 2004 projections: 36,223 pregnant women

28 IIP-JHU | Retrospective evaluation of ACSD in Ghana

5. Coverage and family practices

This section of the report presents the results for priority coverage and family practices indicators and their interpretation. Section 2 describes the methodology used for the analysis, and appendix D defines the priority coverage indicators. We present the results in graphical form for selected priority coverage indicators within each intervention package. Two graphs are presented for each package. The first shows time trends in indicator levels in the HIDs. Arrows in these graphs indicate the ACSD coverage targets as adapted for Ghana. We present data from the ACSD-CDC survey conducted in 2003 in shades of grey and without confidence limits because, as explained in the methods section, these estimates are of lesser quality and should be interpreted with caution.

The second graph for each intervention component presents indicator levels in 1998-9 (baseline) and 2006-7 (endline), with an intermediate point in 2003 for the HIDs and comparison area. The number at the bottom of each bar in the graph is the percent coverage of the indicator and the black lines in these graphs represent the 95 percent confidence limits. We carried out differences-in-differences statistical tests for these comparisons and the results are presented in the text. Appendices G and H present the full results for HIDs and comparison groups.

Tables in appendix I present coverage results from the 2007 Supplemental MICS survey in the HIDs by district, urban versus rural residence, socio-economic status and age of the child. Here we present only statistically significant results on differences in coverage by sub-populations. For certain indicators and sub-populations, the results should be interpreted with caution due to the small sample sizes for some cells. Chapter 8 includes further results in the context of equity.

5.1 Results

EPI+.

Vaccinations and vitamin A supplementation. Figure 6 shows the time trends in measles and DPT vaccinations and one dose vitamin A supplementation in the previous six months in the HIDs in Ghana, based on the two DHS, two MICS with an additional point estimate drawn from the ACSD survey carried out in mid-2003. Coverage levels for these three indicators increased significantly between 1998-9 and 2006-7 (p<0.001). The results suggest that gains in vaccination were gradual over the evaluation period, while gains in vitamin A supplementation mostly occurred between 1998-9 and 2003. All indicators were at or above their ACSD program targets of 80 percent coverage, indicated in the graph with an arrow.

Appendix tables I2 and I3 provide further information on coverage levels for vaccinations and vitamin A supplementation by sub-groups of the population in the HIDs in 2007. No significant differentials in sub-groups of the population were observed for vaccination. Vitamin A supplementation in the previous six months varied significantly (p<0.001) by district, with children in Kasena-Nankana having the highest coverage (96%) and children in Bongo the lowest coverage with vitamin A supplementation (74%). Coverage with vitamin A was marginally higher in urban areas compared to rural areas (p=0.07) and children aged six to 11 months were less likely to receive vitamin A than older children (p<0.001).

IIP-JHU | Retrospective evaluation of ACSD in Ghana 29

* Figure 6: Coverage levels for measles and DPT3 immunization and receipt of one vitamin A supplement in the preceding six months in the HIDs as measured in DHS (1998-9 and 2003), the ACSD survey (2003), and MICS (2006 and 2007), Ghana. 100 ACSD vaccination & 93 95 vitamin A objective 91 90 80 84 86 82 80 76 73 68 68 60 66 65 …. 60 40 Coverage (%) 20 § § § § § § 0 Measles DPT3 immunization Vitamin A immunization supplementation (one dose)

DHS 1998-9 ACSD 2003 DHS 2003 MICS 2006 Sup MICS 2007

§ Estimate based on less than 100 children * Vitamin A coverage data only available for children 6-32 months in 2003 ACSD survey Note: Measles and DPT3 indicators are calculated based on MICS protocol, where the distribution of children reporting vaccination before 12m in vaccination card is applied to all other children reported as vaccinated.

Figure 7 shows coverage levels for vaccinations and vitamin A supplementation in the HIDs and the comparison area in 1999 and 2006-7, with a midpoint in 2003. Measles vaccine coverage increased significantly between 1998-9 and 2006-7 in both the HIDs and comparison area; the increase in the HIDs was not significantly different from gains in the comparison area. DPT3 coverage also increased in both areas, with greater increases in the HIDs (p<0.001). Vitamin A supplementation with one dose in the previous six months increased by 25 and 74 percentage points (pp) in the HIDs and comparison area, respectively. Increases in vitamin A coverage over time in both the HIDs and comparison areas were statistically significant (p <0.001). Baseline levels of vitamin A coverage were significantly less in the comparison area, and the increase in coverage in the comparison area was significantly greater relative to the increase in the HIDs (p<0.001).

30 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure 7: Coverage levels for measles and DPT3 immunization and receipt of one vitamin A supplement in the preceding six months in HIDs and the comparison area as measured in DHS in 1998-9 and 2003, and MICS 2006-7, Ghana.

§ § § §

§ Estimate based on less than 100 children Note: Figures inside bar represent percentage coverage

Insecticide-treated bednets (ITNs). Figure 8 shows the time trends in the use of ITNs in the HIDs in Ghana, based on the two DHS, two MICS with an additional point estimate drawn from the ACSD survey carried out in mid-2003. The 1998-9 DHS did not collect information pertaining to bednet use; thus, no comparable indicators for ITN use among children were available at baseline. However, if ITN use among children is assumed to be close to zero in 1998-9, there were significant increases between 1998-9 and 2007. In 2007, ITN use among children (58 pp) exceeded the ACSD target of 50 percent coverage. The MICS 2006 and supplemental MICS 2007 did not collect information about ITN use among pregnant women, precluding the examination of this indicator in our analyses.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 31

Figure 8: Coverage levels for insecticide-treated nets in the HIDs as measured in DHS (1998-9 and 2003), the ACSD survey (2003), and MICS (2006 and 2007), Ghana.

100

80 ….

60 ACSD target (ITN child) 50 58 40

Coverage (%) Coverage 43

20 27 23 0 0 ~0 ITNs (child)

DHS 1998-9 ACSD 2003 DHS 2003 MICS 2006 Sup MICS 2007

Appendix table I4 provides further information on coverage levels for ITNs in 2006-7 in the HIDs by district, urban/rural residence, child’s sex and age, and wealth quintile. ITN use among children was significantly higher in Builsa, Bongo and Talensi-Nabdam districts relative to the other districts (p<0.001). Higher proportions of children aged zero to 35 months slept under ITNs as compared to children aged 36 to 59 months (p<0.001)

Figure 9 shows coverage levels of ITN use in the HIDs and the comparison area in 1999 and 2006-7, with a midpoint in 2003. The proportion of children sleeping under an ITN increased by approximately 58 pp in the HIDs and 24 pp in the comparison area, if coverage in 1998-9 is assumed to have been close to zero in both areas (both trends p<0.001). The rates of increase over time, between 1999-9 and 2006-7 and between 2003 and 2006-7, were significantly greater in the HIDs relative to the comparison area (p< 0.001).

32 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure 9: Coverage levels for insecticide-treated nets in HIDs and the comparison area as measured in DHS in 1998-9 and 2003, and MICS 2006-7, Ghana.

Absolute change in percentage points between High-impact districts (UER) 1998-9 and 2006-7 Comparison area ~+24 ~+58 100

80

60

Coverage (%) 40

20

§ ~0 ~0 § 23 3 58 24 0 1998-9 2003 2006-7 ITN

Note: Figures inside bars represent percentage coverage

IMCI+.

The IMCI+ package includes case management and nutrition assessment and counseling, and the provision of locally adapted messages to improve family practices related to child survival. In this section, we focus on results related to the case management of childhood illness and child feeding practices.

Case management. Figure 10 shows time trends in the HIDs in the administration of any antimalarial for the management of childhood fever (presumed to be malaria in this highly endemic country), appropriate care-seeking for suspected pneumonia and oral rehydration therapy and continued feeding for diarrhea. The measurement of these indicators is based on reports by mothers of children who reported these illness symptoms in the two weeks prior to the survey. Additional data are available in appendix tables I5, I6 and I7. About three-fourths of mothers of febrile children reported giving their child an antimalarial at baseline, and this decreased significantly in 2007, with only about half receiving an antimalarial. However, in 2007 mothers reported that only nine percent of febrile children received artesunate-amodiaquine, the first line antimalarial in Ghana since 2004 (appendix table I6). Approximately one-half of children with probable pneumonia were taken to a health facility in both 1998-9 and 2007, with no significant change during this period. The proportion of children with diarrhea receiving oral rehydration therapy or increased liquids to prevent dehydration, along with continued feeding, decreased over time, from 39 percent in 1998-9 to 28 percent in 2007. Case-management indicators stagnated or declined over time; none of the ACSD case management coverage targets (indicated with arrows) were met by 2007.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 33

Figure 10: Coverage levels for case management indicators in the HIDs as measured in DHS (1998-9 and 2003), the ACSD survey (2003), and MICS (2006 and 2007), Ghana. 100 ... ACSD target (antimalarial & pneumonia) 80 78 71 ACSD taregt (oral 60 67 66 rehydration & con't 61 feeding) 50

54 .. 53 50 40 39 35 37 32 30 20 28 § § § § § § § §

0Sample size too small Proportion of ill children managed for illness (%) 0 Antimalarial treatment Careseeking for Oral rehydration and pneumonia continued feeding

DHS 1998-9 ACSD 2003 DHS 2003 MICS 2006 Sup MICS 2007

§ Estimate based on less than 100 children

Appendix I presents further details and analyses stratified by sub-population (where sample-sizes permit) for the management of fever and diarrhea, as well as careseeking for pneumonia, in 2007 in the HIDs. Treatment of febrile children with an antimalarial varied significantly by district (p<0.001); only 42 percent of febrile children living Garu-Tempane district received an antimalarial, while in Bolgatanga municipality 85 percent of febrile children received an antimalarial. Antimalarial coverage among urban children was higher than among their rural counterparts (p<0.01). Fifty-eight percent of boys received an antimalarial, while coverage among girl children was 48 percent (p=0.02). Girls were marginally more likely to be adequately managed for diarrhea than boys were (p=0.05). Sample sizes were too small to perform all stratified analyses for indicators related to careseeking for suspected pneumonia and diarrhea management.

Figure 11 presents coverage levels for the case management of childhood illness in the HIDs and the comparison area in 1999 and 2006-7, with a midpoint in 2003. Levels of treatment with any antimalarial for fever decreased significantly in the HIDs, while stagnating in the comparison area. The difference in the changes between 1998-9 and 2006-7 in the HIDs relative to the comparison area was statistically significant (p<0.001). However, if the indicator is defined as “treatment of fever with an effective and nationally recommended antimalarial” there was a precipitous drop in coverage in both the HIDs and the national comparison area, because chloroquine was no longer recommended at the end of the period, but use remained frequent (appendix table I6). Care seeking for pneumonia remained relatively stable in the HIDs, while increasing 14 pp in the comparison area and the difference-in-differences was significant (p=0.04). Correct home management practices for diarrhea decreased in the HIDs, while increasing

34 IIP-JHU | Retrospective evaluation of ACSD in Ghana

seven pp in the comparison area; the difference between the trends in the HIDs and comparison area was statistically significant (p<0.01).

Figure 11: Coverage levels for case management indicators in HIDs and the comparison area as measured in DHS in 1998-9 and 2003, and MICS 2006-7, Ghana. High-impact districts (UER) Absolute change in percentage points btw. 1998-9 to 2006-7 Comparison area +1 +14 +7

100 -25 -4 -11 (%)

80

60

40

20

Proportionof ill childrenmanaged for illness § § § § § § 78 60 71 67 53 61 54 22 66 40 50 36 39 23 32 38 28 30 0 1998-9 2003 2006-7 1998-9 2003 2006-7 1998-9 2003 2006-7 Antimalarial treatment Careseeking for Oral rehydration and pneumonia continued feeding

§ Estimate based on less than 100 children Note: Figures inside bars represent percentage coverage

Figure 12 presents the use of antibiotics for suspected pneumonia among children aged 0-59 months in the HIDs and the comparison area in 1999 and 2006-7. Coverage with antibiotic treatment was low at baseline in the HIDs (2%), and increased to just over 50 percent in 2006-7. Use of antibiotics for suspected pneumonia also increased in the comparison area, although only by 15 pp. These results are inconsistent with the trends in careseeking for pneumonia presented in the previous graphs, which showed stagnation in the HIDs. Secondary analyses (appendix table I7) found that 12 percent of children with suspected pneumonia in the HIDs were taken for care at a private drug vendor in 2007, while no mother reported this behavior in the comparison area. Similarly, approximately 25 percent of mothers in the HIDs reported obtaining the antibiotic for their child’s pneumonia outside of an appropriate health facility, with 21 percent of antibiotics obtained at a drug shop and 4 percent obtained from drug peddlers. Antibiotic distribution at the community level is not authorized in Ghana through community-based distributors or drug peddlers.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 35

Figure 12: Coverage levels for use of antibiotics for suspected childhood pneumonia in HIDs and the comparison area as measured in DHS in 1998-9 and MICS 2006-7, Ghana.

Absolute change in percentage points between High-impact districts (UER) 1998-9 and 2006-7 Comparison area +15

100 +49

80

.. 60

40 2 antibiotics

20 § 2 20 51 35 0 1998-9 2006-7

% of children with suspected pneumonia receiving pneumonia children with suspected of % § § Antibiotic treatment for pneumonia

§ Estimate based on less than 100 children Note: Figures inside bars represent percentage coverage

Feeding, including breastfeeding. IMCI+ in the context of ACSD also included promotion of appropriate infant and young child feeding practices (Box 1). Figure 13 shows the prevalence of selected feeding behaviors as reported by mothers of children less than one year of age at the time of the survey. Breastfeeding behaviors tend to be relatively stable over time, so apparent fluctuations should be interpreted with caution as they may reflect differences in how the questions were posed, the answers recorded or statistical error due to small sample sizes. The proportion of mothers reporting initiation of breastfeeding within one hour of birth increased significantly between 1998-9 and 2007 (p<0.001), with a large, unexplained fluctuation in the 2003 DHS. The prevalence of exclusive breastfeeding of infants less than six months steadily increased over time, from 28 percent of mothers reporting this practice in 1998-9 to 55 percent of mothers reporting this practice in the 2007 survey (p<0.01). This exceeded the ACSD objective of 50 percent coverage by five pp. Complementary feeding among children six to nine months of age remained stable between 2003, 2006 and 2007, with approximately half of children in this age group reported to have received complementary feeding and continued breastfeeding throughout the period. Sample sizes in the 1999-8 and 2003 DHS were too small to provide valid estimates of complementary feeding from these surveys. Prevalence of appropriate complementary feeding practices fell well short of the ACSD target of 90 percent. Mothers reported continued breastfeeding 84 percent of children aged 20-23 months in 2007 in the HIDs (appendix table G3); sufficient sample sizes were not available for baseline estimation of this indicator.

36 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure 13: Prevalence of infant feeding behaviors as reported by mothers in the HIDs as measured in DHS (1998-9 and 2003), the ACSD survey (2003), and MICS (2006 and 2007), Ghana. ... 100 ACSD target (complementary feeding) 90 80 85

ACSD target (exclusive 60 breastfeeding) .. 50 56 55 .. 52 50 53 53 40 45 43 42 39 28 20 11 Sample sizeSample too small 0 sizeSample too small 0 Prevalencebehaviorof as reportedmothers by (%) 0 Initiation of breastfeeding Exclusive breastfeeding (< Breastfeeding plus within one hour of birth 6 months) complementary food (6-9 months)

DHS 1998-9 ACSD 2003 DHS 2003 MICS 2006 Sup MICS 2007

§ Estimate based on less than 100 children

Appendix table I8 shows the breakdown of breastfeeding practices by selected sub-groups of the population, where sample sizes permit. The proportion of rural mothers initiating breastfeeding within one hour of birth (54%) was marginally greater than that among their urban counterparts (42%); (p=0.05). The proportion of infants exclusively breastfed varied significantly by district, with over 70 percent exclusive breastfeeding in Kasena-Nankana, Bongo and Bolgatanga municipality districts and only 37 percent in Bawku municipality (p=0.02). Exclusive breastfeeding was more common among women residing in urban areas (p<0.01) and in wealthier households (p=0.01) than among women in rural and poorer households.

Figure 14 shows the prevalence of infant feeding behaviors in the HIDs and the comparison area in 1999 and 2006-7, with a midpoint in 2003. Initiation of breastfeeding within one hour of birth significantly increased by 41 pp in the HIDs and increased only eight pp in the comparison area; the difference in the rates of change was significant (p<0.001). Exclusive breastfeeding up to six months of age increased by more than 20 pp in both the HIDs and the comparison area; the difference-in-differences was not significant. Complementary feeding of children six to nine months of age declined by 10 pp in the comparison area. Sufficient sample sizes were not available in the HIDs in 1998-9 or 2003, precluding a comparison in trends.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 37

Figure 14: Prevalence of infant feeding behaviors as reported by mothers in HIDs and the comparison area as measured in DHS in 1998-9 and 2003, and MICS 2006-7, Ghana. High-impact districts (UER) Absolute change in percentage points between 1998-9 and 2006-7 Comparison area

+8 +22 -10

... +41 +27

100

80

60

40 ..

§ small 11 20 * Prevelanceof behavior as reportedby mother(%) * § § § .. small too size Sample 23 85 41 52 31 28 29 43 49 55 51 0 67 too size Sample 64 53 57 0 1998-9 2003 2006-7 1998-9 2003 2006-7 1998-9 2003 2006-7 Initiation of breastfeeding within Exclusive breastfeeding Breastfeeding plus one hour of birth (< 6 months) complementary food (6-9 months)

§ Estimate based on less than 100 children *Estimation based on unweighted data, no 95% confidence intervals presented Note: Figures inside bars represent percentage coverage

ANC+.

The ANC+ package as implemented in Ghana included interventions in both the antenatal and perinatal periods. In this section, we address coverage levels for antenatal interventions and interventions designed to improve maternal and neonatal health during delivery and the post-natal period.

Antenatal care. Figure 15 shows the time trends in coverage of antenatal care in the HIDs. Further details are presented in appendices G, H and I. The proportion of women reporting four or more ANC attendances increased steadily by 17 pp over the evaluation period (p<0.001). The ACSD target of 80 percent coverage of four or more ANC visits was achieved. Intermittent presumptive treatment (IPTp) with two doses of SP for malaria during pregnancy was not measured in 1998-9, and coverage was only four percent in 2003, IPTp coverage increased dramatically between 2003 and 2006, with further increases between 2006 and 2007 (p<0.001). Approximately two-thirds of women reported IPTp, falling just short of the ACSD objective of 75 percent ITPp coverage. Tetanus toxoid (TT2) vaccination, consisting of two doses during pregnancy, remained stable, with approximately one-third of women not reporting two vaccinations during their previous pregnancy. However, in 2007 in the HIDs, 78 percent of women reported full neonatal protection from tetanus toxiod in their previous pregnancy, close to the ACSD target of 80 percent coverage. Comparable information about neonatal protection from tetanus toxoid was not collected in earlier surveys, precluding comparisons of this indicator over time.

38 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure 15: Coverage levels of antenatal indicators in the HIDs as measured in DHS (1998-9 and 2003), the ACSD survey (2003), and MICS (2006 and 2007), Ghana. 100 ACSD target ACSD target (NN (ANC 4+) protection) 80 86 ACSD target 80 81 (IPT) 75 74 .

.. 60 64 64 67 63 61 63 56 40 47

33 Coverage (%) Coverage 20 § 4 § § § § 0 dataNo .. data..No § § § 0 Antenatal care IPTp with SP 2 TT doses (4+ visits)

DHS 1998-9 ACSD 2003 DHS 2003 MICS 2006 Sup MICS 2007

§ Estimate based on less than 100 women

Appendix I9 provides further information on coverage levels of antenatal care in the HIDs as measured in the 2007 Supplemental MICS. The proportion of women reporting four or more antenatal care visits varied somewhat by district (p=0.07), with highest levels in Kasena-Nankana district (94%) and the lowest in Garu-Tempane district (73%). Coverage of ANC interventions was inequitable in 2007 in the HIDs. Significantly more women in the wealthiest households reported four or more ANC visits (93%) as compared to those in the poorest households (75%); (p=0.02). Similar inequities were observed for two doses of tetanus toxiod during the previous pregnancy (p=0.06) and full neonatal protection (p=0.04).

Figure 16 shows reported antenatal care in the HIDs and the comparison area in 1999 and 2006-7, with a midpoint in 2003. The proportion of women reporting of four or more ANC attendances and IPTp with two doses of SP increased significantly in both the HIDs and comparison area (p<0.001). Receipt of two doses of tetanus toxoid vaccination during the previous pregnancy did not increase in the HIDs, while increasing 14 pp in the comparison area. Absolute pp increases were significantly greater in the HIDs for IPTp with SP than in the comparison area (p<0.001). The differences in changes over time between the HIDs and the comparison area for four ANC attendances and TT2 were not statistically significant.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 39

Figure 16: Coverage levels of antenatal indicators in HIDs and the comparison area as measured in DHS in 1998-9 and 2003, and MICS 2006-7, Ghana.

§

§ § §

§ Estimate based on less than 100 women Note: Figures inside bars represent percentage coverage

Skilled attendant at delivery and postnatal care. Figure 17 shows trends in deliveries assisted by a skilled attendant and postnatal supplementation with vitamin A as reported by women giving birth within the 12 months before the survey. Assisted deliveries by trained workers included those attended by a doctor, nurse, midwife or auxiliary midwife. Additional data concerning these indicators are available in appendices G, H and I. Assistance at delivery by a skilled provider increased from 17 percent in 1998-9 to 40 percent in 2007 (p<0.001). However, more than half of women giving birth do not benefit from a skilled attendant at delivery and coverage levels fell far short of the initial ACSD target of 80 percent. Supplementation with vitamin A within 40 days after birth was high at baseline (72%) and declined over the period from 1998-9 to 2007 (p=0.01).

Appendix table I10 presents the breakdown of skilled delivery and postnatal vitamin A supplementation in the HIDs by socio-demographic characteristics as measured in the 1007 Supplemental MICS. More than double the proportion of urban dwellers reported a skilled attendant at delivery (71%) than their rural counterparts (31%); (p<0.001). Coverage levels of skilled delivery were similar among women reporting no formal education or primary school; however, women with secondary or higher education were more likely to deliver with a skilled attendant (p<0.01). Women in the highest wealth quintile were more than three times as likely to have a delivery assisted by a skilled provider (77%) than women in the poorest households (23%); (p<0.001).

40 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure 17: Coverage levels of skilled attendant at delivery and postnatal vitamin A supplementation in the HIDs as measured in DHS (1998-9 and 2003), the ACSD survey (2003), and MICS (2006 and 2007), Ghana. 100

80

72

. 60 .. 4 57 § 51 52 40 47 40

Coverage (%) Coverage 20 27 17 18 § § § § § 0 Skilled attendant at delivery Postnatal vitamin A

DHS 1998-9 ACSD 2003 DHS 2003 MICS 2006 Sup MICS 2007

§ Estimate based on less than 100 women

Figure 18 shows coverage levels of skilled deliveries and postnatal supplementation with vitamin A as reported by women giving birth within the 12 months before the survey in the HIDs and the comparison area in 1999 and 2006-7, with a midpoint in 2003. Deliveries assisted by a skilled health provider increased by 23 pp in the HIDs and by 4 pp in the comparison area; the difference between the HIDs and comparison area in the rates of change was significant (p=0.01). Levels of postnatal supplementation with vitamin A decreased by 15 pp in the HIDs, while increasing by 25 pp in the comparison area (p<0.001).

IIP-JHU | Retrospective evaluation of ACSD in Ghana 41

Figure 18: Coverage levels of skilled attendant at delivery and postnatal vitamin A supplementation in HIDs and the comparison area as measured in DHS in 1998-9 and 2003, and MICS 2006-7, Ghana. High-impact districts (UER) Absolute change in percentage points between 1998-9 and 2006-7 Comparison area

+4 +25

100 +23 -15

80 ...

60 Coverage (%)

40

20 § § § § 17 38 18 35 40 42 72 23 58 36 57 48 0 1998-9 2003 2006-7 1998-9 2003 2006-7 Skilled attendant at delivery Postnatal vitamin A

§ Estimate based on less than 100 women Note: Figures inside bars represent percentage coverage

5.3 Summary and interpretation of results

Table 10 summarizes the main results of the adequacy analyses of time trends in coverage in the HIDs. In table 10, we present indicators showing significant improvement between 1998-9 and 2007 in the HIDs in bold italics. Indicators showing positive trends over time in the HIDs included vaccinations, vitamin A, ITNs, antibiotics for suspected pneumonia, timely initiation of breastfeeding, exclusive breastfeeding, antenatal care, IPTp and skilled assistance at delivery. Indicators that were observed to stagnate or decline included case management of common childhood illnesses, tetanus toxoid vaccination and postnatal vitamin A. In the last column of table 10, we present the stated ACSD targets in Ghana in relation to the coverage levels measured in 2007. Many of the stated objectives were fully met or exceeded; however, management of diarrhea, complementary feeding and skilled delivery were 30 pp or more short of the stated ACSD targets for coverage.

42 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table 10: Summary of ACSD coverage results in HIDs as measured in DHS 1998-9 and 2003 and Supplemental MICS 2007, Ghana, as compared to initial ACSD objectives. MIDLINE ENDLINE BASELINE LEVEL IN LEVEL IN CHANGE ABSOLUTE LEVEL IN 2003 2007 1999 to DIFFERENCE 2007 ACSD BETWEEN COVERAGE 1999 DHS DHS MICS TARGET ENDLINE & INDICATOR (%) (%) (%) P value (%) OBJECTIVE Measles <0.001 vaccine 60 68 80 80 0 DPT3 68 76 95 <0.001 80 +15 Vitamin A to <0.001 child 65 86 90 80 +10 ITN for child ~0 23 58 <0.001* 50 +8 Any antimalarial <0.001 for fever 78 71 53 50 +3 Careseeking for >0.10 pneumonia 54 66 50 50 0 Antibiotics for 2 - 51 <0.001 50 +1 pneumonia Oral rehydration 0.05 for diarrhea 39 32 28 80 -52 Breastfeeding <0.001 initiation 11 85 52 n/a n/a Exclusive <0.01 breastfeeding 28 43 55 50 +5 Complementary - feeding - - 53 90 -37 Antenatal care 64 64 81 <0.001 80* +1 (4+ visits) IPTp with 2+ ~0 4 67 <0.001* 75 -8 SP TT2 in pregnancy 63 33 63 >0.10 n/a n/a Full neonatal TT protection n/a n/a 78 n/a 80 -2 Skilled <0.001 delivery 17 18 40 80 -40 Postnatal 0.01 vitamin A 72 58 57 n/a n/a *Changes calculated assuming 0% coverage at baseline NOTE: Indicators in bold italics represent significant positive changes over time

Table 11 summarizes the main results of the plausibility analysis, comparing time trends in coverage for HIDs and the comparison area. Estimates that showed a positive trend over time in HID that were significantly greater than the comparison area at p<0.05 are shown in bold italics. Vaccination, vitamin A supplementation for children aged 6-59 months, ITN utilization among children, antibiotics for pneumonia, exclusive breastfeeding, four or more antenatal care visits, and IPTp all improved by 10 pp or more in both HIDs and comparison area; ITNs, antibiotics for pneumonia, and IPTp increasing significantly more in the HIDs. Breastfeeding initiation and delivery assisted by a skilled worker improved by more than 10 pp in the HIDs, while increasing less than 10 pp in the comparison area. Two doses of tetanus toxoid during pregnancy, careseeking for suspected pneumonia and postpartum vitamin A stagnated (=/- 9pp) in the HIDs, while increased by more than 10 pp in the comparison area. Appropriate management of childhood fever and diarrhea decreased in the HIDs, while stagnating in the comparison area (difference in difference test were statistically significant). These results suggest that ACSD as implemented in the

IIP-JHU | Retrospective evaluation of ACSD in Ghana 43

HIDs in Ghana had a positive effect on levels of coverage for some of the interventions targeted for accelerated implementation.

Table 11: Summary of ACSD coverage results in HIDs and the comparison area as measured in as measured in DHS 1998-9 and 2003, MICS 2006 and Supplemental MICS 2007, Ghana. DIFFERENCE IN ABSOLUTE CHANGE DIFFERENCES BASELINE MIDLINE (% POINTS) TEST (p LEVEL) VALUE IN VALUE IN 1998-9 2003 COVERAGE 1999-9 2003 1998-9 to 2003 to to to INDICATOR AREA (%) (%) 2006-7 2006-7 2006/7 2006-7 HIDs 60 68 +20 +12 Measles vaccine >0.10* >0.10* Comparison 60 67 +18 +11 HIDs 68 76 +27 +19 DPT3 0.001 <0.001 Comparison 65 76 +14 +3 HIDs 65 86 +25 +4 Vitamin A to child <0.001 0.01 Comparison 22 79 +74 +17 HIDs ~0 23 +58 +35 ITN for child** <0.001 <0.001 Comparison ~0 3 +24 +21 Any antimalarial HIDs 78 71 -25 -14 <0.001 >0.10 for fever Comparison 60 67 +1 -6 Careseeking for HIDs 54 66 -4 -16 0.04 >0.10 pneumonia Comparison 22 40 +14 -4 Antibiotics for HIDs 2 n/a +49 n/a <0.01 n/a pneumonia Comparison 20 n/a +15 n/a Oral rehydration HIDs 39 32 -11 -4 <0.01 >0.10 for diarrhea Comparison 23 38 +7 -8 Breastfeeding HIDs 11 85 +41 -33 <0.001 0.001 initiation Comparison 23 41 +8 -10 Exclusive HIDs 28 43 +27 +12 >0.10 >0.10 breastfeeding Comparison 29 49 +22 +2 Antenatal care HIDs 64 64 +17 +17 >0.10 >0.10 (4+ visits) Comparison 55 58 +11 +8 HIDs ~0 4 ~+67 +63 IPTp with SP** <0.001 <0.001 Comparison ~0 1 ~+31 +30 Tetanus toxoid in HIDs 63 33 0 +30 0.12 0.04 pregnancy Comparison 46 47 +14 +13 HIDs 17 18 +23 +22 Skilled delivery 0.01 0.02 Comparison 38 35 +4 +7 Postnatal vitamin HIDs 72 58 -15 -1 <0.001 0.05 A Comparison 23 36 +25 +12 *P value based on children 12-13 months of age ever receiving measles or DPT3 vaccination **Difference in end-line estimates only assuming 0% coverage at baseline in HIDs and comparison area. NOTE: Indicators in bold italics represent positive changes over time in HID that were significantly greater than the comparison area at p < 0.05

Coverage results: contributions and challenges of ACSD implementation.

Preliminary results were reviewed and discussed with a technical team from Ghana that included those directly involved in ACSD implementation and/or the collection and analysis of the data used in the evaluation. We have incorporated the interpretation of results based on discussions with the Ghana technical team and review coverage in the context of ACSD implementation by each ACSD component, as well as overall ACSD contributions and challenges associated with changes in coverage.

44 IIP-JHU | Retrospective evaluation of ACSD in Ghana

EPI+. ACSD in Ghana had set targets to achieve 80 percent coverage for vaccination and vitamin A coverage and reached these goals; routine monitoring system data in the HIDs mirrored these trends. ACSD started with the EPI+ strategy, which focused primarily on preventative services delivered through campaigns and outreach. Key informants noted that ACSD’s key contributions in achieving the EPI+ targets included the supply of commodities and clear targets at the district level. Vaccination and vitamin A supplementation coverage increased in both HIDs and comparison area, although DPT3 increased more rapidly in the HIDs. The GHS and other development partners supported these activities in the comparison area, discussed in section 3 and below in contextual factors.

The promotion, distribution and re-treatment of ITNs were large components of the ACSD strategy in Ghana and elsewhere. Our results show large increases in coverage with ITNs between 1999 and 2006- 7 in the HIDs, with fewer gains in coverage in the comparison area. ACSD supplied large quantities of ITNs early, with other donors such as the Global Fund also providing support for ITN interventions. Key informants reported widespread stock-outs in ITNs, starting in late 2005 and persisting until late 2006, although other partners, such as the Global Fund provided ITNs at this time. UNICEF, with substantial funding from DFID, supported a national campaign to distribute two million long-lasting ITNs in late 2006, with approximately 90,000 LLITNs distributed in the HIDs.

IMCI+. ACSD efforts included expanding coverage and strengthening existing community-based systems, primarily by training, equipping and supporting CBAs to treat childhood illness in the community in the HIDs. However, we found that case management practices for fever and diarrhea declined or stagnated in the HIDs. Administrative data from communities showed similar trends. Likewise, careseeking for pneumonia to an appropriate health provider did not increase, although antibiotic treatment for pneumonia increased. Secondary analyses revealed that this increase was at least partially driven by antibiotics from shops and drug peddlers. The government and partners scaled up C-IMCI before training facility-based staff in IMCI, which may have lessened the synergistic effect of the full IMCI package.

Key informants and program documents noted important challenges to implementing the case management of fever. The GoG changed the first-line antimalarial policy to ACTs in 2004 due to high levels of resistance to chloroquine in Ghana. Although our indicator for treatment of fever appears to have declined marginally in the HIDs, this does not necessarily represent effective management of fever with the recommended first-line antimalarial. Only nine percent of caretakers reported that febrile children received an ACT in 2007 in the HIDs, and fewer than five percent did so in comparison area. Although health facilities and CHPS centers stocked ACTs in 2005, national policy did not allow ACTs at the community level (through CBAs) until 2007. CBAs retained chloroquine in their health kits until it was retired from the communities at the end of 2006. Roll out of ACTs in the community began just after collection of our endline data.

At the end of 2004, CIDA funding was mostly depleted and this affected the constant supply of commodities, especially drugs for managing sick children. Many CBAs, who had previously focused on treating children with fever did not have antimalarials (chloroquine) and reportedly saw fewer sick children for fever, as well as diarrhea, and ARI referrals. Sporadic stock-outs of antimalarial drugs were also noted at facilities. Program implementers reported that gaps in supplies and the end of incentives for health workers (both facility-based and community-based) linked to the end of CIDA hampered the continuation of some ACSD activities.

Key informants reported that the C-IMCI activities through CBAs were a great strength of ACSD, but challenging to sustain. Sufficient supervision and monitoring of the CBA system were reported as on- going issues. Incentives for CBAs were primarily limited to job aids and bicycles given at the beginning of implementation, and program implementers postulated that lack of on-going incentives and packages to increase CBA motivation limited the impact of C-IMCI activities, including community case management of child illnesses.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 45

Infant and young child feeding practices improved somewhat in the HIDs over the course of the study period. Program implementers noted that there were strong nutritional interventions through other donors and NGOs in the HIDs in the early part of the decade, while ACSD focused more strongly on the EPI+, ANC+ and case management of illness components. Nutritional interventions did not receive much emphasis, and some key informants reported a lack of coordination between ACSD activities and regional nutrition activities.

ANC+. Most ANC+ interventions increased over the period of evaluation. IPTp was introduced in 2004 in the HIDs and nationally, although coverage increased significantly more in the HIDs. This strategy was strongly supported by ACSD and the Global Fund in the HIDs. The proportion of women reporting a skilled attendant at delivery increased in the HIDs, but failed to reach the ACSD of 80 percent coverage. Key informants and documents cited explicitly that delivery care was not a focus of the ACSD approach in Ghana. Reported coverage with postnatal supplementation with vitamin A stagnated in the HIDs, while increasing in the comparison area. Measured levels of coverage were exceptionally high at baseline in the HIDs and we were not able to ascertain if this was due to statistical fluctuations or intervention coverage through previous projects.

Overall contributions and challenges of ACSD implementation. Early in the program, ACSD reinforced and strongly supported outreach activities for vaccination, ITNs and ANC services. ACSD built partnerships and built on what exists; this strategy was often cited as the “value-added” aspect of ACSD. Across the mix of child survival interventions, ACSD was also noted as contributing to capacity building and technical support for program implementers and partners. Key informants recognized that enhanced monitoring was important part of ACSD, but also noted that monitoring was weak and became weaker after initial CIDA funds for ACSD were depleted.

Support for ACSD through Canadian CIDA ceased in 2005, with important gaps in sustained external funding as discussed above. Gaps and delays in funding were cross-cutting, affecting: 1) the constant supply of commodities such as antimalarials and ITNs; 2) continued supervision and motivation of CBAs; 3) insufficient resources for recurring costs such as motorbikes, fuel and incentives for health providers; and 4) delays in the development of health promotion materials for CBAs and radio spots. Despite these constraints, other partners, including the GHS continued to support the ACSD activities, with large infusions of support provided by the government of the Netherlands and DANIDA in 2006.

Contextual factors.

The contextual factors considered in the evaluation were based on those proposed as relevant11 for child survival programs.11 Section 3 and appendices A and M provides a more comprehensive description of contextual factors. Given that the adequacy findings on coverage suggest that ACSD had positive effect on some indicators but not on others, the analysis of contextual factors here examines two questions to better interpret the results:

1. Were there any major disruptions in the HIDs or nationally that could explain why ACSD did not lead to a more marked effect on coverage levels?

2. Were there other activities outside of ACSD in the HIDs or nationally that could have led to increases in coverage in the HIDs?

Major disruptions. To our knowledge, there were no natural disasters or other emergencies in the HIDs from 1998 to present that would have influenced the effect of ACSD on intervention coverage. Flooding occurred during data collection for the Supplemental MICS 2007, used for endline estimates. In order to assess the impact of the flooding on the population, as well as on the MICS survey, we developed an additional questionnaire module to assess household damage and migration due to the flooding. Twenty-eight percent of the households in the HIDs reported affects of the flood and 24 percent reported damage to the household structure. This emergency would not affect our coverage measures of interventions delivered well before the survey data collection, such as vaccination, vitamin A supplementation, ANC visits, delivery care, etc.

46 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Disruptions due to the flooding may have negatively affected indicators related to management of childhood illness, which rely on a two-week recall period or use of ITNs, which pertain to the night before the survey. We performed secondary analyses to assess if household disruptions due to flooding may have affected these indicators. There were no significant differences in management of diarrhea or pneumonia or use of ITNs between households affected by the flood and those not affected. A significantly higher proportion of febrile children in flood-affected households received an antimalarial for fever (61%) than children in non-affected households (50%); (p=0.02). Key informants reported that the increase in malaria treatment in the flood areas was likely associated with an emergency procurement of ACTs funded through the European Commission’s Humanitarian Aid Office–the remaining non-flooded areas did not receive this additional funding. Coverage of case management interventions was also similar in the 2006 MICS, and it is unlikely that the flooding biased our coverage estimates.

Other activities in HIDs and comparison area. As part of the evaluation, we documented other health and development project activities in the HIDs and comparison area between 1999 and 2006-7; section 3 and appendix M provide further details. A number of multilateral, bilateral, and non-governmental agencies, as well as Navrongo Health Research Center, implemented similar and complementary interventions targeting child health and nutrition before and during ACSD implementation. It is difficult to quantify the contributions and population coverage of these programs, but activities widely implemented and supported outside of ACSD included: 1) supplementation with vitamin A for children; 2) ITNs; 3) promotion and support of appropriate infant feeding practices; and 4) community case management for common childhood illnesses. Between 1998 and 2007, coverage of CHPS compounds, also focusing on preventative and primary health care, expanded more rapidly in the HIDs than in comparison area. We were not able to find measures of intervention coverage associated with the CHPS strategy, although the literature suggests that the implementation of this strategy in the Kassena-Nankana district led to greater declines in child mortality than routine services or routine services with community volunteers only.34,36,42

Thus, changes in intervention coverage in the HIDs cannot be attributed to ACSD alone and must be viewed in light of prior and concurrent activities of other partners in the health sector. Additionally, the GoN and DANIDA provided substantial financial support in 2006 and 2007 to the HIDs and other northern regions for the HIRD strategy, the national GHS continuation of the ACSD strategy. Many other development programs in the HIDs focused on education and literacy, agriculture, poverty reduction through micro-credit, and water and sanitation. We would not expect these projects to have a large short- term influence on coverage of maternal and child health interventions, because they do not directly address the interventions.

As described in section 3 and appendix M, over the period of 1998 to 2007, Ghana benefited from massive investments in health at the national level. In the comparison area, USAID, WHO, and others supported child survival activities similar to those promoted by ACSD.

Summarizing the presentation on contextual factors:

• No major humanitarian or natural crises were found that affected the coverage results;

• A multitude of maternal and child health activities were implemented by development partners in the HIDs, some in close collaboration with ACSD;

• The expansion of CHPS compounds may have differentially improved access for preventative and curative care in the HIDs;

• Development partners supported activities similar to those included in the ACSD package in the comparison area.

Methodological Challenges.

Here we present a very brief overview of methodological challenges encountered in the retrospective evaluation of ACSD in Ghana, noting how they may have affected the evaluation results related to coverage. Complementing this section, appendix K provides a more thorough review of methodological challenges, appendix F provides descriptions of surveys included in the evaluation, and appendix D and

IIP-JHU | Retrospective evaluation of ACSD in Ghana 47

E provide indicator definitions and a list of the questionnaire items supporting the measurement of the priority indicators in each survey.

Many of the challenges encountered reflect the retrospective nature of the evaluation. The evaluation team was forced to rely existing data and information, even if imperfect.2. The 1998-9 and 2003 DHS had limited sample sizes for calculation of baseline coverage indicators in the HIDs, especially those indicators measured among small subgroups of the sample such as exclusive breastfeeding or careseeking for pneumonia. These small sample sizes affect the precision of point estimates and the statistical power to detect small differences over time.

Collection of data occurred approximately one year apart for the Supplemental MICS 2007 (used for endline coverage estimates in the HIDs) and MICS 2006 (used for endline coverage estimates in the comparison area). We compared estimates of coverage between 2006 and 2007 in the HIDs to assess if the one-year time lag could have influenced our results. Most coverage indicators remained relatively stable in the HIDs between 2006 and 2007, and were not statistically significant. ITNs for children and IPTp were significantly greater in 2007 as compared to 2006 in the HIDs; coverage with any antimalarial for fever was significantly less in 2007. For these three indicators, we reran statistical tests using the 2006 MICS as our endline estimate to identify any possible bias introduced by using the 2007 MICS survey only in the HIDs. Statistical inferences were the same for trends over time and differences in changes over time in the HIDs and comparison area.

The DHS and MICS use slightly different methodologies to collect data. DHS ask only biological mothers of young children about intervention coverage, while MICS questions caretakers of children, even if not biologically related, about intervention coverage. Appendices D and E note differences in the DHS and MICS questions used for indicator calculations; appendices F and K review the differences between the surveys. These differences were minimal and we would not expect them to affect the findings.

The data available in the 1998-9 DHS did not allow for calculation of all priority indicators for the evaluation, which are identical to those used for monitoring progress toward the Millennium Development Goals (MDG).7,8 In the 1998-9 DHS, several essential questions were not included: use of bednets by children or pregnant women, timing of antimalarial administration for febrile children, SP taken as part of IPT for pregnant women, or full neonatal tetanus toxoid protection. For the evaluation of time trends between 1998-9 and 2006-7, we used indicator definitions that could be calculated from the 1998-9 data to ensure comparability with indicator estimates in 2006-7 (see appendices D and E). These proxy indicator definitions were less stringent than the priority indicator in all cases; coverage estimates from 2006-7 using the more stringent, MDG priority coverage indicators are presented in appendices G, H and I.

Taken together, these methodological issues are not likely to influence the endline comparisons between the HIDs and national comparison area. Differences in the conduct of the survey, the DHS and MICS questionnaires and interviewers’ style of asking questions may have introduced some bias into the comparison of coverage levels between 1998-9, 2003 and 2006-7. However, these methodological challenges are not likely to change the main evaluation findings or conclusions in any substantial way.

48 IIP-JHU | Retrospective evaluation of ACSD in Ghana

6. Nutrition

In this section, we describe the differences in nutritional status of young children between the HIDs (comprised of the UER) and comparison area; the latter includes the rest of the country with the exception of the HIDs and urban Greater Accra and Ashanti regions (Accra and Kumasi).

Anthropometric data from the HIDs are available for the 1998-9 and 2003 DHS, and for the main MICS in 2006 and the supplementary MICS in 2007. The same surveys – except for the 2007 supplementary MICS – also provide data for the national comparison area. We used data from the 2006 MICS for the comparison area and from the 2007 supplemental MICS for the HIDs for the endline results. Section 2 explains the rationale in more detail.

Three indicators of undernutrition prevalence were calculated from these surveys: prevalence of stunting (low length for age for children below 24 months; low height for children 24-59 months of age), wasting (low weight for length/height), and underweight (low weight for age). We used the minus two z-score cutoff based on the 2006 WHO Growth Standards,12 to identify children with moderate or severe undernutrition; for severe undernutrition we used the minus three z-scores cutoff. Mean z scores of the three indices were also calculated. Appendix J presents a schematic of the inclusion and exclusion criteria for children included in the analysis.

We present results for all children less than five years of age. For stunting, results are also presented for children aged 24-59 months, the age group with the highest prevalence of this condition 43. Likewise, wasting results are described for children aged less than 24 months. Table 12 presents the numbers of children included in the analyses.

Presentation of the results follows the approach used in the section on coverage indicators. First, the adequacy findings are presented (time trends in the HIDs), followed by the plausibility results (comparison between HIDs and the rest of the country). Appendix J presents full nutrition results for sub-groups in both areas.

6.1 Results

Figure 19 shows that stunting decreased over time in the HIDs. Wasting and underweight remained relatively unchanged over time, with a peak observed in the 2003 DHS, possibly due to seasonality of surveys.

Table 12 and Figure 20 show results for the HIDs as well as the comparison area, in the 1998-9, 2003 and 2006-7 surveys.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 49

Figure 19: Time trends in stunting (children 24-59 months), wasting (children 0-23 months) and underweight (children 0-59 months) in the ACSD “high-impact” districts as measured in DHS and MICS in 1998-9, 2003 and 2006-7, Ghana. 100 ...

80

60 56 40 42 37 35 30 20 27 25 23 21 18 Prevalence of moderate and severe (%) and severe moderate of Prevalence 14 14 § § § § § § 0 Stunting (24-59m) Wasting (0-23m) Underweight (0-59m)

DHS 1998-9 DHS 2003 MICS 2006 Sup MICS 2007

§ Estimate based on less than 100 children

Stunting.

From 1998-9 to 2006-7, there was a reduction of 21 percentage points (pp) for children 24-59 months in the HIDs (p<0.001), compared to a four pp decline in the comparison area. The decline in overall stunting (moderate or severe) in the HIDs was mostly due to the reduction in the prevalence of severe stunting, which fell from 26 to 10 percent. The reduction in the comparison area was from 17 to 13 percent. Mean height/length for age also improved more markedly in the HIDs than in the comparison area. Similar patterns were also observed when all under-five children were analyzed. Despite the small baseline sample size in the HIDs, the difference in difference tests showed that the decline in the HIDs was significantly greater than in the comparison area (p<0.001).

Of the 21 pp reduction in stunting among children 24-59 months observed between 1998-9 and 2007 in the HIDs, the largest drop - of 14 pp - seems to have occurred between 1998-9 and 2003, before ACSD was fully implemented (p<0.01). The seven pp reduction between 2003 and 2007 was not significant (p>0.10). Nevertheless, the confidence intervals for these estimates are wide due to the small sample sizes.

50 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table 12: Summary of anthropometry results in ACSD “high-impact” districts and comparison area as measured in DHS 1999, 2003 and MICS 2006-7, Ghana.

High impact Comparison: districts: Change in Change in 1998 DHS 2003 DHS 2007 Sup. MICS 2006 MICS baseline to 2007 baseline to 2007

HIGH IMPACT GEOGRAPHIC HIGH IMPACT GEOGRAPHIC HIGH IMPACT GEOGRAPHIC 1999 to 2003 to 1999 to 2003 to DISTRICTS COMPARISON¥ DISTRICTS COMPARISON¥ DISTRICTS COMPARISON¥ 2007 2007 2007 2007 % / 95% CI % / CI / % / 95% CI % / 95% CI % / 95% CI % / 95% CI NUTRITIONAL INDICATOR n mean / (SD) n mean (SD) n mean / (SD) n mean / (SD) n mean / (SD) n mean / (SD) p p p p Stunting (height for age) 24-59 months % stunted (< -2 SD) 50-63 38-45 28-56 44-50 32-38 33-40 <0.001 >0.10 0.04 <0.001 97 56 1138 41 90 42 1316 47 1336 35 1385 37 % severely stunted (< -3 SD) 26 20-32 17 15-19 18 11-25 18 16-21 10 9-12 13 11-15 <0.001 <0.01 0.01 <0.01 mean Z score (sd) -2.1 (1.4) -1.8 (1.3) -1.8 (1.2) -1.9 (1.3) -1.6 (1.2) -1.6 (1.2) <0.001 0.08 <0.01 <0.001 0-59 months % stunted (< -2 SD) 38-49 30-35 29-44 37-41 27-31 29-33 <0.001 0.03 >0.10 <0.001 168 44 2068 32 143 36 2277 39 2192 29 2429 31 % severely stunted (< -3 SD) 19 13-24 13 12-15 14 9-20 15 14-17 9 8-11 11 9-13 <0.001 0.03 0.05 <0.001 mean (sd) -1.6 (1.6) -1.4 (1.5) -1.4 (1.5) -1.6 (1.5) -1.3 (1.3) -1.3 (1.4) <0.001 >0.10 >0.10 <0.001 Wasting (weight for height) 0-23 months 7-21 15-20 18-36 11-15 11-17 7-11 >0.10 <0.001 <0.001 0.01 % wasted (< -2 SD) 71 14 952 17 48 27 955 13 865 14 1075 9 % severely wasted (< -3 SD) 5 0-10 4 3-6 10 3-18 4 2-5 4 3-6 2 1-4 >0.10 0.02 0.04 >0.10 mean (sd) -0.9 (1.3) -0.8 (1.3) -1.1 (1.4) -0.5 (1.4) -0.7 (1.3) -0.5 (1.2) >0.10 0.04 <0.001 >0.10 0-59 months % wasted (< -2 SD) 4-12 9-12 7-18 6-9 7-10 5-7 >0.10 0.04 <0.001 0.02 171 8 2186 11 137 13 2264 8 2226 8 2603 6 % severely wasted(< -3 SD) 3 1-5 2 2-3 4 1-7 2 1-2 2 1-3 2 0-2 >0.10 >0.10 0.03 >0.10 mean (sd) -0.6 (1.1) -0.5 (1.2) -0.7 (1.2) -0.2 (1.2) -0.5 (1.1) -0.2 (1.1) >0.10 >0.10 <0.001 >0.10 Underweight (weight for age) 0-59 months % underweight (< -2 SD) 25 21-30 22 20-24 30 21-38 20 18-22 21 19-23 15 13-17 0.03 <0.01 <0.001 <0.001 173 2121 145 2317 2230 2504 % severely underweight (< -3 SD) 7 5-9 6 5-8 9 6-13 5 4-6 5 4-6 4 3-5 0.10 <0.01 0.01 >0.10 mean (sd) -1.3 (1.2) -1.1 (1.2) -1.4 (1.2) -1.0 (1.2) -1.1 (1.1) -0.9 (1.1) 0.02 <0.01 <0.001 0.03 ¥Comparison area comprised of Ghana national level, minus urban Greater Accra and Ashanti regions

IIP-JHU | Retrospective evaluation of ACSD in Ghana 51 Figure 20: Prevalence of stunting, wasting and underweight, and absolute change in percentage points in the HIDs and comparison area as measured in DHS in 1998-9 and 2003, and MICS 2006-7, Ghana.

Absolute change in percentage points between 1998-9 to 2006-7 High-impact districts (UER) Comparison area -4 -8 -7 100

... -21 0 -4

80

60

40

§ 20

Prevalence of moderate and severe (%) § § § 56 41 42 47 35 37 14 17 27 13 14 9 25 22 30 20 21 15 0 1998-9 2003 2006-7 1998-9 2003 2006-7 1998-9 2003 2006-7 Stunting (24-59m) Wasting (0-23m) Underweight (0-59m)

§ Estimate based on less than 100 children

Wasting.

Time trends in wasting should be interpreted with caution because of small sample sizes in the HIDs that are reflected in the wide confidence intervals. For children under 24 months, the prevalence of wasting declined between 1998-9 and 2006-7 by eight pp in the national comparison, but there was no decline in the HIDs. This difference was statistically significant (p=0.03) in favor of the comparison area.

There was a marked peak in prevalence in 2003 in the HIDs, but this estimate is based on only 48 children. Trends in severe wasting and in mean weight-for-length showed similar patterns. No significant differences were observed between the time trends in intervention and comparison area for the period 2003 to 2006-7.

Underweight.

The analyses of underweight included all under-five children. Using the 1998-9 baseline, there was a decline of four pp in the HIDs and seven pp in the comparison area. As for wasting, underweight prevalence showed a peak in 2003 in the HIDs, but not in the comparison area. After this peak, prevalence declined by nine pp in the HIDs and by five pp in the comparison area. Trends in severe underweight and in mean weight-for-age showed similar patterns. The difference-in-differences tests were not significant for the 1998-9/2006-7 period or for the 2003/2006-7 period.

5.3 Summary and interpretation of results

Summary.

Stunting. According to national surveys, the prevalence of stunting in under-five children in the comparison area increased between 1998-9 and 2003 and declined between 2003 and 200616,44 The decline between 2003 and 2006 is probably related to overall socioeconomic progress and improvements in health care and coverage of preventive interventions as described in section 3. Stunting is primarily influenced by dietary quality and quantity, as well as by the incidence and severity of infections. Coverage of interventions for preventing infections, such as ITNs and vitamin A, increased substantially in the HIDs and comparison area. There are insufficient data to assess time trends in complementary feeding, so that these cannot be related to changes in stunting prevalence.

Our results suggest that there was a substantial decline in stunting prevalence in HIDs between 1998-9 and 2003, compared to absence of a decline in the comparison area during the same period. After 2003, the reductions in stunting prevalence were maintained and improved upon, although the rates of decline appear to have been similar in HIDs and comparison area.

Wasting. From 1998-9 to 2006, the prevalence of wasting among children less than five years of age has steadily declined in Ghana as a whole, from 10 percent in 1998-9 to 5.4 percent in 2006.16,44 Our results show an apparent increase in the prevalence of wasting in the HIDs, but not in the comparison area, between 1998-9 and 2003, followed by a reduction of the same magnitude. Looking at the whole period, from 1998-9 to 2006-7, there was a reduction by almost half in the prevalence of wasting in the comparison area that was significantly different from the lack of progress observed in the HIDs.

Underweight. For Ghana as a whole, the prevalence of underweight has shown similar trends as wasting, declining from 25 percent in 1998-9 to 18 percent in 2006.16,44 We observed these declining trends in both HIDs and the comparison area, with no significant differences between the two.

Cross-cutting implementation and contextual factors.

Here we review factors that may affect the comparisons described above.

Socio-economic status. Poverty is associated with stunting (see section 8 on equity for further discussion) and as discussed in section 3, the HIDs were significantly poorer than comparison area. Thus, our stunting results could be affected by this imbalance of poverty. We used direct standardization to estimate the stunting prevalence in HIDs in 2007, had these areas presented a similar wealth distribution to that in the comparison area. The standardized prevalence in 2007 in the HIDs became 31 percent, compared to the crude prevalence of 35 percent in HIDs and of 37 percent in the comparison area. The small number of children available in the earlier surveys in the HIDs does not allow breakdown by socioeconomic position, but because the HIDs are historically poorer than the rest of the country, the time trends are unlikely to be affected.

Presence of other nutritional interventions or programs in the HIDs. Understanding of the role of nutritional interventions requires a discussion of the timing of growth faltering. The active process of stunting, or growth faltering, occurs up to the age of 24 months, and thereafter prevalence remains constant up to five years of age. The most sensitive indicator, therefore, is the prevalence of stunting among children age 24-59 months, who are already fully “stunted.” However, for ACSD to have an impact on stunting, children should be exposed to it during their first two years of life when active faltering, or stunting, is occurring. For this reason, there is a lag between the time of the intervention and the time when an impact on height-for-age can be measured. Most of the reduction in stunting in the HIDs appears to have happened between 1998-9 and 2003 (figure 19), which means that whatever caused this reduction must have happened at least a couple of years before the 2003 survey. Implementation of ACSD started in 2002, thus ACSD activities cannot explain this reduction. Interviews

IIP-JHU | Retrospective evaluation of ACSD in Ghana 53

with key informants and review of documentation showed that strong nutritional interventions (presented in section 3), such as the establishment and support of feeding and nutritional rehabilitation centers and the LINKAGES project which focused exclusively on infant feeding practices,45 were present in the HIDs before the launch of ACSD. These activities are a possible explanation for the marked reduction in stunting prior to 2003.

Natural occurrences. We investigated reasons for the apparent increase in wasting prevalence in HIDs between 1998-9 and 2003. Unlike stunting, changes in wasting can occur soon after a change in causal factors, because it usually reflects acute weight loss. A potential cause of sharp increases in wasting is food shortage, but our interviews with key informants and reviews of the documentation did not indicate that this was the case. It is possible that the apparent increase in wasting is due to statistical fluctuation given the small sample size of fewer than 50 children in the HIDs in 2003.

Summing up, after consideration of other factors, there was still no evidence of a differential impact of ACSD on any of the three nutritional indicators studied. As will be discussed below (section 9 on conclusions), this is consistent with the finding on coverage of interventions with a potential impact on nutrition.

54 IIP-JHU | Retrospective evaluation of ACSD in Ghana

7. Mortality

This section reports on changes in child mortality in the HIDs and mortality trends in the comparison area. The methods used to estimate mortality for the results presented in this section differ from those used for coverage and nutrition analyses. The methods also differ by HIDs and comparison area due to data availability. Below we provide a brief review of the analyses used for mortality estimation for the HIDs and comparison area separately.

Mortality estimation in the HIDs.

We used the full birth history data collected in the 2007 Supplemental MICS to estimate child mortality in the HIDs both before (baseline) and after ACSD became operational (endline). There are two reasons why we elected to use the 2007 survey as the basis for estimating mortality throughout the evaluation period. First, the use of a full birth history allows the calculation of period estimates of mortality from the previous year to 10 or more years in the past because a child’s birth and/or death is very significant to the mother and generally can be recalled reliably. Second, using a single survey to estimate mortality for the two periods – baseline and endline – builds on the correlation between periods arising from use of the same sample of households. This usually reduces the sampling error of the difference in mortality between the two periods, enabling smaller differences to be measured more precisely. Third, this method reduces the impact of non-sampling errors since there is generally more consistency of non-sampling errors within a survey than between surveys.

Whether one or more surveys are used to estimate mortality, larger sample sizes are associated with more precise estimates of mortality. Thus, we want to maximize the sample size by selecting longer time periods for mortality estimation. These periods need to be consistent with ACSD implementation and the baseline period should not extend far into the past, as this would result in a higher mortality estimate before initiation of ACSD in contexts where mortality levels are declining over time. We calculated mortality for two periods of the same length, 3½ calendar years each, before and after ACSD implementation in the HIDs. As shown in Figure 21, based on the documentation of ACSD implementation, we defined the baseline period as July 1998 to December 2001, and the full implementation period as January 2004 to July 2007, with a phase-in period in between baseline and full implementation.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 55

Figure 21: ACSD implementation time periods in Ghana for the retrospective mortality analysis using full-birth history data, based on documentation of ACSD implementation.

YEAR (from full birth history) Time periods used in mortality analysis 1998

1999

2000 A. BASELINE: before implementation ACSD Jul 1998-Dec 2001 2001

2002 B. PHASE-IN: start of 2003 ACSD interventions Jan 2002-Dec 2003 Compare U5MR 2004

2005 C. ENDLINE- Full implementation ACSD 2006 Jan 2004-Jul 2007

2007

A. BASELINE: No ACSD implementation: start of period chosen for symmetry with period C B. PHASE-IN: Start reinforcement of EPI & cold chain; donation of vehicles, motos & refrigerators; ITNs Vitamin A; reinforcement of ANC activities C. ENDLINE = EPI + ITNs; Vitamin A; ANC; CHW training & deployment; Facility IMCI

The under-five mortality rate (U5MR) is our priority indicator for measuring changes in mortality in the HIDs, because the primary goal of the ACSD project was to reduce it by 25 percent by the end of 2006.46 One benefit of using U5MR relative to other measures of child mortality (see Box 3) is that it provides the largest sample size and is less sensitive to errors in reporting age than infant or neonatal mortality. Although we present findings for specific age groups within 0 to 59 months, we have considered U5MR as the primary indicator of mortality impact.

Mortality estimation in the comparison Box 3: area. Measures of child mortality (expressed as deaths per 1,000 live births) In Ghana, there is no single recent household survey with a full birth history to Neonatal mortality The probability of dying between birth generate comparable direct child mortality (NN) and the first month of life estimates for the comparison area (defined as the rest of Ghana minus the HIDs, and Post-neonatal The probability of dying between the urban Greater Accra and Ashanti regions). mortality exact age of one month and the exact The most recent national survey in Ghana age of one year with a full birth history is the 2003 DHS. The probability of dying between birth The 2006 Ghana MICS used the Brass- Infant mortality (IMR) and exact age one year type questions on children even born and children surviving. These questions only Child mortality (CMR) The probability of dying between provide indirect estimates of child mortality, exact ages one and five years which cover a period of up to 15 years before the survey. Under-five mortality The probability of dying between birth (U5MR) and exact age five years Thus, we use the available data from both the 2003 DHS and 2006 MICS to estimate and project trends in the under-five mortality rate for the

56 IIP-JHU | Retrospective evaluation of ACSD in Ghana

comparison area. We focus on U5MR since this is most robust indicator of child mortality, as described above. Yearly direct mortality estimates were calculated from the DHS 2003 and then averaged over a two-year period. The indirect estimates of under-five mortality (from DHS 2003 and MICS 2006) had their most recent two points excluded. The most recent point is always excluded by demographers as being too inaccurate, and the second most recent is recognized as often being biased, usually to higher levels of mortality. We then fit a trend line to all the available data points to estimate increases or declines in mortality in the comparison area, described further in section 7.1.

7.1 Results

Figure 22 presents the annual direct mortality estimates of U5MR in the HIDs from 1997 to 2007, as well as the estimated mortality trend in the comparison area. Mortality over the last 10 years is declining in the HIDs (95% confidence intervals shown in dashed, red lines). The U5MR in the comparison area is estimated to have stagnated over the last 10 years, although trends from 2004 forward are projected from available data (shown in blue dashed bar) and should be interpreted with great caution.

Figure 22: Annual estimates of under-five mortality rates in the HIDs as measured in the 2007 Supplemental MICS and estimated levels of mortality in the comparison area, 1997-2007, Ghana. 180

160

140

120

100

80

60 U5MR (deaths per 1000 births) 40

20 High-impact High-impact - 95% confidence bounds National comparison - estimated National comparison - projected 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Year

Note: Projected mortality in comparison area shown in dashed blue bar

HIDs.

Table 13 presents several age-specific mortality rates in the periods before ACSD implementation and after full implementation, as well as the absolute reduction over time expressed as deaths per thousand births. We present the 95 percent confidence limits for these estimates, as well as the p-value for comparisons of estimate between baseline and endline.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 57

Table 13: Mortality rates by time period and changes between baseline and endline periods in the HIDs, Ghana. HIGH IMPACT ZONES

DIFFERENCE IN BASELINE AND ENDLINE (A minus C) A. JUL 1998 - C. JAN 2004 - Absolute MORTALITY MEASURES DEC 2001 JUL 2007 difference p value Priority Indicator Under-five mortality (5q0) 106.7 86.2 20.6 0.1 95% CI (87.9 - 125.6) (72.2 -100.1) (-4.3 to 45.5) Age-specific indicators Neonatal mortality (NN) 38.4 26.3 12.1 0.07 95% CI (28.5 - 48.4) (17.8 - 34.8) (-1.3 to 25.5)

Postneonatal mortality (PNN) 20.4 26.9 -6.4 >0.10 ACSD Phase-in period period ACSD Phase-in 95% CI (13.9 - 27.0) (18.2 - 35.6) (-17.8 to 5.0)

Infant mortality (1q0) 58.9 53.2 5.7 >0.10 95% CI (47.5 - 70.2) (41.4 - 64.9) (-12.0 to 23.4)

Child mortality (4q1) 50.9 34.9 16 0.05 95% CI (37.3 - 64.4) (26.7 - 43.0) (-0.2 to 32.2)

The U5MR decreased over time in the HIDs, from 106.7 in the period of July 1999 to December 2001 to 86.2 in the period of January 2004 to July 2007, representing a decline in UM5R of approximately 20 percent. This decline in U5MR just failed to reach statistical significance (p=0.10). Neonatal and child mortality showed the fastest relative declines (32% and 31%, respectively, with p levels of 0.07 and 0.05), while infant mortality decreased by 9.7 percent (p>0.10). Postneonatal mortality was observed to increase, although this change was not statistically significant (p>0.10).

Comparison area.

As described above, no comparable data was available for the comparison area. Figure 23 presents the trend line as estimated from available direct and indirect mortality estimates. The U5MR remained approximately constant between 1994 and 2003 at 115 deaths per 1000 live births. We can project this estimate forward to 2006 to cover the ACSD period; however, the uncertainty of this projection increases as it gets further from 2003. Nevertheless, these estimates based on available data suggest that U5MR has been constant in the comparison area for much of the period of the ACSD, but with considerable uncertainty in the period since 2004.

58 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure 23: Estimated and projected under-five mortality rate in the comparison area as measured using indirect and direct mortality estimates from DHS 2003, and indirect estimates from MICS 2006, Ghana.

160

140

120

100

80

60

Excluding women aged 20-24 yrs data in trend line

U5MR (deaths per 1000 births) 40

DHS 2003 direct DHS 2003 indirect 20 MICS 2006 indirect Linear trend

0 1990 1992 1994 1996 1998 2000 2002 2004 2006 Year

7.2 Summary and interpretation of results

Based on these findings, the U5MR in the HIDs implementing ACSD in Ghana declined by 20 percent in the period from 1998 to 2007—from 107 to 86 per 1,000 live births (p=0.10). Based on available data, the U5MR was estimated to stagnate at approximately 115 per 1000 live births through 2003 in the comparison area where ACSD was not implemented.

Cross-cutting implementation and contextual factors.

We considered the implementation and contextual factors that might have offset the impact of ACSD, with special attention to factors that would have influenced the HIDs and comparison area differentially.

As presented in section 3, we observed that the HIDs were significantly poorer than the comparison area. Available data suggested higher rates of mortality in the less poor comparison area and mortality in the HIDs did not show a strong social gradient (see next section), thus it is difficult to assess how socio- economic status might affect our estimates of mortality.

As discussed in section 3 and previous results sections, many child health and nutrition activities took place outside of routine services in the HIDs before and during ACSD. These activities, in addition to other development activities, such as improvements in water and sanitation, most likely contributed significantly to the observed declines in mortality. Evidence reported in the literature suggests that the CHPS strategy as implemented in the Kassena-Nankana district (one of the HIDs) led to declines in child mortality.34,36,42 Coverage of CHPS in the HIDs greatly expanded over the evaluation period, although with lower population coverage than in the original experiment conducted by the Navrongo Health Research Center taking place in Kassena-Nankana district.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 59

Thus, as noted above, the observed declines in mortality in the HIDs must be interpreted as a result of broad efforts to improve child survival and health in these areas. Despite large-scale health programming in the southern and central regions of the country and substantial investments in health at the national level, available data suggest that mortality in the comparison area stagnated.

Methodological Challenges.

There are important methodological issues that may have affected the results of this retrospective estimation of the effect of ACSD on under-five mortality. Appendix K provides a complement to this section with a more detailed discussion of these issues.

Mortality data in comparison area. Our primary methodological limitation is the lack of comparable data in the comparison area. We were not able to estimate mortality, using direct estimates from the full birth history, for the same time periods in the comparison area as in the HIDs. We estimated mortality trends in the comparison area using all available data (direct estimates from the DHS 2003 and indirect estimates from the MICS 2006 and DHS 2003) to assess trends. Although this provides an estimate of mortality trends, it is a composite, based on different surveys and different methodological assumptions in the calculation of mortality. Additionally, these surveys only provide data that refer to mortality experiences up to 2004; the overall trend was projected to 2007, but with the large uncertainty associated with any projection. The incomparability of the data and methodologies between the two areas also precludes statistical comparisons of changes in mortality between the two areas. More appropriate comparison data would consist of direct estimation of mortality through full-birth histories collected in the neighboring districts in the Northern and Upper West regions during the Supplemental MICS 2007—the same survey and methodology used for estimation in the HIDs. In spite of numerous attempts, we were unable to obtain such data in a timely fashion.

Definition of the “before” and “after” periods of ACSD A second methodological challenge was the definition of the “before” and “after” periods of ACSD implementation. Documentation of implementation is difficult in a retrospective evaluation, and is based by necessity on records maintained for other purposes and the subjective recall of project implementers. The two periods defined for the purpose of this evaluation were discussed and agreed to with in-country teams composed of ACSD implementers and national counterparts, and we believe that they accurately distinguish between periods before ACSD was implemented and periods during which ACSD was “fully implemented” in the views of those responsible.

In summary, despite these methodological challenges, there is sufficient evidence to conclude that that there was a reduction in child mortality in the HIDs from before to after ACSD was implemented, that just failed to reach statistical significance. The 20 percent reduction in U5MR between the two periods comes close to the reduction goal for the ACSD project of 25 percent by the end of 2006. At the same time, available data suggest that U5MR has stagnated in the comparison area, at least through 2004. In our conclusions, we discuss how these findings relate to the results on coverage and nutrition.

60 IIP-JHU | Retrospective evaluation of ACSD in Ghana

8. Equity of coverage, nutrition and mortality

In addition to evaluating the impact of ACSD implementation on indicators of coverage, undernutrition and mortality, it is also important to assess whether or not the strategy helped reduce inequities in health.

In this chapter, we describe within-population inequalities according to socioeconomic position, place of residence and sex, separately for the HIDs and for the comparison area. Socioeconomic position was categorized according to wealth quintiles, obtained from an index based on ownership of household assets and building characteristics (described in Appendix D). The definition of urban or rural residence was based on survey classification, derived from the 2000 Ghana Housing and population census.

The sample size available in 1998-9 and 2003 DHS for the calculation of baseline indicators in the HIDs was small. Because equity analyses require breakdown of these already small samples into two to five subgroups, it was not advisable to carry these out for the coverage and nutrition data. Given the over- sampling of the HIDs at endline in the Supplemental MICS in 2007, it was possible to carry out equity analyses for the post-implementation period. Our analyses, therefore, will be restricted to documenting how inequalities differ between the two areas after ACSD implementation. We also attempted to investigate ethnic group inequalities, but except for the Mole-Dagbani ethnicity, no other group accounted for more than 10 percent of the sample in both the HIDs and comparison area, and thus it was not possible to compare ethnic inequalities across areas.

Families in the HIDs were markedly poorer than those in the national comparison area (see figure 4 and table 2 in section 3). For example, only nine percent of the under-five children belonged to the wealthiest quintile based on the national sample). The small sample size in the upper quintiles in the HIDs should be borne in mind when interpreting the results.

Appendix I presents the breakdown of all coverage indicators according to sex and wealth quintiles within the HIDs zones in 2007. Due to the imbalance in the number of children in each wealth quintile in the HIDs when the combined samples were used, the analyses in appendix I relied on a different asset index, based exclusively on the HID sample in order to produce quintiles with approximately equal number of children. The results in Appendix I, therefore, may differ from those presented in this chapter.

In this section, we present results for both the HID and comparison area, but restrict the results to six coverage indicators representing the different components of ACSD. These include EPI+ (measles vaccine, ITNs for children and vitamin A to children), IMCI+ (diarrhea management / ORT) and ANC+ (three or more antenatal visits, skilled attendant at delivery). We also carried out equity analyses for the two main indicators of impact: stunting among children aged 24-59 months and under-five mortality rate.

Socioeconomic inequalities.

These results are summarized in figures 24a-h and in table 14. The table presents two summary measures of inequality. The slope index shows the absolute difference between top and bottom of the wealth scale, based on a regression approach the uses data from all quintiles rather than just the two extreme groups. For example for skilled delivery in the HIDs, the index of 65.6 indicates that this is the difference in percentage points (pp) in the coverage between the richest and poorest children. Table 14 also presents the concentration index that summarizes the overall amount of inequity in the population. Concentration indices take values between minus one and one. A value of zero indicates that the outcome is equitably distributed across all wealth groups. A negative value indicates disproportionate concentration of the health variable among the poor, for example in the case of disease or malnutrition, where the poor are more likely to be affected. A positive value indicates that the poor are getting less than would be expected had the distribution been equitable, as often occurs for preventive and curative interventions.iii

iii More information available at: (http://siteresources.worldbank.org/INTPAH/Resources/Publications/Quantitative- Techniques/ health_eq_tn07.pdf)

IIP-JHU | Retrospective evaluation of ACSD in Ghana 61

Figure 24a-h: Socioeconomic inequalities, showing breakdown by wealth quintiles of selected indicators in “high-impact” zones and comparison area, Ghana, 2006-7.

62 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table 14: Summary indices of socioeconomic inequalities for selected indicators in HIDs and comparison area, Ghana, 2006-7. SLOPE INDEX OF INEQUALITY CONCENTRATION INDEX Comparison Comparison INDICATOR HIDs Area HIDs Area Measles coverage -0.6 2.8 -0.001 0.004 Vitamin A (children) 1.0 1.1 0.005 0.008 ITNs (children) -0.5 -0.3 0.000 -0.058 Diarrhea Management* 13.6 13.7 0.077 0.089 ANC (3 visits) 13.5 22.4 0.031 0.049 Skilled delivery 65.6 60.2 0.254 0.230 Stunting -5.1 -11.1 -0.070 -0.129 Underfive mortality -18.8 N/A -0.037 N/A * ORS, recommended home fluids, or increased fluids with continued feeding

There were virtually no inequities for indicators such as measles vaccination, vitamin A and ITNs, which were promoted using campaign and community outreach approaches. In contrast, poor children presented lower coverage levels than their better-off peers for diarrhea management, antenatal and delivery care. The largest gaps refer to skilled attendance at delivery. There was only one significant difference between HIDs and the comparison area in terms of equity in coverage – in the latter, ITN coverage was slightly higher among the poor than among the rich, whereas in the HIDs there was no inequality (p=0.02).

To better understand the equity gap in diarrhea management, we carried out additional analyses of children who received ORS packets from a provider. Inequities were even sharper, with seven percent coverage in the poorest quintile and 42 percent in the better-off. This is in agreement with the finding that interventions requiring contact with a provider – ANC, skilled delivery, etc – tend to be more inequitable than those delivered through community channels.

In terms of stunting and mortality (Figures 24g-h) the slopes are in the opposite direction than for most coverage indicators, that is, levels are higher among the poor than the rich. The summary indices (table 14) take a negative sign under these conditions. The degree of inequality in stunting was lower in the HIDs than in the comparison area, but this was not statistically significant.

For mortality, data on equity are available only for the HIDs, because of the limitations of the data collected in the comparison area (see section 7). For this reason, the quintiles used for the mortality analyses are based on the asset distribution in the HIDs only rather than the joint distribution of assets in HIDs and comparison area. Use of the joint distribution, with very small numbers of HIDs children in the better-off quintiles, would not allow precise estimation of mortality rates for these groups. Our results show that children in the poorest quintile had mortality levels that were substantially greater than those in all other quintiles (Figure 24-h). This pattern is uncommon at high mortality levels such as that observed in the HIDs; when mortality is high, such as observed in Ghana, the better-off quintile usually stands out from the other quintiles with markedly lower mortality levels.

Summing up, the analyses of socioeconomic inequalities show remarkably small gaps between rich and poor for interventions delivered through campaigns and outreach, but there are substantial inequities for those that depend on health services. Appropriate management of diarrhea, which is mostly indicative of family practices, also showed inequities. Inequalities in stunting prevalence are lower in the HIDs, but it is difficult to attribute this finding to ACSD given the lack of differential effect on inequities in coverage indicators.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 63

Gender inequalities.

Table 15 presents data on possible inequities in terms of gender. There is no evidence of a preferential treatment for boys, either in the HIDs or in the comparison area. Gender inequalities were not analyzed for antenatal or delivery care, when the sex of the baby was yet to be known (assuming a low frequency of pregnancy ultrasound). For the impact indicators (table 15), stunting prevalence was similar in both sexes, but mortality was higher for boys than for girls, as is the case in most places in the world, although this difference was not statistically significant.

Table 15: Selected coverage, nutrition and mortality indicators for boys and girls in the HIDs and comparison area, Ghana, 2006-7.

2006 MICS and 2007 SUPPLEMENTAL MICS COVERAGE or NUTRITIONAL MALE FEMALE INDICATOR AREA TOTAL % n % n p Any measles HIDs 98% 98% 187 99% 208 >0.10 Innoculation (12- 23m) Comparison 96% 95% 274 98% 276 0.07

ITN use for HIDs 58% 57% 1119 58% 1137 >0.10 under five children Comparison 24% 24% 1364 23% 1304 >0.10 Vitamin A HIDs 90% 90% 991 90% 984 >0.10 supplementation of children (6- Comparison 96% 96% 1217 96% 1151 >0.10 59m) HIDs 28% 23% 183 34% 174 0.05 ORT for diarrhea Comparison 30% 33% 231 27% 173 >0.10

Moderate & HIDs* 35% 37% 671 32% 665 0.08 severe stunting (24-59m) Comparison 37% 37% 730 35% 654 >0.10

MORTALITY AREA U5MR U5MR Births U5MR Births p Under-five HIDs 86.2 93.8 834 77.6 858 >0.10 mortality

Urban-rural inequalities.

Urban residents accounted for 11 percent of the HID sample and 21 percent of the comparison area. Urban women showed higher coverage of skilled attendance at birth and urban children had lower prevalence of stunting in both the HIDs and comparison area (table 16). In contrast, rural children in the comparison area were significantly more likely to sleep under an ITN. There were no significant urban/rural differentials for under-five mortality rates, nor for coverage with the remaining interventions (measles vaccine, vitamin A, diarrhea managment and ANC). There is no evidence that ACSD implementation affected urban/rural differentials.

In summary, the analyses of inequalities by socioeconomic position, gender and urban/rural residence did not provide evidence that ACSD implementation contributed to improving equity in Ghana.

64 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table 16: Selected coverage, nutrition and mortality indicators for urban and rural areas in the HIDs and comparison area, Ghana, 2006-7.

COVERAGE or 2006 MICS and 2007 SUPPLEMENTAL MICS NUTRITIONAL URBAN RURAL INDICATOR AREA TOTAL % n % n p Any measles HIDs 98% 99% 87 98% 309 >0.10 Innoculation (12- 23m) Comparison* 96% 97% 110 96% 439 >0.10 ITN use for HIDs 58% 53% 418 59% 1838 >0.10 under five children Comparison* 24% 17% 586 25% 2083 0.01 Vitamin A 0.07 supplementation HIDs 90% 93% 368 89% 1607 of children (6- >0.10 59m) Comparison* 96% 97% 511 96% 1857

ORT for HIDs 28% 32% 58 27% 299 >0.10 diarrhea Comparison 30% 36% 86 29% 319 >0.10 Skilled birth <0.001 attendant: HIDs* 40% 71% 94 33% 392 doctor or <0.001 nurse/midwife Comparison* 42% 61% 114 37% 419

3+ visits ANC HIDs 89% 94% 94 88% 385 >0.10 care Comparison* 80% 88% 113 77% 419 0.09 Moderate & HIDs* 35% 27% 228 36% 1108 0.03 severe stunting (24-59m) Comparison* 36% 23% 333 41% 1052 <0.001

MORTALITY AREA U5MR U5MR Births U5MR Births p Under-five mortality HIDs 86.2 83.6 329 86.4 1363 >0.10

IIP-JHU | Retrospective evaluation of ACSD in Ghana 65

66 IIP-JHU | Retrospective evaluation of ACSD in Ghana

9. Conclusions

In this section, we summarize the findings of the evaluation, addressing two separate questions:

a. Was ACSD implementation associated with improvements in coverage, nutrition and mortality over time?

b. If so, was progress in the ACSD districts faster than observed for the rest of the country?

As described in section 2 (methods), there was no true baseline survey in the HIDs and comparison area that met the quality criteria for coverage and nutritional data. Implementation of ACSD started in mid 2002, and therefore the 1998-9 DHS was too early for a baseline, and the 2003 DHS survey a bit too late. It is important to keep this issue in mind when interpreting the evaluation results.

Figure 25 summarizes these trends in the HIDs and comparison area during the period from 1998-9 to 2006-7. The horizontal axis shows the change in coverage in the HIDs and the vertical axis the corresponding changes in the comparison area. All dots are on the right side of the y-axis, that is, the indicators showed an increase in the comparison area (although not all of these increases were statistically significant). Most dots are also above the x-axis, meaning that the indicators increased in the HIDs. A few interventions, mainly related to case management, showed some degree of decline in the HIDs.

Figure 25: Summary of absolute changes between 1998-9 and 2006-7 in coverage and family practices in “high-impact” districts and comparison area, Ghana. Absolute percentage change in coverage Comparison area -40 -20 0 20 40 60 80 80 IPTp 60 Antibiotics for pneumonia ITN (child)

40 BF within 1 hr DPT3 EBF Skilled delivery Vitamin A (child) 20 ANC4+ Measles HIDs TT2 0 Careseeking pneumonia

ORT & Vitamin A -20 feeding (postnatal)

Absolute percentage change in coverage coverage change percentage in Absolute AM for fever -40

Key: ANC+ interventions EPI+ interventions Infant feeding Case management

When the indicator increased (or decreased) to a similar extent in both areas, the points are close to the diagonal. Indicators that are above the diagonal showed better performance in HIDs than in comparison

IIP-JHU | Retrospective evaluation of ACSD in Ghana 67

area. The reverse is true for those below the diagonal. A larger number of indicators improved faster in the HIDs than in the comparison area than the reverse, but key indicators of case-management tended to increase faster in the latter areas, except for antibiotics for pneumonia. A caveat of the results shown in Figure 25 is that they do not reflect baseline levels. For example, vitamin A to the child shows a larger increase in the comparison area than in the HIDS, but it started out from baseline levels of 65 percent in the HIDs and 22 percent in the comparison area, so that the scope for improvement was much greater in the latter.

Relative to the two questions posed at the beginning of the chapter, the answers for coverage indicators are:

(a) Most coverage indicators improved over time in the HIDs and reached the target coverage levels set by ACSD, although declines were observed for case management indicators.

(b) Comparison with the rest of the country showed mixed results, although more indicators showed faster increase in the HIDs than in the comparison area.

Turning to nutritional status, the answers to the two basic questions are:

(c) The HIDs showed a reduction between 1998-9 and 2007 in underweight and stunting prevalence, but not in wasting.

(d) Over the period from 1998-9 to 2006-7, stunting declined significantly faster in HIDs than in the comparison area. Because of the time lag between the implementation of ACSD-promoted nutritional interventions and the detection of an impact on stunting, it is unlikely that ACSD can account for much of the observed reduction in this indicator. Wasting, on the other hand, declined significantly in the comparison area while remaining stable in the HIDs.

Reducing under-five mortality by 25 percent by 2006 was the primary goal of the ACSD strategy. Our analyses showed that:

(c) There was a reduction of 20 percent in under-five mortality in the HIDs, close to the ACSD goal of 25 percent. This trend was ascertained through the full birth history technique, and the reduction was close to reaching statistical significance (p=0.10).

(d) Data on under-five mortality trends in the comparison area were available from a different source than those for the intervention area, with an endpoint in 2004. Other analyses suggest that mortality levels remained stable at around 115 deaths per thousand live births, in contrast to the 20 percent reduction by 2007 in the HIDs. The different endpoints and analytical techniques used in the two time series preclude a more accurate comparison.

Because of the small sample sizes in the HIDs at baseline, analyses of inequalities in coverage and nutrition indicators were limited to comparisons at the end of the study period. Our conclusions are:

(a) Only small socioeconomic inequalities were observed for interventions delivered through campaign approaches such as vaccination, vitamin A and ITNs. Diarrhea management, four or more ANC visits, stunting and mortality showed intermediate magnitudes of inequalities, whereas large rich-poor gaps were observed for skilled delivery care. Inequalities between boys and girls were virtually non-existent. Urban-rural inequalities were small, except for skilled attendance at delivery and for stunting.

(b) When HIDs were compared to the rest of the country, there was no evidence of differences in patterns of health inequalities for intervention and coverage indicators.

The retrospective nature of the evaluation imposed a number of important constraints that may have affected our findings. These include the fact that no true baseline data were available, as discussed above. Secondly, the available “near-baseline” samples were very small in the HIDs, precluding the precise measurement of coverage and nutrition indicators. Third, the methods and timelines for mortality assessment were different in the two areas being compared. Finally, HIDs were markedly poorer than the rest of the country, so that comparability is affected; a more appropriate comparison area would consist of neighboring districts in the Northern and Upper West regions, but in spite of numerous attempts we were

68 IIP-JHU | Retrospective evaluation of ACSD in Ghana

unable to obtain such data in a timely fashion.

The joint interpretation of findings on coverage, nutrition and mortality in the HIDs is limited by the different time spans for the coverage and nutrition indicators (1998-9 to 2006-7) and for mortality estimation (1998-2001 to 2004-2007). The main causes of under-five deaths in Ghana in 2003 were malaria (33%), neonatal causes (29%), pneumonia (15%) and diarrhea (12%). The highest coverage levels in the endline surveys in the HIDs were achieved for vaccinations, vitamin A supplementation to children, antenatal interventions (including IPTp and TT) and ITNs. Exclusive breastfeeding also showed large increases. One would expect these interventions to have a greater impact on deaths due to malaria, neonatal causes and diarrhea. None of these preventive interventions, except for HiB vaccine and exclusive breastfeeding, would be expected to affect pneumonia deaths. Reported careseeking for pneumonia stagnated while antibiotics for pneumonia significantly increased. Further analyses showed that over one-fourth of the reported antibiotics came from drug shops and itinerant vendors, making it difficult to interpret the impact of this practice. Key informants reported that mothers might have chosen to go to drug vendors when community-based workers encouraged mothers to seek care for cases of childhood pneumonia. Case-management interventions against malaria and diarrhea showed low and declining coverage levels in the HIDs. Taken together interventions showing large gains in coverage had only limited impact on the main causes of death, and hence are compatible with a moderate decline in mortality levels, similar to the 20 percent reduction observed in the HIDs.

When contrasting trends in the ACSD and comparison area, it is important to consider that a large number of international, bilateral and Ghanaian agencies have been operating in both areas, before as well as during the study period. The sections on background characteristics and implementation (sections 3 and 4) show that many of the interventions promoted by ACSD had been actively implemented by other agencies, some well before ACSD was formally launched in 2002, and others in collaboration with ACSD. These included, but were not limited to; the Navrongo Health Research Center (vitamin A, ITNs and CHPS strategy), Ghana Red Cross Society (mother-to-mother support groups and community activities), World Food Program feeding programs, the LINKAGES project (infant feeding interventions), World Vision and CRS (nutritional rehabilitation centers and education), CHPS centers (access to primary health care), etc. The coverage of the CHPS strategy, posting community nurses to improve preventative and curative primary health care, expanded greatly in the HIDs over the ACSD implementation period. ACSD worked with many of these partners to achieve further increases in coverage.

Building upon what exists is a key ACSD strategy, and although this makes strong programmatic sense, it renders it difficult if not impossible to attribute specific coverage gains to ACSD per se. Thus, the results must be interpreted in light of combined efforts to improve child survival in the region. The Ghana implementation team noted on various occasions several key ACSD contributions, including: 1) the program’s ability to concentrate on a package of effective interventions; 2) additional resources for commodities, equipment and human resources; 3) clearly stated targets; 4) establishment of productive partnerships and synergies across institutions; and 5) achievement of strong commitment from the GoG, GHS and other donors. The Ghana team also noted key lessons learned from the ACSD experience and recommendations for future child health programming, shown in box 4.

At the same time, other development partners, including UNICEF, provided massive investments in the rest of Ghana, again making it difficult to ascertain the additional impact of ACSD by comparing the HIDs with other geographical regions.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 69

Box 4: The way forward: Lessons learned in ACSD Ghana according to national counterparts

1. The ACSD strategy did not focus strongly on interventions to improve child undernutrition  New ACSD-like programs need a “nutrition-plus” component to ensure that efforts and resources are devoted specifically to nutritional interventions.

2. Community-based activities, including training and supporting CHWs, were an integral, but challenging component of ACSD  More attention needs to be given to the motivation of community-based agents and their supervisors at the sub-district level  Support for adequate supervision, monitoring and incentives and an uninterrupted supply of commodities will be essential to sustain adequate levels of motivation and quality

3. Gains for behaviors related to management of childhood illness, skilled assistance at delivery, and nutritional practices were less than expected  Changing behaviors is complicated and time consuming. More efforts and resources should be devoted to behavior-change strategies, especially face-to-face counseling and mother’s support groups

4. Supervision and monitoring system are often weak and untimely, particularly at the sub-district and community levels. Problems, such as stock-outs of ORS and antimalarials, were picked up by the current system, but only after persisting for long periods.  Importance needs to be given to supervision and M & E systems, developing systems that function in real time.

5. Increased supply of commodities was a contribution of ACSD; however, stock-outs of essential commodities associated with weak supply management, gaps in funding and changes to national policies hindered potential gains in intervention coverage.  Ensuring an adequate and continuous supply of essential commodities will strengthen future program efforts;  Commodity security should be included in program planning and monitoring; alternative approaches should be explored to strengthen commodity security.

6. ACSD was integrated into the planning processes at the regional level; it now needs to be better integrated into the national and district-level planning processes

7. Government ownership of the program was an on-going issue; the program is still often viewed as an externally driven project outside of the routine health services.

8. External evaluation results can be used to improve new ACSD-like programs  ACSD successes should provide an impetus for scale-up of similar packages and new interventions;  Introduction of interventions should be done incrementally with early review;  Evaluation results should convey a sense of urgency of all that remains to be done, especially for nutrition and case management of childhood illness

The findings reported above should not detract from the fact that remarkably high and equitable levels of coverage with key child survival interventions were achieved in the HIDs, and that ACSD coverage goals were met for a majority of these indicators, in a region of extreme poverty when compared to the rest of the country. Stunting prevalence declined markedly over time, although much of the improvement seems to have occurred prior to ACSD implementation, likely associated with strong nutrition programs present in the HIDs for a number of years. Under-five mortality also showed a 20 percent reduction, which just failed to reach statistical significance. The fact that gains in intervention coverage were greater in the HIDs than in the comparison area lends plausibility to the hypothesis that some part of the mortality reduction found in the HIDs may be attributed to activities linked to ACSD.

70 IIP-JHU | Retrospective evaluation of ACSD in Ghana

References

1. UNICEF. Accelerating child survival and development: A results-based approach in high under-5 mortality areas. Final Report to CIDA. New York, 2005.

2. Bryce, J., Gilroy, K., Black, R.E., Jones, G. & Victora, C.G. A Retrospective Evaluation of the Accelerated Child Survival and Development Project in West Africa; Inception Report. Baltimore, MD, Johns Hopkins University Institute for International Programs, 2007.

3. Bryce, J., Victora, C.G., Habicht, J.P., Black, R.E. & Scherpbier, R.W. Programmatic pathways to child survival: results of a multi-country evaluation of Integrated Management of Childhood Illness. Health Policy Plan. 20 Suppl 1: i5-i17 (2005).

4. UNICEF. Accelerating early child survival and development in high under-five mortality areas in the context of health reform and poverty reduction: a results-based approach. UNICEF funding proposal (2002-2003). New York, 2002.

5. Darmstadt, G.L., Bhutta, Z.A., Cousens, S., Adam, T., Walker, N. & de Bernis, L. Evidence- based, cost-effective interventions: how many newborn babies can we save? Lancet. 365 (9463): 977-88 (2005).

6. Jones, G., Steketee, R.W., Black, R.E., Bhutta, Z.A. & Morris, S.S. How many child deaths can we prevent this year? Lancet. 362 (9377): 65-71 (2003).

7. Bryce, J., Terreri, N., Victora, C.G., Mason, E., Daelmans, B., Bhutta, Z.A. et al. Countdown to 2015: tracking intervention coverage for child survival. Lancet. 368 (9541): 1067-76 (2006).

8. UNICEF. UNICEF/WHO Meeting on Child Survival Survey-based Indicators. Summary List of Child Survival Indicators. New York, 2004. Available at www.childinfo.org.

9. Bryce, J., Boschi-Pinto, C., Shibuya, K. & Black, R.E. WHO estimates of the causes of death in children. Lancet. 365 (9465): 1147-52 (2005).

10. Efem, I., Jehu-Appiah, C., Anemana, S.D., Addai, E., Awittor, E. & Ankrah, V. Report of the review of the Accelerated Child Survival and Development Programme in the Upper East region of Ghana, November 2004. Accra, Ghana, UNICEF, Ghana, 2004.

11. Victora, C.G., Schellenberg, J.A., Huicho, L., Amaral, J., El Arifeen, S., Pariyo, G. et al. Context matters: interpreting impact findings in child survival evaluations. Health Policy Plan. 20 Suppl 1: i18-i31 (2005).

12. W H O Multicentre Growth Reference Study Group & de Onis, M. WHO Child Growth Standards based on length/height, weight and age. Acta Paediatr. 95 (s450): 76-85 (2006).

13. Habicht, J.P., Victora, C.G. & Vaughan, J.P. Evaluation designs for adequacy, plausibility and probability of public health programme performance and impact. Int J Epidemiol. 28 (1): 10-8 (1999).

14. WHO. World Health Statistics. Geneva, 2007.

15. UNICEF. State of the World's Children 2008. New York, UNICEF, 2007.

16. GSS & UNICEF. Multiple Indicator Cluster Survey, Ghana 2006. Accra, Ghana, Ghana Statistical Service and UNICEF-Ghana, 2006.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 71

17. Berry, L. Ghana : a country study, 3rd ed ed. Washington, D.C. , Federal Research Division, Library of Congress 1994, pp. 382.

18. U.S. Census Bureau. International Data Base - Country Summary: Ghana, 2008.

19. Ghana Statistical Service (GSS) and Macro International Inc (MI). Ghana Demographic and Health Survey 2003. Calverton, Maryland, GSS and MI, 2003.

20. Bryce, J., Requejo, J. & 2008_Countdown_Working_Group. Tracking progress in maternal, newborn, and child survival: the 2008 report. (avilable at http://www.countdown2015mnch.org). New York, UNICEF, 2008.

21. Coulombe, H. Ghana Census-based Poverty Map: District and Sub-District level Results, Ghana Statistical Service and Department for International Development, UK, 2005.

22. FAO. Gateway to Land and Water Resources, Ghana obtained from: http://www.fao.org/ag/agl/swlwpnr/reports/y_sf/z_gh/gh.htm#overview, accessed 10 August 2008., 2004.

23. Appawu, M., Owusu-Agyei, S., Dadzie, S., Asoala, V., Anto, F., Koram, K. et al. Malaria transmission dynamics at a site in northern Ghana proposed for testing malaria vaccines. Trop Med Int Health. 9 (1): 164-70 (2004).

24. Kleinschmidt, I., Omumbo, J., Briet, O., van de Giesen, N., Sogoba, N., Mensah, N.K. et al. An empirical malaria distribution map for West Africa. Trop Med Int Health. 6 (10): 779-86 (2001).

25. Gemperli, A., Sogoba, N., Fondjo, E., Mabaso, M., Bagayoko, M., Briet, O.J. et al. Mapping malaria transmission in West and Central Africa. Trop Med Int Health. 11 (7): 1032-46 (2006).

26. Landgraf, B., Kollaritsch, H., Wiedermann, G. & Wernsdorfer, W.H. Plasmodium falciparum: susceptibility in vitro and in vivo to chloroquine and sulfadoxine-pyrimethamine in Ghanaian schoolchildren. Trans R Soc Trop Med Hyg. 88 (4): 440-2 (1994).

27. Ehrhardt, S., Mockenhaupt, F.P., Agana-Nsiire, P., Mathieu, A., Anemana, S.D., Stark, K. et al. Efficacy of chloroquine in the treatment of uncomplicated, Plasmodium falciparum malaria in northern Ghana. Ann Trop Med Parasitol. 96 (3): 239-47 (2002).

28. Koram, K.A., Abuaku, B., Duah, N. & Quashie, N. Comparative efficacy of antimalarial drugs including ACTs in the treatment of uncomplicated malaria among children under 5 years in Ghana. Acta Trop. 95 (3): 194-203 (2005).

29. Afari, E.A., Akanmori, B.D., Nakano, T. & Ofori-Adjei, D. Plasmodium falciparum: sensitivity to chloroquine in vivo in three ecological zones in Ghana. Trans R Soc Trop Med Hyg. 86 (3): 231-2 (1992).

30. Vitamin A supplementation in northern Ghana: effects on clinic attendances, hospital admissions, and child mortality. Ghana VAST Study Team. Lancet. 342 (8862): 7-12 (1993).

31. Otten, M., Kezaala, R., Fall, A., Masresha, B., Martin, R., Cairns, L. et al. Public-health impact of accelerated measles control in the WHO African Region 2000-03. Lancet. 366 (9488): 832-9 (2005).

32. Bossert, T.J. & Beauvais, J.C. Decentralization of health systems in Ghana, Zambia, Uganda and the Philippines: a comparative analysis of decision space. Health Policy Plan. 17 (1): 14-31 (2002).

72 IIP-JHU | Retrospective evaluation of ACSD in Ghana

33. Ghana_Health_Service. Regional Detailed Health Statistics, obtained from: http://www.ghanahealthservice.org, accessed 7 August 2008., 2008.

34. Binka, F.N., Bawah, A.A., Phillips, J.F., Hodgson, A., Adjuik, M. & MacLeod, B. Rapid achievement of the child survival millennium development goal: evidence from the Navrongo experiment in Northern Ghana. Trop Med Int Health. 12 (5): 578-83 (2007).

35. Binka, F.N., Kubaje, A., Adjuik, M., Williams, L.A., Lengeler, C., Maude, G.H. et al. Impact of permethrin impregnated bednets on child mortality in Kassena-Nankana district, Ghana: a randomized controlled trial. Trop Med Int Health. 1 (2): 147-54 (1996).

36. Phillips, J.F., Bawah, A.A. & Binka, F.N. Accelerating reproductive and child health programme impact with community-based services: the Navrongo experiment in Ghana. Bull World Health Organ. 84 (12): 949-55 (2006).

37. Nyonator, F.K., Awoonor-Williams, J.K., Phillips, J.F., Jones, T.C. & Miller, R.A. The Ghana community-based health planning and services initiative for scaling up service delivery innovation. Health Policy Plan. 20 (1): 25-34 (2005).

38. Belch, R. Working paper: background conditions in the Upper East region, Northern Ghana, 2005. A background document for the Land Conservation and Smallholder Rehabilitation Project (LACOSREP) - Phase II Evaluation. Wa, Ghana, International Fund for Agricultural Development, 2006.

39. Browne, E.N., Maude, G.H. & Binka, F.N. The impact of insecticide-treated bednets on malaria and anaemia in pregnancy in Kassena-Nankana district, Ghana: a randomized controlled trial. Trop Med Int Health. 6 (9): 667-76 (2001).

40. Binka, F.N., Hodgson, A., Adjuik, M. & Smith, T. Mortality in a seven-and-a-half-year follow-up of a trial of insecticide-treated mosquito nets in Ghana. Trans R Soc Trop Med Hyg. 96 (6): 597-9 (2002).

41. UNICEF. Stratégies accélérées de survie et du développement du jeune enfant (SASDE) au Mali: Un recherche opérationnelle mise en œuvre dans le cadre du PRODESS dans six cercle de démonstration. Bamako, Mali, 2005.

42. Pence, B.W., Nyarko, P., Phillips, J.F. & Debpuur, C. The effect of community nurses and health volunteers on child mortality: the Navrongo Community Health and Family Planning Project. Scand J Public Health. 35 (6): 599-608 (2007).

43. Shrimpton, R., Victora, C.G., Onis, M., Lima, R.C., Blossner, M. & Clugston, G. World wide timing of growth faltering: implications for nutritional interventions. Pediatrics. 107 (5): E75 (2001).

44. GSS, NMIMR & ORC_Macro. Ghana Demographic and Health Survey 2003. Calverton, Maryland, Ghana Statistical Service (GSS), Noguchi Memorial Institute for Medical Research (NMIMR), and ORC Macro, 2004.

45. AED. LINKAGES Project Ghana: Final Report (1997-2004). Washington, DC, Academy for Educational Development 2004.

46. UNICEF. Accelerating early child survival and development in high under-five mortality areas in the context of health reform and poverty reduction: a results-based approach. UNICEF proposal to Canadian CIDA. New York, 2002.

IIP-JHU | Retrospective evaluation of ACSD in Ghana 73 ACCELERATING CHILD SURVIVAL AND DEVELOPMENT (ACSD)

Final Report The Retrospective Evaluation of ACSD: Ghana

APPENDICES

Submitted to UNICEF Headquarters on 7 October 2008

Institute for International Programs Johns Hopkins Bloomberg School of Public Health Baltimore, MD

A. Description of Ghana and “high-impact” districts B. Methodology for documentation of implementation activities and contextual factors C. Documentation of implementation D. Definition of key indicators E. Survey Questions F. Methodologies of surveys in Ghana 1998-2007 G. Tables presenting priority coverage indicators over time for ACSD high-impact districts H. Tables presenting comparisons of priority coverage indicators over time in ACSD high-impact districts and the comparison area I. Tables presenting 2007 MICS results for key coverage indicators in the ACSD high-impact districts by socio-demographic characteristics of the population J. Additional tables for nutrition K. Methodological challenges L. References for the appendices M. Mapping of partners’ activities in “High-impact” districts (Upper East region) and nationally

APPENDIX A Description of Ghana and “high-impact” districts Geography Ghana, located in West Africa, maintains three international boarders and a coast off the Gulf of Guinea. Togo is situated to the East, Cote d’Ivoire to the West, and Burkina Faso to the North and Northwest. Formerly known as the Gold Coast, Ghana achieved independence from Great Britain in 1957. Ghana’s 23,383,000(2) people are distributed over 238,500 km, with their capital in Accra. Divided into 10 political regions, 17% of the population resides in the three northern regions (Upper West, Upper East, and Northern Regions), which along with Brong Ahafo region comprise Ghana’s Savannah ecologic zone.(3) Greater Accra and part of Central Region encompass the Coastal zone, while Ashanti, Volta, Western and Eastern Regions are predominantly in the Forest zone.(4)

Fig A1: Ecological map of Ghana Fig A2: Map of Ghana showing the regions

Population Of Ghana’s 23,383,000 people, 38% are younger than 15 years old.(2) In 2000, 47% of households were in urban areas, but given the higher average number of people per household in rural areas, an estimate 41% of the population is urban.(5) However, some regions, like the Upper East Region, host as much as 87% of the population in rural areas.(3) The overall male-female ratio is 100.2:100, but distribution is unequal with more women living in rural areas than men. The estimated growth rate is currently 1.9% with a total fertility rate of 3.8 births per woman.(5)

A2 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Diversity typifies Ghana’s population with more than 50 languages and dialects spoken throughout the country. According to the 2000 Census, Akans comprise 45.3% of the population, Mole-Dagbon 15.2%, Ewe 11.7%, Ga-Dangme 7.3%, and less than 5% each of Guan, Gurma, Mande-Busanga and others. In terms of religion, approximately 69% are Christian, 15-30% Muslim, and the other faiths include traditional African religions and Judaism.

Economy Well endowed with natural resources, Ghana’s per capita output is twice that of neighboring West African countries. In 2005 and 2006, the Gross Domestic Product grew at a rate of 6%. The domestic economy is based on subsistence agriculture, while gold, timber and cocoa earn most foreign exchange. Even though 60% of the labor force is involved in agriculture, it only contributes 34% of the GDP. The services industry, employing 25% of the labor force, supplies 41% of the GDP.(6, 7)

Despite prosperity relative to its neighbors, Ghana maintains a 5.7 billion (US$) debt, 26% of the Gross National Income. In 2001, the unemployment rate rested at 20% with no recent updates. According to a 2003 Poverty Profile, roughly 40% of Ghanaians live below the poverty line (900,000 cedis), and 27% are in extreme poverty (less than 700,000 cedis). However, rural areas suffer the brunt of poverty with an average of 55% below the poverty line.(8) The northern savannah regions are most affected with 70-88% of households in poverty.(3)

Table A1: Percentage of Population below Poverty Lines(8) National Average Rural Average Urban Average Poverty (>900,000 c) 39.5% 54.6% 23% Absolute poverty 26.6% 40.7% 14.2% (>700,000)

Education

Fifty-eight percent of Ghanaians 15 years or older are literate. Literacy among youth (15-24 years) is higher, but the gender disparity continues with rates of 76% for males and 66% for females. The primary school net enrololment ratio, the number of school-aged children enrolled divided by the number of school-aged children in the population, is 69-70% for boys and girls. The gross enrollment ratio, which includes children outside the age-appropriate limits, is 93-94%, suggesting that children older than official primary school-aged enroll as well.(9) Similar to the poverty trends, the 3 northern regions have the lowest adult literacy rate at 24%.(3)

Health

Primary health program expansion since 1978 has reduced childhood mortality rates. Additional sub- district health facilities and trained personnel, along with the Expanded Programme on Immunization (EPI) initiated in 1976, have contributed to health gains that now provide Ghanaians with an average life expectancy of 59 years.(9) Additionally, health reforms in early 1990s focused on an achieved reduction in early childhood mortality. However, progress plateaued by 2003. Table 3 provides a 1998-2003 comparison of various health indicators.

Of particular concern, the infant mortality rate (IMR) rose from 57 deaths of infants less than 12 months per 1,000 live births in 1998 to 63 deaths per 1,000 as measured in the 2003 DHS. The Upper West and Northern Regions were estimated to have the highest infant and child mortality rates in the 10 years previous to the 2003 DHS. Rising vaccination coverage and health care seeking behavior prompted the Ghanaian Ministry of Health to investigate explanations for the mortality stagnation. The Upper East Region was identified as an anomaly to the trend, and their child survival interventions are currently being explored.

IIP-JHU | Retrospective evaluation of ACSD in Ghana A3

UPPER EAST REGION

Given the disparities observed among health, education and economic indicators, the Upper East Region (UER) has become one of the focal points for collaborative programs and intervention projects.

Geography Located in the north-eastern corner of Ghana, UER shares international borders with Burkina Faso and Togo, along with internal borders with Upper West Region and Northern Region. Cross-border movement is common and gives way to difficulties in disease surveillance and control. UER is divided into 8 districts, each autonomous regarding planning, budgeting and implementation of projects. Bolgatana is the Region’s capital. Overall, there are 911 communities both dispersed and sometimes overlapping. Of the Region’s 1017 km of road, 69% is considered motorable. Unlike the Ghana’s southern regions, UER has 2 seasons instead of 4, with particular drought hazards between January and March.

Fig A3: District Map of UER

Population According to the 2000 census, 4.8% of Ghanaians live in UER, which represents only 3.7% of the country’s landmass. Despite a higher than average population density, 87% of the population is rural. The 2006 estimated population was almost 983,000, up from 920,000 in 2000. The growth rate is 1.1%, below the national average. 56.3% of males and 49% of females are 0-19 years old. The proportions reflect an excess of adult females compared to national averages. Out-migration of men is the predominant explanation. (10) Ethnically, 74.5% are Mole-Dagbon, 8.5% Grusi, 6.2% Mande-Busanga, and 3.2% Gurma. The Region’s main languages are Gurune, Kusal, Kasem, Buili and Bisa. 46% of the population practices traditional religions, 28% Christianity, and 23% Islam.

Economy UER is has the highest percent of the population living below the poverty among Ghana’s 10 Regions. 88% earn less than 900,000 cedis annually. According to the Ghana Living Standards Survey, poverty worsened between 1992 and 1999 in UER. Over 80% of the economically active population engages in agriculture, predominantly grains and cattle. Only one industry, a cotton ginnery, is active.

Education UER also has the highest level of illiteracy in the country. Seventy-eight percent of adults 15 years or older are not literate in either English or a Ghanaian language. The disparity between male and female literacy is most exaggerated in UER as well. The Region supports 449 primary schools, 177 junior secondary schools, and 23 senior secondary schools, but 71.8% over the population 6 years and older have never attended school. More males than females have attended school, 35.3% versus 23.6%

A4 IIP-JHU | Retrospective evaluation of ACSD in Ghana

respectively of the 6 years and older population; however the proportion varies by district. 34% of children 7-14 years work full-time, over half of which are boys. (7)

Table A2: Educational Attainment for Those Who Ever Attended School Education Level Percent of >6 years old population who have Attained attended school and attained given educational level Primary 48.1% Middle/JSS 20.8% Secondary/SSS 12.5% Vocational/Technical 4.2% Post-secondary 4.5%

Health UER’s health situation presents an interesting mix of improvements and declines. In 1998, DHS information indicated that UER had the highest early childhood mortality rates in Ghana, as shown in Table 3. However, by the 2003 DHS, UER reduced both infant and under-5 mortality rates were estimated to have dramatically decreased, despite the lack of progress in neighboring regions. However, many other health indicators remained above national averages, such as the percent of malnourished children. UER mothers actually received less antenatal care (ANC) in 2003, while the percentage of professionally assisted births rose. However, the majority of UER deliveries were still unattended professionally, at slightly over half the national average. In comparison with Upper West Region and Northern Region, which have basically similar populations in terms of culture, socio-economic conditions, health determinants and human resource difficulties, the Upper East Region far exceeded mortality rate expectations. Fig A4: Average annual rainfall in Ghana Sentinel site data indicated that 44% of UER’s mortality burden was attributed to deaths of children younger than 5 years old. The primary under-5 mortality contributors were: malaria, anaemia, diarrhoea, malnutrition, acute respiratory infections, measles and neonatal complications. Child survival interventions, such as the Integrated Management of Childhood Illnesses, may be related to UER’s deviant results. Several organizations have supported projects in the Region; such as the Dioscesan Health Service (1998-2006), Ghana Red Cross Society (1999-2006), World Vision International (1996-2007), Community Water and Sanitation Agency (1973-2005), US Agency for International Development (1998-2007), Japanese International Cooperation Agency (2003-2007), World Health Organization (2003- 2007), Opportunities Industrialization Centres International (2003-2006), and Danish International Development Agency (2003- 2007).(10)

Source: FAO Gateway to Land and Water Resources, Ghana(1)

IIP-JHU | Retrospective evaluation of ACSD in Ghana A5

APPENDIX B Methodology for implementation of ACSD activities and contextual factors

Various techniques were employed to collect information retrospectively about the implementation of ACSD activities and contextual factors in the “high-impact” zones. Much information was gathered from colleagues at the UNICEF-Ghana field office, who have been collaborating on the retrospective evaluation throughout the process. Field visits, key informant interviews and working meetings to review of the preliminary coverage results all provided information pertaining to details of ACSD implementation and contextual factors. Details of these discussions are provided in table B1. During these encounters, the JHU evaluation team requested any documents providing more details on ACSD and other partner’s activities.

A6 IIP-JHU | Retrospective evaluation of ACSD in Ghana

TEAM VALUATION s and E PARTICIPANTS Jennifer Bryce, Jennifer Bryce, Robert Black, Gareth Jones and Kate Gilroy Kate Gilroy and Elizabeth Hazel Gareth Jones and Elizabeth Hazel Gareth Jones , Gareth Jones Kate Gilroy Gareth Jones, Kate Gilroy Gareth Jones, Kate Gilroy

Full documentation available available documentation Full ACSD-Ghana ii) challenges for ACSD-Ghana ii) challenges ring data processing; performed ring data processing; OPICS COVERED T

Full documentation available upon request available upon Full documentation

Full documentation available upon request available upon Full documentation sessions carried out to document ACSD implementation activitie ACSD implementation out to document carried sessions Discussion of ACSD and mortality rates in Ghana & UER UER Ghana & in rates ACSD and mortality of Discussion

Areas of focus: (i) impressions of (i) impressions of focus: Areas from those and (iii) communications the evaluation; to was of the visit The objective the program. implementing for and identify promising avenues process begin the learning analysis and further documentation training du ring interviewer team assistance Provided technical Provided on-site MICS survey. supplemental for the 2007 UNICEF and to interviewer team and feedback of supervision GSS. du Provided logistical assistance provided feedback to data quality; of review a comprehensive GSS. UNICEF and upon request upon request and processing MICS 2006 – implementation, of Discussion sources data progress; other available

ARTICIPANTS P Ghana MOH Ghana MOH UNCEF, MOH, Danida/DFID, Danida/DFID, UNCEF, MOH, district and regional representatives and UNICEF, GSS, Macro MICS interviewer teams and Harvard UNICEF, GSS Health of Public School Ghana Statistical Service Ghana Statistical GHS – Public Health Div. GHS – Public Health Div. ACSD of Discussion

ESCRIPTION D To collect To collect about information Ghana’s experience and with ACSD, on the available data outcomes process, impact and potential of ACSD in Ghana. Training and survey of supervision teams Review of data processing Meeting Meeting Meeting Description of field visits, key informant interviews, and work and informant interviews, of field visits, key Description

ATE D Table B1. Field Visits Nov 2006 Aug & Sept 2007 Jan 2008 Key informant interviews / discussions 16 – 17 Nov, 2006 contextual factors contextual factors

IIP-JHU | Retrospective evaluation of ACSD in Ghana A7

TEAM VALUATION E PARTICIPANTS Jennifer Bryce, Jennifer Bryce, Robert Black, Gareth Jones, Kate Gilroy Jennifer Bryce, Robert Black, Gareth Jones, Kate Gilroy Jennifer Bryce, Robert Black, Gareth Jones, Kate Gilroy Jennifer Bryce, Robert Black, Gareth Jones, Kate Gilroy Jennifer Bryce, Robert Black, Gareth Jones, Kate Gilroy Jennifer Bryce, Jennifer Bryce, Robert Black, Gareth Jones, Kate Gilroy Jennifer Bryce, Gareth Jones, Kate Gilroy Kate Gilroy

OPICS COVERED T of Ghana & other West Africa countries countries West Africa other & of Ghana Presentation of the evaluation overview; general discussion of general discussion overview; of the evaluation Presentation ACSD Samata up with C ACSD in Bongo district; Follow of Discussion concerning documents for & A Anyinato Azaba, G Alcolba, of ACSD and annual reports implementation and health roles in outreach CHOs and CBAs of Description child birth at health of promotion of description promotion; facility; review of kit boxes of CHO in community; and work CHPS centre of Description centre of CHPS and records; review review of CHPS registers supplies out at present carried and past research projects of Description of exercise of mapping of DSS; discussion Navrongo; description UER Discussion of implementation of ACSD, IMCI; discussion of IMCI; discussion of ACSD, implementation of Discussion etc re: CCM, IMCI, national policies including nutritional ACSD in Bawku West district; of Discussion and ITNS centers

ARTICIPANTS P UNICEF GHS – Bongo District Offices GHS – Bongo district, Zorkor facility GHS – Bongo district, facility Kodoro CHPS Navrongo health research center GHS – Reproductive & child GHS – Reproductive & child health unit district West – Bawku GHS offices

ESCRIPTION D Informal discussion Informal discussion Service Ghana Statistical Meeting / informal Poverty maps discussion Mtg Site visit to health centre Site visit to CHPS center Site visit to Navrongo Meeting / interview Meeting / interview Mtg with district

ATE D 16 – 17 Nov, 2006 20-24 Nov. 2006

A8 IIP-JHU | Retrospective evaluation of ACSD in Ghana

TEAM VALUATION E PARTICIPANTS Jennifer Bryce, Jennifer Bryce, Gareth Jones, Kate Gilroy Jennifer Bryce, Kate Gilroy Gareth Jones Jennifer Bryce, Gareth Jones, Kate Gilroy Jennifer Bryce, Gareth Jones, Kate Gilroy Gareth Jones Jennifer Bryce, Gareth Jones, Kate Gilroy Jennifer Bryce, Jennifer Bryce, Kate Gilroy

s of training & monitoring of s of training OPICS COVERED T Description of CHOs and CBAs roles in outreach and health and health roles in outreach CHOs and CBAs of Description review of services; of antenatal health promotion; Description notebooks monitoring center register; review of CBAs of and collection reports annual & available data of Discussions and strategies communications of discussions data/reports; discussion materials; collection of of pertinent documents CBAs; collection of and availability mapping activity of Discussion protocol/questionnaire/data of ACSD Discussion and progression; challenges, implementation etc from findings/perceptions of preliminary Informal presentation & challenges successes of Discussion site visit; and possibility of further MICS 2006 survey of Discussion Region (also on 24-11-2006) sampling in UER, UWR & Northern from findings/perceptions of preliminary Informal presentation & challenges successes of Discussion site visit; Identification of documents with UNICEF inputs & timeline for inputs & with UNICEF of documents Identification ACSD

ARTICIPANTS P GHS – Bawku West – GHS – Bawku center Sapelliga health East Regional GHS – Upper office Navrongo health research center UNICEF UNICEF

ESCRIPTION D Site visit to health Site visit to health center collection F-U mtgs; of keydocs Follow-up mtg UNICEF Discussion Debriefing of site visit Informal discussion UNICEF Debriefing of site visit Informal discussion Informal discussion UNICEF

ATE D 20-24 Nov. 2006

IIP-JHU | Retrospective evaluation of ACSD in Ghana A9

TEAM VALUATION E PARTICIPANTS Gareth Kate Gilroy Gareth, Kate Kate TBD

toring of CBAs; collection toring OPICS COVERED T

Discussion of mapping exercise project for UER project for mapping exercise of Discussion General timeline of implementation; collection of further collection of implementation; General timeline documentation and data sources of further data sources available; discussion and Collection files identification of mico-data Discussion of training and moni documentation pertinent KNUST preliminary and mortality nutrition of coverage, Full review results

ARTICIPANTS P GHS, MOH, KNUST< KNUST< GHS, MOH, GSS UNICEF, IN-DEPTH

ESCRIPTION D Informal discussion Informal discussion UNICEF Informal discussion UNICEF Mtg Informal discussion Informal discussion KNUST Presentation of results preliminary

ATE D 20-24 Nov. 2006 results preliminary Work sessions to review and interpret July 2008

A10 IIP-JHU | Retrospective evaluation of ACSD in Ghana

APPENDIX C Documentation of ACSD implementation activities

UNICEF-ACSD, consolidating the efforts of previous programs, presented a package of cost effective intervention and a strategy of service delivery for scale-up region-wide(11). UNICEF acted as the facilitator with partner support and the Ghana Health Service (GHS) implemented the interventions at the regional, district and sub-district levels(11). The UER has a strong history of community-based health volunteers. UNICEF-ACSD utilized this resource for intervention delivery: developing a community health curriculum and recruiting and training Community Based Agents (CBAs)(11). De-worming, PMTCT and IPTp programs were introduced as part of ACSD(11).

The ACSD implementation activities are described in more detail here, expanding on the information provided in the main body of the report. Timelines of implementation activities for each ACSD component are presented in tabular format with brief explanatory text. In order to estimate the magnitude of implementation, the population projection for adults, children and infants in 2004 was used to standardize and provide a coverage estimate.

Delivery of ITNs In late 2002, distribution of ACSD ITNs began in the Upper East region (table C1). The start of implementation was variable by district, some districts adopted ITN delivery before the overall ACSD program(11). Bed nets were distributed to the district offices, then to the volunteers and then to the communities(11). Multiple strategies of community delivery were used: 1. ITN sale of nets to target groups at health centers 2. Volunteer sales agents accompany nurses on health outreach session to sale nets while the nurses work 3. Community based agents (CBA) trained in management of childhood illnesses began distributing and retreating ITNs(11). 4. Retreatment and distribution campaigns All volunteers, CBAs and nurses involved in the ITN program have been trained on ITN distribution and retreatment (11).

Treated bed nets were sold at a reduced price to families with children under five and pregnant women through a chit (voucher) system(11). However as demand for nets increased, the subsidized nets were sold using the 20/80 rule(11). Eighty percent of the nets are sold to the target groups for 5000 cedis and the remaining 20% are sold to anyone for 23,000 cedis(11). At the time of purchase, the customers are advised by the volunteers to retreat every six months, through the health centre or a volunteer(11). Retreatment cost is 2000 cedis per net and the ITN volunteer agents receive a 1000 cedis per net sold or retreated(11).

From November 2002 to September 2004, volunteers in UER sold 156,510 out of 236,500 (66%) ITN nets received (Table C1). In the UER, a reported 36,223 ITN nets were distributed for an estimated 38,450 (94%) pregnant women and 109,579 ITN nets distributed for an estimated 144,187 (76%) under five children (Table C1). Distribution occurred at ANC, PNC, CWS and delivery service points (12). At the end of 2004, 100% coverage is reported for children under five and pregnant women(13).

The first household retreatment campaign occurred in May of 2003 (11) and the second retreatment campaign was planned for April of 2004(12). The retreatment campaign in 2004 was delayed until June, leading to concerns that the campaign occurred too late in the rainy season(12). Retreatment was also integrated into Child Health Week in May of 2004(11, 12). The campaigns that occurred during 2003 and the 2004 Child Health week were free of charge(11). After Child Health week in 2004, the nets were retreated for 2000 cedis(11). During this second retreatment exercise, there was poor turnout as cost was a barrier for many people(14). Also there were limited chemicals for retreatment(11). The cumulative outputs from these campaigns are listed in Table C1. By mid-2004, depending on the source 12,000 – 13,000 nets were retreated out of an estimated 169,965 nets in the community, approximately 10 percent (table C1).

IIP-JHU | Retrospective evaluation of ACSD in Ghana A11

ITN distribution and retreatment activities continued into 2005. By December 2005, most of the districts had stock-outs of ITN nets, except for Bongo and Bawku Municipal districts funded through the Global Fund and Roll back Malaria(15). Therefore only modest gains in cumulative ITN distribution for pregnant women and under five children are reported since 2004 (table C1). Ghana submitted a proposal to UNICEF for addition bed nets in 2005 (15). In order to estimate the number of bed nets required, Ghana completed a bed net inventory that was integrated into the filariasis treatment campaigns(15).

Retreatment of bed nets in 2005 was planned for the spring but then delayed until Child Health Promotion Week(15). There are also several accounts of first and second quarter funds delayed until later in the year(16, 17) although ITN distribution was reportedly ongoing (15). At the end of the year approximately 25,000 were retreated out of an estimated 244,000 nets in the community (Table C1). Retreatment levels are still relatively low (25%) however the Upper East region received approximately 40,000 KO tablets to help with retreatment activities(17).

The proposal for UNICEF was for one billion, one millions cedis for additional nets(18). At the first quarter of 2006, the region had 40,000 KO tablets but no nets to distribute(18). Instead the region procured nets from the Global Fund to distribute while awaiting the UNICEF nets(18). The Global Fund nets were significantly more expensive at 20,000 cedis compared to the UNICEF nets which were sold for 5,000 cedis(19). Retreatment campaigns began during Child Health Promotion Week and continued for several months in order to address the issue of low retreatment levels (18). At this point, CBA-IMCI volunteers are the primary mechanism for retreatment campaigns(19). The mid-2006 totals for bed net distribution and retreatment are quite low compared to earlier years (Table C1). However as Ghana moves towards exclusive use of long-lasting ITN nets, the retreatment campaigns will become less important(11).

A12 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Notes

Received: 236,500 ITN;Total sales: Cumulative nets retreated since June Different totals source reported by Kick-off date

▪ 156,510 ITN Cumulative ITN distribution listed under second retreatment, 2004 May-June

▪ 2003 ▪ ) ) ) estimated 169,965 12 14 20 under-five children targeted under-five children targeted 1 1 pregnant targeted women pregnant targeted women Intensity Intensity 1 1 ITN retreatment integrated into Child Health Week activities( 12 ) 12,012 re-treated out of nets in the community( 109,579 / 153,799 144,187 / 153,799 36,223 / 36,223 13,766 re-treated ( 13,353 retreated ( 38,450 / 36,223 Area Regional Regional

12 ) ) 13 ) Regional 11, 12 11, ) 20 ) ) Activity Activity 11, 20 12 ) 11, 20 Nov: began distribution of ITNs( first retreatment of May-July: ITN( Health May: Week( June: second retreatment of ITN ( Cumulative ITN Distribution from 2002 to Sept. 2004 ( (Dec. 2004) ( Annual ITN Distribution( Months Jan-Mar Apr-Jun Jul-Sept Oct-Dec Jan-Mar Apr-Jun Jul-Sept Oct-Dec Jan-Mar Apr-Jun Jul-Sept Oct-Dec Annual Total/ Total/ Annual Unspecified Total/ Annual Unspecified Total/ Annual Unspecified Overview of ITN distribution and retreatment in the Upper East region, 2002-2006 2002-2006 East region, retreatment in the Upper and of ITN distribution Overview 2002 2003 2004 Year Table C1: Table C1:

IIP-JHU | Retrospective evaluation of ACSD in Ghana A13

Notes

Used 2005 estimated for number of Received: 236,500 ITN;Total sales: Different totals source reported by

▪ nets in community ▪ 184,069 ITN

▪ 807,447 adultsand children>5y807,447 ) 16 under-five children targeted 1 pregnant and women 1 pregnant targeted women under-five children targeted Intensity Intensity 1 1 pregnant targetedwomen 1 17 ) nt women & 38,450 children 0-11m ; 0-11m children 38,450 & nt women 6,829 retreated out of an estimated 244,000 nets in the community 132,270 / 153,799 684 per 190,022 5,489 / 153,799 2,541 / 36,223 children under five 40,576 / 36,223 25034 retreated out of an estimated 244,000 nets in the community( 25,812 retreated in early May before the rainy season ( Area Regional Regional Regional 40,000 KO tablets have been delivered in 2005

) Regional 16 ) Regional 13, 13, ) 13 ) 13, 18 Activity Activity ) : 153,799 under five children, 36,223 pregna children, five under : 153,799 22 ) 15 ) ChildMay: Health Week ( Cumulative ITN Distribution from 2002 to Dec. 2005 ( Mid year report: ITN retreatment( Mid year report: bed net distribution and retreatment( ITN sales agent trained( 21 Months Jan-Mar Apr-Jun Jul-Sept Oct-Dec Jan-Mar Apr-Jun Jul-Sept Oct-Dec Annual Total/ Total/ Annual Unspecified Annual Total/ Unspecified 2005 2006 Year 1 – Estimated population from 2004 projections( 2004 from Estimated population 1 –

A14 IIP-JHU | Retrospective evaluation of ACSD in Ghana

EPI+ The strengthening of the preexisting EPI+ program involved many partners and it is difficult to decipher the exact UNICEF contribution(23). UNICEF-ACSD mostly focused on developing strategies to improve defaulter tracing (11). Using community based surveillance systems, the CBA volunteers used a register to trace children due for vaccinations(11). Mop-up campaigns occurred after National Immunization Days (NID) to vaccinate “zero dose” children identified by polio vaccinators during the NID(11). UNICEF also worked through GHS, District Assemblies and the BESFA and BUCO rural banks to allocate funds for the purchase of iodized salt to women in microcredit groups in the Builsa and Bawku East districts(12).

In 2003, monitoring data shows 45% BCG, 38% measles and 35% Penta3 coverage for an estimated 23,207 children under five targeted (Table C2). Concerns during implementation of EPI+ in 2003 include: poor quality of district-level EPI data, continuing high levels of wastage and defaulting, incomplete or late EPI form submissions and irregular submission of Cold Chain inventory reports(23).

Vitamin A campaigns were initiated during the 2004 Child Health Promotion Week(24). Monitoring data from 2004 shows an estimated 44% of children 6-11m and 27% of children 12-59m received vitamin A(25). In 2004, two campaigns for de-worming children under five and pregnant women occurred during the NID in March and September, resulting in very high coverage for the target populations (12). Four rounds of NID occurred in the Upper East region in February, March, October and November (26). The monitoring data from the first quarter of 2004 shows large gains compared with 2003: 36% BCG, 30% Measles and 31% for Penta3 (Table C2). The NID campaigns are successful; polio coverage is above 100% and no wild-type polio has been detected since September 2003(14). However the issue is still adherence to immunization schedule: initial contact with EPI+ is high, but continuation is poor, effecting the quality of vaccination (14). Intermittent shortages of measles vaccine were reported in BW district(12, 26)

In addition to campaigns, the Upper East regional office also provided cold chain equipment, logistical assistance and monthly or quarterly feedback at the district level (23). UNICEF contributed to these efforts along with other partners. The issues identified at the end of the year by the UNICEF for the EPI+ programs are late release of funds, inadequate incentives for volunteers, inadequate information of immunization schedule and inadequate monitoring and supervision(14).

In 2005, the NID continued, four rounds in February, April, November and December(27). UNICEF procured 26.2 millions doses of OPV to assist the GHS in polio eradication efforts(17). The EPI+ program received adequate quantities of routine vaccines: MOH and GAVI purchased all vaccine using UNICEF procurement services(17). All districts in the Upper East region have a community-based registration system in place: 992+ CBS volunteers have been trained in all 8 districts and deployed with community registers(17). Four districts in the UER have functional defaulter tracing and outreach services(17). The CBS volunteers received defaulter tracing refresher training and conducted a market immunization and defaulter tracing exercise, results on Table C2 (16) (27).

In addition to the defaulter tracing training and exercises, Bawku West district reported strong EPI+ activity for 2005: routine monthly static and outreach immunizations, quarterly mop-up immunizations and training of Health Staff on increasing immunization coverage at sub-district levels (27). A GHS National EPI report shows high coverage for the UER at the end of 2005 (Table C2). Continuing measles shortages is reported for the BW district(19). No Vitamin A coverage data for this year although continued distribution. De-worming campaigns for children under five continued with very high coverage rates, almost 100% (Table C2).

In 2006, an EPI survey completed in four districts, shows defaulter tracing and mop-up campaigns need to be strengthened(18). Defaulter tracing exercises are to be done monthly, but some districts are not in compliance(19). A GHS National EPI report shows high coverage for the UER at the end of 2006 (Table C2). There was no quarterly mop-up of iodization campaign although market supplementation occurred (Table C2).

IIP-JHU | Retrospective evaluation of ACSD in Ghana A15

28 ) Notes USAID 0.5 million USD grant ( First quarter EPI+ results for 2004

▪ ▪ much higher than 2003 annual results

targeted 1 targeted targeted targeted 1 1 1 targeted targeted 1 1 targeted targeted 1 1 targeted de-wormed de-wormed 1 1 1 targeted 1 Intensity Intensity

DPT: 8177 / 38,450 Cold chain expansion Increase demand for EPI services 170,736 / 153,799 177,533 / 153,799 DPT: 711 / 38,450 BCG: 10462 / 38,450 Measles: 8735 / 38,450 BCG: 8357 / 38,450 Measles: 7037 / 38,450 Measles: 33,927/ 38,450 Penta3: 33,395/ 38,450

BCG: 41,528/ 38,450

54,803 / 153,799 Area Regional (district level available) Regional (district level available) Regional (district level available) Regional (district level available) Regional (district level available) Antigen

BCG, Measles, DPT3 Vitamin A NID: Antihelminth NID: Antihelminth BCG, Measles, DPT3 National

25 ) ) ) 14 ) 12, 14 12, 14 12, 20 ) Activity Activity 20 ) Cumulative immunization for 2003( National Immunization (NID)( Days Jan: ACSD EPI+ activities begin ( 11 ) March: First round of de- worming( EPI+ first quarter totals( ChildMay: Health Promotion Week( Oct: second round of de- ( worming Months Timeline of implementation of EPI+ activities in the Upper East region, 2002 - 2006 Upper East region, 2002 Timeline of implementationthe of EPI+ activities in Annual Total/ Unspecified Annual Total /Unspecified Jan-Mar Apr-Jun Jul-Sept Oct-Dec Jan-Mar Apr-Jun Jul-Sept Oct-Dec Jan-Mar Apr-Jun Jul-Sept Oct-Dec 2002 2003 2004 Year Table C2:

A16 IIP-JHU | Retrospective evaluation of ACSD in Ghana

r Notes

Since Oct 2004 Market immunization and defaulte No mop-up campaign held during

GHS EPI report GHS EPI report tracing ▪ ▪ ▪ this quarter

targeted 1 targeted 1 targeted 1 targeted 1 targeted 1 targeted 1 807,447 adultsand children>5y807,447 Intensity Intensity

Penta3: 1074 / 38,450 UER: BCG: 111% Penta3: 96% Measles: 90% UER: BCG: 111% Penta3: 93% Measles: 96% 177,553 / 153,799 Yellow fever: 1509 /38,450 Measles: 1503 / 38,450 BCG: 1140 / 38,450 OPV3: 1052 / 38,450 Supplemented 100 bags of salt in Bolga market

nt women & 38,450 children 0-11m ; 0-11m children 38,450 & nt women Area National Regional (district level available) Regional (district level available) Antigen

National National

Antihelminth fever Yellow BCG Penta3 OPV3 Measles

) found in district report Only 27 15 ) 15 ) ) Iodized salt salt 13 ) Iodized ) : 153,799 under five children, 36,223 pregna children, five under : 153,799 22 ) 24 ) Activity Activity Annual total de-worming for under fives( EPI+ annual totals( Annual regional summary of EPI activities( 24 ) Months Annual regional summary of EPI activities( Jan-Mar Apr-Jun Jul-Sept Child Health Week( Oct-Dec Jan-Mar salt( Apr-Jun Jul-Sept Iodized Oct-Dec Annual Total /Unspecified

2005 2006 Year 1 – Estimated population from 2004 projections( 2004 from Estimated population 1 –

IIP-JHU | Retrospective evaluation of ACSD in Ghana A17

IMCI

Before ACSD, IMCI interventions were implemented on a smaller scale by partners such as Ghana Red Cross and Catholic Relief Services (11). ACSD pulled together the experiences of these pilot programs and presented a common framework for scale-up region-wide(11). Community IMCI was implemented in 2003 however the majority of community based agents (CBA) did not begin service until 2004(11). The C-IMCI model utilizes trained CBAs on a voluntary basis to provide the following services: appropriate infant feeding practices(12), health education to mothers, fever treatment with pre-packed chloroquine, diarrheal treatment with ORS, recognition and referral of ARI, promote immunization and iodized salt and mobilize the community for participation in de-worming, NIDs and other programs(11). UNICEF collaborated with KNUST, Ghana Red Cross and Ghana Health Services to develop a CBA training program. CBA volunteers were equipped with bicycles, educational materials and health kits containing Kinaquine junior and infant (chloroquine), ORS sachets and hand washing material(11). In order to continue motivation and commitment, the CBA volunteers earn a percentage of sales(11). For instance, a CBA earns 100 cedis on every ORS sachet sold(11). Based on past experience in the regions with volunteers, female volunteers are preferred over men(11).

Monitoring and supervision was carried out by the KNUST team and the regional office(12). However the most of the time it is integrated into routine supervision with Regional Health Management Team (RHMT) members(12). However there are issues with integration, sub-district supervisors are reluctant to carry out CBA supervision without additional funds for fuel, for instance(29).

In the first half of 2003, IMCI scale-up activities and CBA volunteer recruitment took place (Table C3). CBA volunteers were trained at the district and regional level and also at KNUST Community Health Department of the School of Medical Science (11). The first CBA training session occurred from May- June 2003, in which 1039 or 1118 (depending on the source) volunteers were trained (Table C3). Training of Trainers (TOT) sessions are reported in July 2003 at the sub-district level (Table C3). Additional volunteer training sessions occurred in July and November and December of 2003 in all districts although there is no record on the number of volunteer trained (20). The number of children seen and referred for illness for 2003 is reported in Table C3.

In March of 2004, an additional 744 CBAs were trained bring the total to 1780 or 1892, depending on the source (Table C3). The first two quarters of 2004 show poor supervision of CBAs at the district and sub- district level (12). In June 2004, KNUST conducted CBA supervisory visits, covering one-third of the sub- districts in the region in one month (30). KNUST visited the CBAs and the households that had accessed services from the CBAs. Some of the results were promising: there was an itinerary and plans for continued supervisory visits, CBAs are undertaking follow-up visits for illness treatment and CIMCI implementation has strong community support(30). However supervision is weak, CBA coverage is inadequate and distribution of inputs is incomplete and behind schedule(30). For childhood illnesses, the KNUST team found that mothers are accessing CBAs too late (more than 24 hours after onset of symptoms) and many adults are consuming the drugs meant for children(30). Also it was found that CBAs focused more on drug treatment than the health education messages(30). Finally KNUST noted that many areas in the region will be inaccessible to CBAs during the rainy season.

KNUST participated in CBA training in October of 2004, training approximately 100 CBAs in the Bolga and BE districts(31, 32). Also in October, UNICEF held a TOT session for 30 extension field staff from GHS, Department of Community Development and Environmental Health of Sanitation Unit (33). The TOT session focused on ACSD activities so that the leaders could return their communities and train representatives to disseminate the information(33). Later in November, the 30 participants in this TOT exercise trained a total of 300 representatives on the ACSD objectives. The purpose of these workshops was to strengthen the capacity of community members on ACSD activities and promote the use CBAs (33) In November, UNICEF conducted an ACSD sensitization workshop for 36 political authorities including council chairpersons, opinion leaders, assembly-persons, market queens and political representatives(33). Another component of CIMCI is training of clinicians. UNICEF reported 5 clinicians were trained in the CIMCI module although it is difficult finding trainers in the UER(14).

A18 IIP-JHU | Retrospective evaluation of ACSD in Ghana

In October, KNUST undertook another CBA supervisory visit this time focusing on CBAs to discuss the implementation process; this was the fourth such supervision exercise since May of 2004 (32). KNUST found that many of the CBAs had not received refresher training since their initial training, 11 months ago and weak supervision continues to be an issue. However many of the supervisors had been introduced to various M&E tools. The quality of training differed by district, Bolga had the most well organized training and Bawku the poorest(32). Also at the district level, there is a lack of the computer literacy for proper data processing, analysis and interpretation. KNUST found CBAs continuing to emphasis drug treatment over health education and had difficultly filling out paperwork.

The number of children seen and referred for illness for 2004 is reported in Table C3. In 2004, health education sessions were only taking place in three districts: Bawku West, Bolga and Bongo. The number of adults and adolescent children receiving health education from CBAs is shown on Table C3.

As of 2004, five district hospitals have been designated as Baby Friendly as part of the Baby Friendly Health Initiative Facilities (BFHI). UNICEF- ACSD supports BFHI by developing guidelines on exclusive breastfeeding, training Mother to Mother Support Groups (MtMSG) and providing training materials. A district report from Bawku West offers more details on BFHI training: meeting with midwives and sub district leaders on BFHI, 20 health staff trained on lactation management and complementary feeding and 100 TBAs trained on exclusive breast-feeding(26).

In 2005, UNICEF continued clinical staff training exercises, 48 prescribers and 3 regional staff were trained in the UER, 20 RHMT and NGO partners sensitized to CIMCI, 20 district level TOT sessions and sub-district TOT and CBA training(16, 17). At there end of the year, UNICEF reports a total of 1780 CBAs providing CIMCI servicing 922 communities(17). Other reports state 1810 CBAs trained(16) and 1892 total trained(29).

Throughout 2005, KNUST and the regional staff supervised the CBAs. Sub-district supervision was found to continue to weak and there is an inadequate supply of logistics such as kits, training materials and bicycles(15). There was also an issue with expired Kinaquine(16). There are been some CBA drop- out: either the CBAs left, found other jobs or the women got married(16). Replacement of CBA staff affecting CBA supervision and new staff claim they are not trained in CIMCI(19). The number of children seen and referred for illness for 2005 is reported in Table C3. In 2005, health education sessions were only taking place in three districts: Bawku West, Bolga and Bongo. The number of adults and adolescent children receiving health education from CBAs is shown on Table C3.

BFHI activities continued; UNICEF focused on exclusive breastfeeding training and 16 health facilities were assessed and 15 qualified as Baby Friendly(16).

In mid 2006, TOT sessions are continued with 10 clinicians and 3 regional focal persons participating(13). In October, 24 prescribers were trained(34). The monitoring team evaluated trained prescribers and found high non-compliance with the ACSD objectives (13) Monitoring and supervision of CBAs are ongoing but not to expectation(13). In 2006, a total of 1982 CBAs are reported (19). The regional team conducted supervision visits to 1366 of the 1982 CBAs in the region. The team found poor supervision at the district and sub-district level for instance many supervisors were not inquiring about ACSD activities during their supervisory visits. In some districts the supervisors did not know where their volunteers were or even how many CBAs were in their jurisdiction. The CBAs complained of missing or irregularly paid commissions and lack of mobility due to broken bicycles(19). The regional team also found poor integration of ACSD activities into routine services(19).

IIP-JHU | Retrospective evaluation of ACSD in Ghana A19

Notes

Reported program kick-off date Different number reported than elsewhere Different number reported than elsewhere “Action plan scale up& budget meeting IMCI No data on number of volunteers trained No data on number of volunteers trained ▪ ▪ ▪ ▪ KNUSTwith Home Management of Malaria: Feb 2003” ▪ ▪

1

1

1 1 Intensity Intensity

1 Under-five children 1039 trained per 153,799 Under-five children Malaria: 50,760 cases per 153,799 Under-five children ; 970 referred Diarrhoea: 21,444 cases per 153,799 Under-five children ; 268 referred Under-five children ;all referred 1118 trained per 153,799 ARI: 549 cases per 153,799 Source UNICEF Annual review IMCI Monitoring report Area Regional Regional Regional Regional

) Regional ) ) 29 ) 20 ) ) 12, 14 12, 20 Activity Activity 20 Community entryand volunteer recruitment ( 20 ) 11 TOTfor sub-districts July: begin( Community volunteer training( May-June: volunteerMay-June: training ( volunteerJuly: training( Nov/Dec: volunteer training ( Results of community health volunteers: treatment and referrals ( 13 ) Timeline of implementation of IMCI+ activities in the Upper East region, 2002 - 2006 Upper East region, 2002 Timeline of implementationthe of IMCI+ activities in Months

Jan-Mar Planning meeting( 20 Jan-Mar Planning Apr-Jun Jul-Sept Oct-Dec Annual Total / Unspecified

Year Year 2003 Table C3:

A20 IIP-JHU | Retrospective evaluation of ACSD in Ghana

13 ) Notes ) and in a 2006 brief resume( 25

From GHS, Dept. of Community Development From GHS, Dept. Community representatives trained were on Poor supervision from district to sub-district Cumulative CBAs reported trained 1118 also reported trained in 2003 different Slightly (smaller) numbers in annual Number of adults and adolescent children

▪ & Environmental Health of Sanitation Unit ▪ ACSD objectives by 30 TOTmembers trained in Aug 2004

▪ quarters level for the first two ▪ ▪ ▪ report( ▪ received health education from community health

1

1 1

1 Intensity Intensity population over 5y

1 1 RI: 549 cases per 153,799 30 extension field staff 300 participants 36 leaders participated 1036 (2003) + 744 (2004) = 1780 cumulative trained per 153,799 5 clinicians trained per 153,799 Under-five children Under-five children Malaria: 14,003 cases per 153,799 Diarrhoea: 21,444 cases per 153,799 Under-five children A Diarrhoea/sanitation: 8895 per 807,447 Under-five children Under-five children Source Training briefing report Training briefing report Training briefing report

Area Regional, held in Tamale Regional, held in various locations Regional KNUST report Regional KNUST Regional Regional Regional Regional (district available) Bolga, BE KNUST report ~100 CBAs in various sub-districts Regional KNUST report Bawku West, ) ) ) 31, 32 31, ) ( 29 33 33 20 ) 33 ) ) Oct: KNUST Activity Activity Oct: TOT for TOT Oct: Nov: Community 12 st th th Nov: Sensitization nd general ACSD activities( representative workshops( 25 -31 26-29 Oct: : KNUST CBA supervisory visits( 32 ) 2 exercise for political authorities( CBA volunteer training ( 14 ) 13 25 IMCI Clinical Training ( 14 ) Results of community health volunteers: treatment and referrals ( 12 ) CBA training( Quarterly supervisionQuarterly report( June: KNUST CBA supervisory visits( 30 ) receivedJuly: Kinaquine prepacks( 25-26 Results of community health volunteers: health Months

Jan-Mar Apr-Jun Jul-Sept Oct-Dec Annual Total/ Unspecified

Year Year 2004

IIP-JHU | Retrospective evaluation of ACSD in Ghana A21

Notes Number of different health volunteers, more to Lists different number of CBAs = 1892

▪ be trained. ▪

1

1 1

1 1 1

1

1

pregnant 1 1 population 1 pregnant 1 Intensity Intensity

1 RI: 1643 per 807,447 Under-five children CBA-surveillance: 1522 per 153,799 staff regional 3 children; Under-five trained women CBA-agents: 1892 per 153,799 Under-five children ITN volunteers: 684 per 153,799 Malaria: 20,189 cases per 153,799 Under-five children ; 1,556 referred Diarrhoea: 11,839 cases per 153,799 Under-five children ; 968 referred Under-five children ; 944 referred women TBA: 495 per 36,223 Under-five children Midwives: 173 per 36,223 48 prescribers trained per 153,799 ARI: 944 cases per 153,799 population over 5y over 5y 5 health institutions qualify for BFHI; 3 district hospitals and 18 health centres to follow. Infant feeding: 4665 per 807,447 Immunization: 6688 per 807,447 Malaria: 8158 per 807,447 A population over 5y population over 5y Source Volunteer database Area

Regional Regional (district available) Regional Regional

) 14, 16, 17 Activity Activity Baby Friendly Health Health Friendly Baby Initiative (BFHI) facilities Database for community based volunteers and midwives ( 35 ) Oct: Training for prescribers ( Results of community health volunteers: treatment and referrals ( 15 ) Months

Jan-Mar Apr-Jun Jul-Sept Oct-Dec

Year Year 2005

A22 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Notes )

34 training date, in addition to 48 trained st Strongest level of health education in Bawku No education reported in other districts Nov 21 642/1366 no bicycleswith 816/1366 had no kits 460/1366 had no reporting formats

▪ West ▪ ▪ in Oct 2005 (

▪ ▪ ▪ 1

1 1 1 1

1 1

1 population 1 Intensity Intensity 450 children 0-11m; 807,447 adults and children >5 and adults 807,447 0-11m; children 450 population over 5y 1 over 5y over 5y Under-five children population over 5y Diarrhoea/sanitation: 24,401 per 807,447 Infant feeding: 24,140 per 807,447 Immunization: 23,447 per 807,447 Malaria: 21,319 per 807,447 population over 5y ARI: 7,068 per 807,447 Malaria: 10,377 cases per 153,799 Under-five children ; 513 referred Diarrhoea: 9.897 cases per 153,799 Under-five children ; 172 referred Under-five children ; all referred Under-five children; 3 regional focal focal regional 3 children; Under-five persons population over 5y 10 clinicians per 153,799 24 prescribers trained per 153,799 supervision1366 met by team out 1982 CBAs trained ARI: 166 cases per 153,799 Source

Meeting report & IMCI supervisory report Area

Bawku West, Bolga, Bongo Not specified Not specified Regional district (by available) Regional 19 ) ) : 153,799 under five children, 36,223 pregnant women & 38, & women 36,223 pregnant children, five under : 153,799 22 ) report( ) ) 15 ) 13 13 Activity Activity Results of community health volunteers: health education( Midyear report:Midyear IMCI TOT management case training( Mid year report: results of health community volunteers: treatment and referrals( Mid year report: IMCI+ ( 13 ) Months

Jan-Mar Apr-Jun Jul-Sept Oct-Dec Supervision

Year Year 2006 1 – Estimated population from 2004 projections( 2004 from Estimated population 1 –

IIP-JHU | Retrospective evaluation of ACSD in Ghana A23

ANC+

ANC services, known as the IPT package, were offered through the Antenatal clinics to pregnant women. The IPT package includes Vitamin A supplements, Iron and folic acid supplements, antihelminths and IPT of malaria with sulphadoxine pyremethamine. Two districts were supported by the Global Fund starting May 2004: Bongo and Bawku East(11). In June 2004, ACSD extended IPT to the remaining four districts(11). Tetanus Toxoid immunization of pregnant women is also included, but as part of the EPI+ program and many of the immunizations were done during NIDs(23). UNICEF supported TT immunization along with other partners(23).

In May 2004, a National TOT for IPT activities took place in all six UER districts(20) . District level IPT training occurred in June 2004(11). Also this year sensitization activities took place in ANC clinics by services provider(20). TT immunization was scaled up to all districts, the first round of TT supplementary immunization activities (SIA) took place in early 2004 with 46% coverage of the target population (Table C4). The second round of TT SIA occurred later in the year with an estimated 62% coverage(23). The third round of TT SIA only covered two districts with 73-95% estimated coverage (Table C4). Distribution of postnatal vitamin A and SP to pregnant women began mid-2004, with service through ANC delivery centres(12). Annual postnatal vitamin, SP and de-worming coverage for 2004 can be seen in Table C4. UNICEF-ACSD completed a market immunization and defaulter tracing exercise for EPI+ antigens, TT was included in the exercise (Table C4). Bawku West district reported iron deficiency and anaemia control training for midwives and other health workers (Table C4).

In 2005, de-worming activities were scaled up in Bawku West. Annual de-worming rates for the region are reported in Table C4. BW reported three rounds of TT SIA in 2005 and even though we found no other evidence of TT SIA in other UER districts, it is assumed that it was ongoing because TT coverage is estimated at 76% for the UER in 2005(24). Data in postpartum vitamin A distribution is variable during 2005, but up to 2217 women were dosed (Table C4). SP distribution continued with no reported side effects (Table C4) although there are high drop-out rates after first and second dose(15). Use of SP by pregnant women was promoted through radio health education discussions(16). TBAs and CBAs were trained on distribution of postpartum Vit A (16)

The 2006 annual monitoring data for vitamin A, SP, de-worming and TT are reported in Table C4.

A24 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Notes No data on number of TOT No data on scale of July-Dec 2004 2004 July-Dec

▪ participants ▪ sensitization activities MW, CHN doing ANC, CHO ?? ▪

pregnant targeted women pregnant targeted women 1 1 Intensity Intensity dose: 7,320 / 36,223 dose: 5,973 / 36,223 st nd

1 2

Area

BW, Bol, Buil, KN

Regional (district available)

23 ) Regional 105,879 / 230,700 targeted women ) ( 25 ) 20 ) 20 ) ( 25 ) 11, 20 12 12 Activity Activity PMTCT sensitization activities ( 7 – 12 June: District IPT training( De-worming of pregnant women began( Post natal Vitamin A supplementation ( 12 ) May: ACSD fundedMay: IPT begins TOTwith training ( Postnatal July: vitamin A supplementation begins( Months Jan-Mar Apr-Jun Jul-Sept Oct-Dec Annual Total/ Unspecified Jan-Mar Apr-Jun Jul-Sept Oct-Dec Annual Total/ Unspecified Jan-Mar TT SIA round 1( Apr-Jun Jul-Sept Oct-Dec Timeline of implementation of ANC+ activities in the Upper East region, 2002 – 2006 – 2006 Upper East region, 2002 Timeline of implementationthe of ANC+ activities in Year 2002 2003 2004 Table C4:

IIP-JHU | Retrospective evaluation of ACSD in Ghana A25

) 14 June 2005 th 16 ) 13 ) Notes -24 th 12 ) ( 26 May – Dec 2004 May

▪ BW: 20 Regional: No dates or numbers specified ( 2217 received PP Vit A but number of dose not specified. ( 1649 reported in 2006 Brief Resume ( No deworming in Bawku West( Training on iron deficiency and anaemia control in pregnancy.

1 pregnant targeted women pregnant targeted women pregnant targeted women pregnant targeted women Intensity Intensity 1 1 1 1 pregnant targeted women 1 pregnant women targeted pregnant women targeted 1 1 pregnant targeted women 1 pregnant women targeted pregnant women 1 dose: 645/ 36,223 : 7717 / 36,223 : 4843/ 36,223 dose: 645/ 36,223 : 10700 / 36,223 st nd st nd rd Twenty nine midwives and CHNs trained out of 5190 1 3 Bol: 54,307 / 57,372 targeted women Buil: 13,816 / 18,899 targeted women 312 /6708 pregnant targeted women pregnant targeted women 1 IPT1: 18,197 / 36,223 IPT2: 11,115 / 36,223 IPT3: 5,945 / 36,223 2 2 7320 / 36,223 Area BW BW BW Regional, district available

Bolga and Builsa only Bolga, Builsa Bawku, West Bawku (other districts are nil) Regional (district available) Regional (district available) ) 12 23 ) Regional 143,954 / 230,700 targeted women 23 ) 15, 26 ) 12 ) Activity Activity 15 ) Iron Deficiency Anaemia control Training ( Jan: deworming of pregnant women began in BW( 27 ) Three rounds ofTT SIA coverage Annual total for IPT since May 2004( 16 ) IPT for pregnant women ( TT SIA round 2 ( TT SIA round 3 ( Deworming of pregnant women( Results of market immunization, TT for pregnant women Annual total postpartum Vitamin A( Months Annual Total/ Unspecified Jan-Mar Apr-Jun Jul-Sept Oct-Dec Annual Total/ Unspecified Year 2005

A26 IIP-JHU | Retrospective evaluation of ACSD in Ghana

13 ) Notes

5,937 reported in 2006 Brief resume report (

807,447 adults and children >5y pregnant targeted women Intensity Intensity 1 pregnant targeted women pregnant targeted women 1 1 24 ) pregnant targeted women pregnant targeted women pregnant women targeted pregnant women 1 1 1 nt women & 38,450 children 0-11m; 0-11m; children 38,450 & nt women UER: TT2 – 78% ( IPT1: 10,698 / 36,223 6,019 / 36,223 IPT2: 7,793 / 36,223 IPT3: 5,555 / 36,223 6711 / 36,223 Area

Regional 5,051 / 36,223 GHS EPI report

Regional Regional Regional, district available ) 15 ) : 153,799 under five children, 36,223 pregna children, five under : 153,799 22 ) 24 ) 13 ) 13 ) 13 ) Activity Activity Mid year report:Mid year deworming of pregnant women totals ( Annual regional EPI of summary activities( Annual total Deworming for pregnant women( Mid year report: IPT totals ( report:Mid year Postpartum vitamin A totals ( Months Jan-Mar Apr-Jun Jul-Sept Oct-Dec Annual Total/ Unspecified 1 – Estimated population from 2004 projections( 2004 from Estimated population 1 – Year 2006

IIP-JHU | Retrospective evaluation of ACSD in Ghana A27

Table C5: Summary of inputs of supplies and logistics for the ACSD program, UER Date Supplies Number Notes Year 1 Motor bikes (20) 12 42 motor bikes planned/requested Year 1 Pickup truck (20) 1 June 2003 Kinaquin prepacks (20) January 2004 Kit boxes(20) 300 January 2004 Bicycles (20) 216 May 2004 Bicycles (20) 300 Bicycles (16) 800 Reported in 2005 No date given FIMCI training Reported in 2005. Missing the manuals: (16) 60 “Treat the child” and Introduction 60 “Management of sick infant 1 Assess and class sick 60 wk to 2 m” manuals child 60 Identify treatment 60 Assessment chart Counsel to mother Bicycles (19) 814 Reported in 2006 Kit boxes (19) 1,400 Reported in 2006 Motor bikes (19) 6 Reported in 2006 Reporting booklets 2,022 Reported in 2006 (19) Breast feeding 1,440 Reported in 2006 posters (19) Kinaquine junior (19) 645,900 Reported in 2006 Kinaquine infant (19) 100,200 Reported in 2006 ORS (19) 645,900 Reported in 2006 Facility training 420 Reported in 2006 manuals (19) Mosquito nets from 287,850 Reported in 2006 UNICEF (19) Global Fund provided 80,000 nets Mosquito nets from 40,000 Reported in 2006 revolving fund (19) KO tablets (19) 287,850 Reported in 2006 Scales mother/child 20 Reported in 2006 (19)

A28 IIP-JHU | Retrospective evaluation of ACSD in Ghana

PROTOCOL FOR MISSING/UNKNOWN DATA EXCLUDE CASES: Missing mother's report and no entry on vaccination card IMPUTE TIMING 1: Missing card or vaccination on card: use mother's report & imputedistribution timing with of known vaccination dates IMPUTE TIMING 2: Missing/invalid date on card: impute timing with distribution of known vaccination dates EXCLUDE CASES: Unknown/missing mother's report and no card data

¹ ² rst year isrst year assumed to be the same as proportion DOMINATOR DOMINATOR All children 12-23m, MB include still alive, still 6-59m, All children alive, include MB All children 12-23m, MB include still alive,

All children 12-23m, MB include still alive, rs for the evaluation of ACSD

3 NUMERATOR

3 Eligible children received received Eligible children measles inoculation before 12 months of age; according to immunization card, mother's report or receipt of vaccination during national campaign received Eligible children DPT3 before 12 months of age; according to immunization card or mother's report Hib3 received Eligible children before 12 months of age; according to immunization card receiving Eligible children vitamin A in previous 6m according to mother's report or immunization card

Child Child Child Child DATAFILE DATAFILE ore 12m of age, the proportion of vaccinations in the fi given e and family practice indicato

INDICATORS INDICATORS Percentage of children aged 12-23 months received who measles vaccine before first birthday Percentage of children aged 12-23 months received who 3 doses of DPT vaccine before first birthday. Percentage of children aged 12-23 months received full HIB vaccination doses) (3x before first birthday. Percentage of children 6 - received59m at least who one high dose vitamin A supplement the last 6 within months

EPI+ ACSD TARGET TARGET D Measles immunization coverage § DPT3 immunization coverage § Hib3 immunization coverage § Coverage of 6 vitamin A in last months § Definition of priority indicators and protocols for missing data and protocols indicators Definition of priority 1 2 3 4 Definition of priority coverag NO. PPENDIX To estimate the children without a card to have rec'd vaccine bef Indicator Coverage Consensus § International birth children ¹ MB=Multiple birth: include all multiple ² CDC 2003 - valid data due issues, included children with for indicator variables to data quality were all 3 of children with an of children with immunization card rec'd thewho vaccine before 12m of age (MICS manual)

A Table D1:

IIP-JHU | Retrospective evaluation of ACSD in Ghana A29

PROTOCOL FOR MISSING/UNKNOWN DATA

EXCLUDE CASES: Reported treatment of child but missing for specific medications used EXCLUDE CASES: Reported treatment of child but missing for specific medications used EXCLUDE CASES: Reported treatment of child but missing for specific location of treatment EXCLUDE CASES: Reported treatment of child but missing for specific medications used

¹ ² cough, labored cough, labored cough AND cough AND DOMINATOR DOMINATOR

Children (0-59) with reported fever in previous two weeks, include MB, exclude deceased Children (0-59) with reported fever in previous two weeks, include MB, exclude deceased Children (0-59) with: DHS: labored breathing MICS: breathing and chest congestion weeks,in previous two include MB, exclude deceased Children (0-59) with: DHS: labored breathing MICS: breathing and chest congestion weeks,in previous two include MB, exclude deceased NUMERATOR Eligible children received received Eligible children appropriate antimalarial medication according to in previous two national policy weeks any received Eligible children antimalarial medication during weeks illness in previous two seen at were Eligible children appropriate health care facility: excluding pharmacy and other drug vendors given Eligible children antibiotics

Child Child Child Child DATAFILE DATAFILE INDICATORS INDICATORS

Percentage of children aged 0-59 months with fever receiving appropriate antimalarial drugs Percentage of children aged 0-59 months with fever receiving any antimalarial drugs Percentage of children aged 0-59 months with suspected pneumonia taken to an appropriate health care facility. Percentage of children aged 0-59 months with suspected pneumonia antibiotics receiving IMCI+ ACSD TARGET TARGET ACSD Case management malaria (effective) Case management malaria- programmatic (programmatic) Care seeking pneumonia § Antibiotic treatment of pneumonia

5 6 7 8 NO.

A30 IIP-JHU | Retrospective evaluation of ACSD in Ghana

PROTOCOL FOR MISSING/UNKNOWN DATA EXCLUDE CASES 1: Reported treatment of child but missing for ORS, RHF and IF and positive/missing for continued feeding EXCLUDE CASES 2: Reported treatment of child but positive/missing for ORS, RHF or IF and missing for continued feeding

¹ ² DOMINATOR DOMINATOR Children (0-59) with reported diarrhoea in previous two weeks, include MB, exclude deceased NUMERATOR received Eligible children ORS, RHF or increased fluids AND continued feeding Child DATAFILE DATAFILE

, or more increased (MICS) continued OR more INDICATORS INDICATORS AND AND RHF RHF about the same somewhat less Rec'd somewhat ORS packets ORS

Rec'd fluids feeding Percentage of children aged 0-59 months with ORS diarrhoea receiving OR (MICS) ORS ORT/RHF recommended home fluids IMCI+ Continued feeding ACSD TARGET TARGET ACSD Increased fluids (IF) ORS/RHF/increased fluids for children diarrhoeawith + continued feeding §

9 NO. § International Consensus Coverage Indicator Indicator Coverage Consensus § International birth children ¹ MB=Multiple birth: include all multiple ² CDC 2003 - valid data due issues, included children with for indicator variables to data quality were all

IIP-JHU | Retrospective evaluation of ACSD in Ghana A31

PROTOCOL FOR MISSING/UNKNOWN DATA EXCLUDE CASES: Reported ever of initiation breastfeeding, but missing timing EXCLUDE CASES: still Missing for breastfeeding and HH test EXCLUDE CASES: Missing salt no salt with test EXCLUDE CASES: Missing salt EXCLUDE CASES 1: Missing for all feeding still for AND positive/missing variables breastfeeding EXCLUDE CASES 2: Negative/ missing for all feeding variables AND missing for still breastfeeding EXCLUDE CASES 1: Missing for all feeding still for AND positive/missing variables breastfeeding EXCLUDE CASES for all 2: Positive/missing feeding variables AND missing for still breastfeeding

¹ ² Child (20-23) Child (20-23) Child (0-5) with Child (6-9) with most recently most recently most recently most recently most recently DOMINATOR DOMINATOR Women a birthwith in previous 12m completed All HH with surveys Children (20-23m): DHS: born (include only one & living still alive MB) mom.with MICS: completed with questionnaire completed All HH with and salt surveys available for testing Children (0-5m): DHS: born (include only one & living still alive MB) mom.with MICS: completed questionnaire Children (6-9m): DHS: born (include only one & living still alive MB) mom.with MICS: completed questionnaire. NUMERATOR Women initiated breastfeeding the first hourwithin after delivery with Eligible HH has salt >=15ppm iodine Eligible children still Eligible children breastfeeding still Eligible children breastfeeding and did not liquids or foods in receive any previous 24h still Eligible children breastfeeding and received in the solid/semisolid foods previous 24hr

HH HH Child Child Child Women DATAFILE DATAFILE HH HH exclude HH include HH foods) INDICATORS INDICATORS are breastfed and Percentage of infants aged 6-9 months who food (solid or semisolid receive complementary Percentage of newborns put to the breast within one hour of birth; most recent live birth previous 12m Percentage of children aged 20-23 months who are currently breastfeeding Percentage of households consuming iodized salt: Percentage of households consuming iodized salt: no salt with with no salt with Percentage of infants aged 0-5 months who breastfed are exclusively ACSD IMCI+ TARGET TARGET Timely initiation Timely of breastfeeding § Exclusive breastfeeding through 6 months (0-5m) § Breastfeeding and complementary feeding (6-9 months) § Continued breastfeeding (20-23 months) § Consumption of iodized salt Consumption of iodized salt

10 11 12 13 14 15 NO. § International Consensus Coverage Indicator Indicator Coverage Consensus § International birth children ¹ MB=Multiple birth: include all multiple ² CDC 2003 - valid data due issues, included children with for indicator variables to data quality were all

A32 IIP-JHU | Retrospective evaluation of ACSD in Ghana

PROTOCOL FOR MISSING/UNKNOWN DATA EXCLUDE CASES: Unknown/missing for slept under a bed net last night EXCLUDE CASES 1: Missing ITN data (a) Net obtained <=12m prior AND obtained missing if treated when (b)Treated net obtained AND missing months ago obtained (c) Treated the net after obtaining but missing months ago treated EXCLUDE CASES 2: Unknown/missing for a,b & c and for slept under positive/missing/unknown a net last night EXCLUDE CASES 3: Positive/missing for a, b & c AND for unknown/missing slept under a bed net last night

¹ ² DOMINATOR DOMINATOR pregnant All eligible women All children under five, still living pregnant All eligible women NUMERATOR slept Eligible pregnant woman under a mosquito net last night slept under an Eligible child ITN mosquito net last night slept Eligible pregnant woman under an ITN mosquito net last night

Woman Woman Woman DATAFILE DATAFILE INDICATORS INDICATORS Percentage of pregnant women sleeping under any mosquito net last night Percentage of children aged 0-59 months sleeping under an insecticide treated mosquito net (Use trt'd <=12 months due to heaping at 12m) Percentage of pregnant sleeping women under an insecticide treated mosquito net last night (Use trt'd <=12 months due to heaping at 12m) ITNs ACSD TARGET TARGET ACSD Use of bednets by pregnant women Effective use of bednets by § children < 5yr Effective use of bednets by pregnant women

16 17 18 NO. § International Consensus Coverage Indicator Indicator Coverage Consensus § International birth children ¹ MB=Multiple birth: include all multiple ² CDC 2003 - valid data due issues, included children with for indicator variables to data quality were all

IIP-JHU | Retrospective evaluation of ACSD in Ghana A33

PROTOCOL FOR MISSING/UNKNOWN DATA EXCLUDE CASES: Unknown/missing for number prenatal AND positive/missing visits skilled HCW for EXCLUDE CASES: Received medicine during pregnancy for malaria but unknown. of medicine missing type EXCLUDE CASES: Unknown/missing if received TT or received TT but unknown dosage EXCLUDE CASES: Unknown/missing if received iron or iron but for unknown time period

EXCLUDE CASES: Unknown/missing data for birth attendant EXCLUDE CASES: Unknown/missing if received TT or received TT but unknown dosage or date of most recent injection

¹ ² DOMINATOR DOMINATOR with women All eligible a pregnancy resulting in a live birth the previous 12m NUMERATOR 2+ does during pregnancy 1 dose during pregnancy + 2+ doses prior, the most 3+ doses prior, the most 4+ doses prior, the most 5+ lifetime doses Eligible women received 4+ Eligible women a prenatal care visits with trained health care worker received at Eligible women doses of SP during least two the pregnancy received at Eligible women doses of tetanus least two toxoid during the pregnancy received any Eligible women of the following: ▪ ▪ doses before any pregnancy ▪ recent 3 years before pregnancy ▪ recent 5 years before pregnancy ▪ recent 10 years before pregnancy ▪ received iron Eligible women for at supplementation daily least 90 days with delivered Eligible women a trained health care worker. Women Women Women Women Women Women DATAFILE DATAFILE INDICATORS INDICATORS Most recent live birth within previous 12m Percentage of pregnant women reportwho at least 4 a skilled prenatal visits to doctor, health worker: or auxiliary nurse/midwife midwife Percentage of pregnant receiving intermittent women preventative treatment for malaria during pregnancy Percentage of newborns protected against tetanus: Mother rec'd at least 2 doses of TT during pregnancy Percentage of newborns protected against tetanus: Mother rec’d immunity through injections previous to pregnancy Percentage of pregnant receiving 3 months of women iron supplementation. Percentage of births attended by skilled health doctor,worker: nurse/midwife midwife auxiliary or ANC+ ACSD TARGET TARGET ACSD 4+ prenatal visits, trained health care worker Intermittent malaria treatments in pregnancy coverage TT2 during pregnancy § Full TT coverage Pregnant women take 3 months iron supplements Skilled attendant § at delivery

19 20 21 22 23 24 NO.

A34 IIP-JHU | Retrospective evaluation of ACSD in Ghana

missing skilled HCW skilled missing PROTOCOL FOR MISSING/UNKNOWN DATA EXCLUDE CASES: Unknown/missing if A received vitamin EXCLUDE CASES 1: Unknown/missing place and no of delivery data for postnatal care EXCLUDE CASES 2: Non-institutional positive/ and delivery and positive/missing received postnatal care

¹ ²

previous 12m in a live birth in the in a live birth the DOMINATOR DOMINATOR a pregnancy resulting All eligible women with with women All eligible NUMERATOR (a) Eligible women delivered at (non- an institutional facility domestic) Eligible women who (b) delivered domestically received at least one postnatal 3 of checkup within days with a trained health delivery care worker Eligible women received Eligible women vitamin A supplementation 2 months of delivery within Women Women DATAFILE DATAFILE INDICATORS INDICATORS Most recent live birth within previous 12m Percentage of newborns receiving a postnatal visit by trained worker (doctor, a or auxiliary nurse/midwife of 3 days within midwife) delivery. Percentage of women A vitamin receiving 2 supplementation within months of birth ANC+ ACSD TARGET TARGET ACSD Postnatal visit 3 dayswithin of trained delivery, health worker Postnatal supplementation Vitamin A § with

25 26 NO. § International Consensus Coverage Indicator Indicator Coverage Consensus § International birth children ¹ MB=Multiple birth: include all multiple ² CDC 2003 - valid data due issues, included children with for indicator variables to data quality were all

IIP-JHU | Retrospective evaluation of ACSD in Ghana A35

§

OF CASES PROTOCOL FOR EXCLUSION values for with improbable Cases from are excluded height-for-age as defined improbable analysis; of Z deviations +/- 4 standard to the overall relative score the value from score median Z datafile crude values for with improbable Cases are excluded weight-for-height improbable from analysis; +/- 4 standard defined as to relative of Z score deviations value Z score the overall median datafile from the crude values for with improbable Cases from are excluded weight-for-age as defined improbable analysis; of Z deviations +/- 4 standard to the overall relative score the value from score median Z datafile crude N/A DOMINATOR§ DOMINATOR§ missing) birth month & year birth missing) measure anthropometric survey before the anthropometric measure measure anthropometric survey before the month & year birth missing) measure anthropometric survey before the 1. Have a reported (non- Have a reported 1. 2. (non-missing) Have a valid 3. the night Slept in the house 1. (non-missing) a valid Have 2. Slept in the house night (non- a reported Have 1. 2. (non-missing) a valid Have 3. Slept in the house night months aged 24-59 Children who: who: 0-23 months aged Children who: 0-59 months aged Children Expressed as 1000 live births ) 36 36 ) 36 ) ) ) ) Children <-3 z with Children <-3 z with Children <-3 z with years

36 36 36 NUMERATOR Severe: and Moderate <-2 z scores Children with for height for age based on the 2006 WHO growth curves( Severe: and Moderate <-2 z scores Children with for height based weight on the 2006 WHO growth curves( Severe: and Moderate <-2 z scores Children with for for age basedweight on the 2006 WHO growth curves(

Severe: scores for height for age based on the 2006 WHO curves(growth Severe: for scores for weight height based on the 2006 curves( WHO growth Severe: scores for for age weight based on the 2006 WHO curves(growth The probability of dying between birth and exact age five file from Birth history history women’s extracted Household Household Household DATAFILE DATAFILE

INDICATORS INDICATORS Stunting (low height for age) among children 24- 59 months of age* for (weight Wasting height) among children 0- 23 months of age* for (weight Underweight age) for children 0-59 months of age* Under-five mortality rate Definition of priority impact indicators impact indicators Definition of priority 1 2 3 4 NO.

Table D2:

A36 IIP-JHU | Retrospective evaluation of ACSD in Ghana

1) Unprotected well, 2) Unprotected 1) moved and the indices calculated for calculated the indices moved and ndex value, based on its reported assets assets ndex value, based on its reported d into quintiles based on their index on based d into quintiles r the ACSD evaluation, the urban areas areas urban r the ACSD evaluation, s, 4) vendor provided water, 5) Bottled water, provided s, 4) vendor

DESCRIPTION OF DEFINITION DESCRIPTION OF

Improved sanitation facilities include: sanitation facilities Improved sewer, to a public 1) Connection system, to a septic 2) Connection latrine, 3) Pour-flush 4) Simple pit latrine, pit latrine. 5) Ventilated improved . sanitation facilities include: Unimproved latrine, or shared 1) Public pit latrine, 2) Open 3) Bucket latrine. value. fo quintiles of wealth For the calculation are re Great Accra region of Ashanti and All household assets and utilities are dichotomized into indicator variables. indicator variables. into are dichotomized utilities assets and All household to variables the indicator all using is performed analysis components Principle factor. Each using the first principle the variables of weights the standardize i a weighted then assigned is household then divide are Households and utilities. area only. comparison the HIDs and in households Unimproved drinking water include: sources drinking Unimproved pond Rivers or 3) spring, Unprotected water, 6) Tanker truck water es/mdgreport/definit es/mdgreport/definit rg/mics/mics3/docs/D SOURCE OF DEFINITION MDG water and sanitation definitions definitions sanitation water and MDG (http://www.unicef.org/w ion.php) ion.php) DHS standard calculation of wealth quintiles of wealth quintiles calculation DHS standard (http://www.childinfo.o HS%20Wealth%20Index%20(DHS%20Com parative%20Reports).pdf) MDG water and sanitation definitions definitions sanitation water and MDG (http://www.unicef.org/w ion.php)

Definition of contextual variables in the ACSD evaluation Definition of used

VARIABLE Table D3: CONTEXTUAL Wealth quintiles Improved Water Source Improved Sanitation Facilities

IIP-JHU | Retrospective evaluation of ACSD in Ghana A37

APPENDIX E Survey questions used in the calculation of coverage indicators DHS Questionnaire DHS Questionnaire ACSD Questionnaire MICS Questionnaire NO. ACSD TARGET 1998/99 2003 2003 2006 & 2007/08 EPI+

Have vaccination card Have vaccination card Have vaccination card (IM1); Have vaccination card (q458); Measles innoc. on (q404); Measles innoc. on Measles innoc. on card (IM6); (q443); Measles innoc. on Measles immunization card (q460); Rec'd other card (q405); Rec'd other Rec'd other vaccines (IM10); Mom card (q444); Rec'd other 1 coverage vaccines (q462); Mom vaccines (q407); Mom report of measles innoc (IM17); vaccines (q446); Mom report report of measles innoc report of measles innoc rec'd vaccine during campaign of measles innoc (q447G) (q463G) (q414) (IM19)

Have vaccination card Have vaccination card Have vaccination card Have vaccination card (IM1); DPT3 (q443); DPT3 on card (q458); DPT3 on card (q404); DPT3 on card on card (IM5C); Rec'd other DPT3 immunization (q444); Rec'd other (q460); Rec'd other (q405); Rec'd other vaccines (IM10); Mom report of 2 coverage vaccines (q446); Mom report vaccines (q462); Mom vaccines (q407); Mom DPT(IM15); number of of DPT(q447E); number of report of DPT(q463E); report of DPT(q412); doses(IM16) doses(q447F) number of doses(q463F) number of doses(q413)

Hib3 immunization Have vaccination card (IM1); Hib3 N/A N/A N/A 3 coverage innoc. on card (IM5C)

Have vaccination card Have vaccination card Have vaccination card Have vaccination card (IM1); VitA Coverage of vitamin A in (q458); Mother's report (q458); VitA on card (q460); (q404); VitA on card (q405); on card (IM8a/b); Mother's report 4 last 6 months (q448) Mother's report (q457) Mother's report (q403) (VA1,VA2)

IMCI+ Had fever(q515); gave Had fever(q466); gave meds (q517); what meds Case management Had fever(q449); gave meds Had fever(ML1); gave meds meds (q473); what meds (q518); prescribed meds 5 malaria (effective) (q449A); what meds (q449B) (ML3/ML5); what meds (ML4/ML7) (q474) (q523); what meds prescribed(q524)

Had fever(q515); gave Case management Had fever(q466); gave meds (q517); what meds Had fever(ML1); gave meds malaria-programmatic N/A meds (q473); what meds (q518); prescribed meds 6 (ML3/ML5); what meds (ML4/ML7) (programmatic) (q474) (q523); what meds prescribed(q524)

Suspected pneum. (q450 & Suspected pneum. (q467 & Suspected pneum. (q511 & Suspected pneum. (CA5, CA6, q451); consulted for q468); consulted for q512); consulted for Care seeking pneumonia CA7); consulted for treatment 7 treatment (q452); where treatment (q470); where treatment (q513); where (CA8); where consulted (CA9) consulted (q453) consulted (q471) consulted (q514)

Suspected pneum. (q450 & Suspected pneum. (q511 & Suspected pneum. (CA5, CA6, Antibiotic treatment of q451); consulted for q512); consulted for N/A CA7); consulted for treatment 8 pneumonia treatment (q451A); where treatment (q513); where (CA8); where consulted (CA9) consulted (q451B) consulted (q514)

ORS/RHF/increased fluids for children with Had diarrhea (q454) Had diarrhea (q475) Had diarrhea (q501) Had diarrhea (CA1) diarrhoea + continued feeding

ORS ORS (q461) ORS (q478a) ORS (q506a) ORS (CA2a) 9 ORT/RHF RHF (q461) RHF (q478b) RHF (q506b) RHF (CA2b)

Increased fluids (IF) Increased fluids (q457) Increased fluids (q476) Increased fluids (q504) Increased fluids (CA3)

Continued feeding Continued feeding (q458) Continued feeding (q477) Continued feeding (q505) Continued feeding (CA4)

A38 IIP-JHU | Retrospective evaluation of ACSD in Ghana

DHS Questionnaire DHS Questionnaire ACSD Questionnaire MICS Questionnaire ACSD TARGET 1998/99 2003 2003 2006 & 2007/08

Ever breastfed (q440); Timely initiation of Ever breastfed (q427); Ever breastfed (q323); Timing of Ever breastfed (MN12); Timing of Timing of BF initiation 10 breastfeeding Timing of BF initiation (q425) BF initiation (q3243) BF initiation (MN13) (q441)

Still breasfeeding (q445); 11 Exclusive breastfeeding Still breasfeeding (q447); liquids in last 24h (q492a- Still breasfeeding (q326); Still breasfeeding (BF2); through 6 months (0- liquids & foods in last 24h e); food in last 24h (q493a- liquids/food in last 24h (q331b-g) liquids/foods(BF3) 5m) (q434) j)

Breastfeeding and Still breasfeeding (q447); Still breasfeeding (q445); Still breasfeeding (q326); food in Still breasfeeding (BF2); complementary feeding liquids & foods in last 24h 12 food in last 24h (q493a-j) last 24h (q331g) liquids/foods(BF3) (6-9 months) (q434)

Continued 13 breastfeeding (20-23 Still breasfeeding (q447) Still breasfeeding (q445) Still breasfeeding (q326) Still breasfeeding (BF2) months)

Consumption of iodized 14 salt Iodized salt (q35) Iodized salt (q35) Iodized salt (q29) Salt tested for iodization (SL1) Consumption of iodized 15 salt ITNs

Child slept under net last Child slept under net last night Child slept under net last night night (H32D); How long ago (ML10); How long ago was net (q465C); How long ago was net Effective use of was net obtained (H31); obtained (ML11); Brand of net obtained (q465E); Was a treated bednets by children < N/A Was a treated net obtained (ML12); Was a treated net 17 net obtained (q465F); Was the net 5yr (H31b); Was the net ever obtained (ML13); Was the net ever ever treated(q465G); How long treated(H32A); How long treated(ML14); How long ago ago treated (q465H) ago treated (H32B) treated (ML14)

Currently pregnant (q226); How long ago was net Effective use of obtained (H31); Was a 18 bednets by pregnant N/A treated net obtained N/A N/A women (H31b); Was the net ever treated(H32A); How long ago treated (H32B)

ANC+

Prenatal care and who did Prenatal care (q303);who did you Prenatal care and who did you 4+ prenatal visits, Prenatal care (q407); you consult (q407); Number consult (q304); Number of visits consult (MN2); Number of visits 19 skilled HCW Number of visits (q409) of visits (q409) (q306) (MN2bb)

Intermittent malaria Took meds for malaria Took meds for malaria (q223); Took meds for malaria (MN6A); treatments in N/A 20 (q421); Which meds (q422) Which meds (q224) Which meds (MN6B) pregnancy

TT2 coverage during Rec'd TT (q410); number of Rec'd TT (q415); number of Rec'd TT (q308); number of doses Rec'd TT (TT2); number of doses 21 pregnancy doses (q411) doses (q416) (q309) (TT3)

Pregnant women take 3 Rec'd iron (q417); Number Rec'd iron (q313); Number of days months iron N/A N/A 22 of days took iron (q418) took iron (q314) supplements

Skilled attendant at Assisted with birth (q414) Assisted with birth (q426) Assisted with birth (q320) Assisted with birth (MN7) 23 delivery

Location of delivery (q427); Rec'd postnatal care if non- Postnatal visit within 3 Days after delivery rec'd institutional delivery (q429) 24 days of delivery, skilled care (q417B); who N/A N/A HCW performed care (q417C) Days after delivery rec'd care (q430); who performed care (q431)

Postnatal 25 supplementation with Rec'd vitamin A (q417G) Rec'd vitamin A (q433) Rec'd vitamin A (q322) Rec'd vitamin A (MN1) Vitamin A

IIP-JHU | Retrospective evaluation of ACSD in Ghana A39

APPENDIX F Methodology and implementation of household surveys in Ghana 1998 to 2008

The methodologies and implementation of households surveys re-analyzed for the ACSD retrospective evaluation are presented in table F1. Less documentation of the methods and implementation was available for the ACSD 2003 survey. This survey is presented in the body of the report, but should be interpreted with caution due to questions about the data quality and the exact methodologies utilized. A full report describing data quality issues in the ACSD 2003 survey is available on request from the JHU evaluation team.

A40 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2007 MICS West regions) regions) West Ghana Statistical Upper East Region East Upper supplementary supplementary Northern and Upper Upper and Northern Services and UNICEF; TA (Macro and JHSPH) (also data available in (also data available Sampling methods / size; Sampling frame/ selection/weights; Revised questionnaire Training manual; Interviewer manual; Supervisory field report; Datafile for analysis frame/ manual; Revised analyses analyses Supervisor Interviewer for analysis; Services and Commission) questionnaire; Report of data MICS 2006 & Ghana AIDS / size; Sampling Training manual; manual; Datafile Ghana Statistical (PEPFAR, Macro Ministry of Health of Ministry selection/weights; with UNICEF; TA with Sampling methods

Revised analyses analyses National National National National Ghana DHS 2003 questionnaire; Report of data Training manual; Services & DHS, Ghana Statistical selection/weights; Macro International Supervisor manual; Interviewer manual; Sampling methods / Datafile for analysis; Datafile for analysis; size; Sampling frame/ hana 1998 to 2008 presented in the ACSD evaluation report report in the ACSD evaluation to 2008 presented 1998 hana

English Region Region UNICEF) Upper East East Upper questionnaires; ACSD 2003 Navrongo Health (CDC – Atlanta & Datafile for analysis Datafile for analysis , TA Research Centre Yes Yes Yes Yes Yes Yes Yes Yes Revised analyses analyses National National questionnaire; Report of data Training manual; Services & DHS, Ghana Statistical selection/weights; DHS 1998-99 Macro International Supervisor manual; Interviewer manual; Sampling methods / Datafile for analysis; Datafile for analysis; size; Sampling frame/ Geographic Geographic Coverage Implementing (& TA) Agency Datafile for available reanalysis Survey document- ation available Methodology and implementation of household surveys in G surveys in of household implementation and Methodology Survey Component Survey Table F1: General

IIP-JHU | Retrospective evaluation of ACSD in Ghana A41

20 2007 MICS HID: 173 full listing urban/rural household listing fieldsurvey work 2 stage sampling Done by GSS from Done by only; technical team throughout period of stratified by district & stratified by supplementary supplementary Listing of selected HH recommended standard, 300 UE, UW & early 2006 listing before Done by GSS by Done some re-listed 25 in rural UE, MICS 2006 from household clusters chosen UW & Northern; and urban/rural; survey fieldsurvey work Northern regions May – July 2005;May clusters selection 2 stage sampling, Standards Survey 5; oversampling in from Ghana Living stratified by region stratified by Complete listing in 20 in all other HHs

412 Ghana work point) work other regions 2000 census; DHS 2003 UW, Northern & and urban/rural; household listing clusters selection – JuneMay 2003 2 stage sampling, before survey field stratified by region stratified by Brong Ahafo; 16 in 20 in the UE, UW & oversampling in UE, Done by Macro from from Macro by Done Brong Ahafo regions Complete HH listing; clusters chosen from Northern and 15 in all

83 Unknown Unknown Unknown stratification ACSD 2003 2 stage sampling design; Unknown 400 work regions October 1998 EAs. August – household listing 2 stage sampling chosen from 1984 DHS 1998-99 before survey field Complete listing in in all other regions urban/rural; clusters EAs with <500 HHs; EAs with Done by Macro from from Macro by Done Northern regions; 15 stratified by region & stratified by partial listing in larger 20 in the UE, UW and census; oversampling in UE, UW & Northern selection selection Household Household Stratification Stratification & sampling of clusters of Number clusters of Number households per cluster Mapping/ listing Survey Component Survey Sampling & enumeration

A42 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2007 Kumasi MICS Immunization; Anthropometry Learning; Child National Health Illness; Malaria; Insurance; Birth Newborn Health; Under-five [Men's] Security of Tenure; of Security added to HH quest. Violence; HIV/AIDS; Marriage and Union; Full birth history was Toxoid; Maternal and supplementary supplementary Education; Vitamin A; Registration and Early Breastfeeding; Care of Pretested in peri-urban child mortality; Tetanus & Flooding module was Sociodemographic Info ; Contraception; Domestic added to quest. women’s Household, women's and Household, women's English English Tenure; Malaria; Domestic and Early Vitamin A; June 2006 Security of of Security and Union; Household, Pretested in women's and and Newborn Development; Greater Accra Immunization; Breastfeeding; Contraception; Anthropometry MICS 2006 Care of Illness; Learning; Child SES Info ; child HIV Knowledge; Health; Marriage region in 2 urban Birth Registration Violence; Female and 2 rural EAs in Genital Mutilation; Under-five [Men's] Sexual Behaviour; mortality; Maternal [Men's] module Dagbani Ghana Husband’s Pretest of all DHS 2003 breastfeeding; AIDS & STDs; AIDS & STDs; sexual activity; sexual activity; Contraception; Women’s Work; Work; Women’s local languages; Background and questionnaires in Immunization and Ewe, Nzema, Ewe, and Sociodemographic Akan, Ga, English, Info; Reproduction; urban & rural areas Also pretested AIDS Fertility Preferences;Fertility Household, women's 5-7 May 5-7 May 2003 in all 5 Health; Marriage and Pregnancies, ANC, &

women English women's Unknown Household, status, work of status, work Contraception; ACSD 2003 Hygiene, marital illness and care; & breastfeeding; Sociodemographic Info; Reproduction; Pregnancies, ANC, Immunzation; Child [Men's] Dagbani pretested. Pretest of all breastfeeding; Contraception; languages were Background and 1998; the 5 local Ewe, Hausa,Ewe, and Health; Marriage; Immunization and DHS 1998-99 Height and Weight Sociodemographic Akan, Ga, English, Husband/Partner’s Info; Reproduction; Fertility Preferences;Fertility Household, women's Pregnancies, ANC, & questionnaires in Sept Women’s Work; AIDS; Work; Women’s Modules aires used Questionn- included in Language of Language questionnaire questionnaire questionnaire questionnaire Pre-test / pilot women's/child aires Survey Component Survey Questionn-

IIP-JHU | Retrospective evaluation of ACSD in Ghana A43

driver 2007 JHSPH MICS trainees March 2008 urban Kumasi 4 teams in HIDs discussion of the Sept – Dec 2007 interviews among Interviewer manual 1 supervisor; 1 field Follow-up a few with 2007; TA by Macro & 2007; TA by supplementary supplementary editor; 4 interviewers; 1 editor; 4 interviewers; Two inTwo Aug-Sept weeks Interviewing techniques, 2 days conducted in peri- questionnaires, and mock additional clusters in Feb-

t teams driver guides trainees 9 and mock July, 2006 July, techniques, Interviewing field editor; 4 interviewers; 1 interviewers; 1 supervisor; 1 MICS 2006 3 month period 80 interviewers questionnaires, Standard MICS interviews in 16 discussion of the operators: 17–31 interviews among and 10 data entry urban & rural EAs 3 days conducting 2003 August 2006 September, 2008

guides training. 15 teams Ghana 1 nurse; 4 DHS 2003 October, 2003 Standard DHS Standard DHS Standard DHS 6-27 July 2003 Late July – lateLate July entry operators;entry training including Nurses trained in nurses & 12 data anemia and AIDS anthro. measures. blood collection for 102 interviewers, 23 interviewers; 1 interviewers; driver 1 supervisor; 1 editor;

Unknown Unknown Unknown Unknown Unknown ACSD 2003

training. training. 14 teams Nov 1998. 1 supervisor 1 field editor 3 interviewers 3 interviewers measurement. 1 driver (male) Standard DHS Standard DHS anthropometric Nov – Feb 1999 July – Sept 2003 (male or female) (male or female) (13/14 were male) (13/14 were DHS 1998-99 3 week period, Oct- 3 week Two day training on on training day Two Mid NovemberMid 1998 2003 July July Late Standard DHS guides Unknown up work work Manuals Survey start- Survey Period of field Period of field Logistics & timing Training content Practice admin survey in field team Survey composition of Number teams Survey Component Survey Training Field organization / work

A44 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2008 2007 Jones) MICS Upper East region JHU team for 1 week Similar to MICS 2006 Similar to MICS 2006 Similar to MICS 2006 supplementary supplementary completed early June by GSS, Macro, & IIP- by (Trevor Gareth Croft & Trevor Croft, consultant Similar to MICS 2006, TA Start-up supervision done Data GSS internal by GSS by checking Unknown fieldwork; fieldwork; field teams Completed consistency Field editors consistency by MICS 2006 standard MICS entry operatorsentry concurrent with concurrent with concurrent with concurrent with supervisors & 4 Office editors at November 2006 Data processing Data processing feedback sent to with 2 data entry double and entry edited/cleaned for GSS using CSPro Done according to fieldwork; 10 data fieldwork; secondary editors; GSS Ghana for internal 10 regional Field editors using CSPro DHS 2003 DHS standard concurrent with concurrent with concurrent with concurrent with entry operators;entry Completed mid- Data processing Data processing coordinators and double and entry fieldwork; 12 data sent to field teams Done according to December 2003 by by 2003 December fieldwork; feedback Data edited/cleaned consistency by GSS byconsistency supervised fieldwork statisticians acted as Office editors at GSS consistency checking GSS coordinated and

analysis analysis Unknown Unknown Unknown imputed to “6” ACSD 2003 SPSS & Stata for Files transferred to Missing birth monthMissing birth Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Yes Unknown Unknown Unknown Unknown developed for internal entered and edited using Field editors 1999 by GSS 1999 by DHS standard for DHS surveys. DHS 1998-99 Done according to (ISSA) programme for Survey Analysis Analysis for Survey The data were then Thedata were Data edited/cleaned consistency by GSS byconsistency microcomputers and Office editors at GSS the Integrated System the Integrated System Completed mid-March data team aires Quality Quality Technical Technical Editing of Data entry Data entry questionn- birth dates procedures procedures supervision supervision Data editing control loop control Observation Observation of interviews of interviews Imputation of Imputation of Finalization of Survey Component Survey Supervision Data processing

IIP-JHU | Retrospective evaluation of ACSD in Ghana A45

2007 MICS informants Field visits; key supplementary supplementary from UNICEfF MICS 2006 report; available Ghana MICS 2006 38 ) Ghana DHS 2003 DHS 2003 report(

from UNICEF; discussion with ACSD 2003 Files transferred Howard Goldberg Howard 37 ) DHS 1998-99 DHS 1998 report(

Sources: Survey Component Survey

A46 IIP-JHU | Retrospective evaluation of ACSD in Ghana

CI 92 61 82 97 l. 54 - 79 - 93 - 95% 95% .88 - .88 ) pp %) miss( MICS 2007 Su n% 396 80 0.9 397 95 0.7 1975 90 2 2256 58 0.5 CI 94 52 79 98 85 87 - 34 - 53 - 88 - 78 - 95% 95% vious 12 m not available in data %) miss( No Data Data No MICS 2006 n% 30 66 2 in card s applied to all children reported as vaccinated). 7 131 91 2 6 144 430 1 %) miss( 4 4 3 in previous 6 months only (pre 84 26 27 N/A n% 276 73 17 30 82 0 656 323 66 7 30 93 1.4 1394 CI 30 97 77 87 16 - 75 - 59 - 65 - 95% 95% %) miss( N/A same as DPT of children reported vaccination before 12m 2003 DHS 2003 ACSD-CDC time for ACSD “high impact” districts time 31 [32] 0 n/a 166 39 68 0 39 76 0 185 86 1 ³ 230 23 0.4 1 1 5 78 59 64 N/A 2002 2002 IHNS n% n% CI 71 70 78 59 - 50 - 58 - 95% 95% %) miss( ed on MICS protocols (where distribution N/A tors from IHNS 2002 survey report ( 2) ITN = Insecticide treated net defined as treated within 12 before months the survey or long-lasting net. No Data No Data 1998/1999 DHS 1998/1999 n% 38 60 1 38 68 0 155 65 2 G EPI+ and ITN coverage indicators over time in the “high impact” districts, Ghana (weighted) time in the “high impact” districts, EPI+ and ITN coverage indicators over

Tables presenting priority indicators over PPENDIX 2 (1) IHNS data not available; indica (3) Only available for children 6-32 months of age (4) Includes bednets treated A Table G1. Indicators* of children Percentage months who 12-23 aged DPT of 3 doses received vaccine of children Percentage months who 12-23 aged Hib against immunized are - 6 children of Percentage least at received who 59 A high dose vitamin one last the within supplement 6 months ITNs of children Percentage sleeping months 0-59 aged insecticide an under treated mosquito net (ITN) of pregnant Percentage an under sleeping women insecticide treated mosquito net (ITN) EPI+ of children Percentage months who 12-23 aged against immunized are measles *All vaccination indicators calculated bas *All vaccination indicators calculated

IIP-JHU | Retrospective evaluation of ACSD in Ghana A47

CI 33 58 60 58 24 - 41 - 48 - 95% 95% 41 - 41 0 6 %) Miss( 4 4 ACT % 50 51 .D) .D) n 2007 MICS suppl 2007 MICS 357 28 0.9 554 53 1.6 SEE APP

(

CI 40 76 20 -20 58 -58 95% 95% DHS %) n/a n/a 145 n/a n/a 136 Miss( 4 4 % n/a n/a 2006 MICS n

20 %) Miss( % CDC No data n 431 37 1 31 30 1 367 61 0.3 38 67 3.6 367 59 0.3 38 0.3 3.6 0 - 0.8 554 9 1.6 6 - 12 2003 ACSD- DEFINITION OF PNEUMONIA DIFFERENT FROM NY ANTIMALARIAL MEDICATION CI 46 87 84 18 - 54 - 47 - 95% 95% A MICS

(2) (4) %) Miss(

the “high impact” districts, Ghana (weighted) the “high impact” districts,

] No data No % chloroquine chloroquine ACT (since 2004) 2003 DHS CASES n 43 32 0 44 71 0 44 66 0 29 [66] 0 n/a 206 35 0 20

>24 1 % 24 39 53 BASED ON ; 2002 IHNS n URVEY REPORT S

2002

CI 49 89 72 28 - 68 - 36 - 95% 95% IHNS %) Miss( RESULTS NOT SHOWN : A / No data % N [ chloroquine chloroquine

1998/1999 DHS n 32 2 035 0 - 5 39 2 60 78 0 32 54 2 25-49] INDICATORS FROM FROM INDICATORS ; Illness case management indicators over time in management indicators Illness case 3

BASED ON CASES : DATA NOT AVAILABLE IMCI case case IMCI indicators management 2 Table G2. NCLUDED TREATMENT WITH APPROPRIATE ANTIMALARIAL ACCORDING TO NATIONAL POLICY IHNS I

(1) UNWEIGHTED Percentage of children aged 0-59 months with suspected pneumonia treatedantibiotics with Percentage of children aged 0-59 months with diarrhoea receiving ORS, RHF or increased fluids and continued feeding Percentage of children aged 0-59 months with fever receivingappropriate antimalarial drugs Percentage of children aged 0-59 months with antimalarial receiving fever drugs Ghana antimalarial policy Percentage of children aged 0-59 months with suspected pneumonia taken toan appropriate health provider (3) [

A48 IIP-JHU | Retrospective evaluation of ACSD in Ghana

CI 63 92 56 64 95% 95% 42 - 77 - 47 - 46 - %) miss( n%

2002

CI 55 484 52 0.8 29- 95% IHNS %) miss( 2006 MICS² 2006 suppl² MICS 2007 n% INDICATORSFROM ; %) miss(

.D) CDC n% 2003 ACSD- SEE APP (

DATA NOT AVAILABLE DATANOT DHS IHNS

CI 97 328 45 2 28 42 0 95% 95% 73 - (1)

] %) miss( CASES >24 2003 DHS 2003 21 n/a n/a n/a 93 82 8 25 [92] 0 n/a 125 84 0 28 [43] 0 n/a 168 39 3 32 [56] 0 n/a 258 55 0 18 n/a n/a n/a 94 50 5 30 [53] 0 n/a 159 53 0 BASED ON ; 33 2002 2002 IHNS¹ 95% CI n % n % RESULTS NOT SHOWN : A / UNABLE TO CALCULATE BREASTFEEDING INDICATORS AS N %) ; [

miss( 25-49] 1998/1999 DHS 1998/1999 n% NO FULL BIRTH HISTORY Feeding behaviour indicators over time in the “high impact” districts, Ghana (weighted) impact” districts, Ghana over time in the “high indicators Feeding behaviour

BASED ON CASES : 2006:

MICS

UNWEIGHTED Indicators Percentage of Percentage 20-23 aged children months who are currently breastfeeding 13 n/a n/a n/a Percentage of Percentage 0-5 infants aged months who are exclusively breastfed 39 of Percentage 6-9 infants aged [28]months who are 3 and breastfed receive n/a complementary food 19 n/a n/a n/a Percentage of Percentage the to put newborns one breast within of hour birth 39 11 0 3-19 45 85 0 [ (2) Table G3.

IIP-JHU | Retrospective evaluation of ACSD in Ghana A49

85 82 77 - 73 - 95% CI 95% %) miss( No data No data n% CI 61 485 40 1 35-46 68 480 8257 2 479 78-86 74 67 484 2 61-72 63 1 57-68 62 481 57 1 51-62 98 47895 89 478 2 86-92 81 2 86 482 78 1 33- 40- 48- 41- 85- 95% 95% 37 - 77 - 64 - 0 1 %) miss( 47 52 No data No data 2006 MICS 2007suppl MICS n% 28 28 23 47 16 %) miss( 27 <1 51 <1 CDC No data No data birth in the previous 12 month for “high impact” districts, birth in the previous 12 month for “high impact” districts, n% 2003 ACSD- 2003 332 322 74 4 28 75 0 332 CI 78 326 74 2 28 86 0 65 281 10 16 47 320 56 4 28 61 1 98 326 82 2 28 91 0 40- 19- 58- 95% 95% 50 - 5-34 10-30 47-70 2 0 0 0 %) miss(

52 18 20 No data WOMEN WITH INSTITUTIONAL DELIVERIES ASSUMED TO HAVE APPROPRIATE POSTNATAL CARE 2003 DHS (2) 45 58 45 33 0 44 45 45 64 0 45 45 4 0 0-10 45 4 0 0-10 318 5 5 23 54 16

45 81 47 2002 2002 IHNS¹ URVEY REPORT S

75 MIDWIFE ORAUXILLIARY MIDWIFE 52 - 52 / 95% CI n % n % 2002

%) miss( NURSE IHNS , No data No data 72 0 61-83 O CTOR 1998/1999 DHS 1998/1999 D No data on # of days n% 39 17 1 7-27 39 19 0 10-27 39 63 0 46-79 39 64 1 INDICATORS FROM FROM INDICATORS ; ³

Antenatal and postnatal care over time among womenAntenatal and with a live

DATA NOT AVAILABLE KILLED HEALTH WORKER DEFINED AS IHNS S

Indicators Percentage ofnewborns receivinga postnatal a visitskilled by health worker of² within3days delivery Percentage ofbirths skilled attended by health worker³ Percentage of pregnant women receiving intermittentpreventative treatment for malaria during pregnancy SP) in previous(any year Percentage of pregnant women receiving intermittentpreventative treatment for malaria during pregnancy in previous(2+ doses) year Percentage ofnewborns protected againsttetanus (2+ doses TTduring pregnancy) Percentage ofnewborns protected fully against tetanus Percentage of pregnant women receiving 3 months of iron supplementation. Percentage ofwomen receiving vitamin A supplementation within 2 months of 39 birth Percentage of pregnant women who reportat least 4 prenatalvisits to a skilled healthworker³ Percentage of pregnant women who reportat least 3 prenatalvisits to a skilled healthworker³ 39 74 1 66-82 45 78 0 Ghana (weighted) Table G4: (1) (3)

A50 IIP-JHU | Retrospective evaluation of ACSD in Ghana

95% CI %) Miss( (weighted)

No Data 2006 MICS 2006 MICS Comparison area ¥ Comparison n% 555 50 0 44 - 56

2006 MICS/ 2007 MICS suppl. suppl. MICS 2007 MICS/ 2006 No Data HID 2007 MICS suppl MICS same as DPT U E R 2 SD “high impact” districts and the SHANTI REGIONS

95% CI95% n % A

. (%) Miss Miss CCRA AND No Data A LASTING NET - REATER REATER n% 2003 DHS 2003 G

] HID area ¥ Comparison No Data No 31 [32] 328 2 0.1 0.5 - 4 230 23 2765 3 0.4 2 - 4 2256 58 2668 24 0.9 21 - 26 CASES >24 95% CI n % rity indicators over time in AC S AND URBAN AREAS OF THE (%) MONTHS BEFORE THE SURVEY OR LONG Miss HID BASED ON ; 12 No Data No Data No Data n% EXCLUDING THE , 1998/1999 DHS 1998/1999 RESULTS NOT SHOWN :

A / N HID area ¥ Comparison [

No Data No Data No Data n% 38 6038 511 68 60 515 1 55 - 64 65 39 1 68 60 - 70 536 39 67 76 0.5 537 64 - 71 76 396 0.3 80 73 - 80 397 549 95 78 1.1 545 76 - 79 79 1.8 78 - 81 155 65 2099 22 3 17 - 26 185 86 2271 79 2 77 - 81 1975 90 2368 96 1 95 - 97 NATIONALLEVEL 25-49]

– HANA G EPI+ and ITN coverage indicators over time in “high impact” districts and comparison areas, Ghana time in “high impact” districts and EPI+ and ITN coverage indicators over H

BASED ON CASES : NSECTICIDE TREATED NET DEFINED AS TREATED WITHIN I

=

Tables presenting comparisons of prio Table H1. comparison area OMPARISON AREA IS ITN

PPENDIX C

UNWEIGHTED Indicators ITN ¹ EPI+ * Percentage of children aged 12-23 12-23 aged children of Percentage against immunized are who months measles 12-23 aged children of Percentage of 3 doses received who months vaccineDPT 12-23 aged children of Percentage against immunized are who months Hib 6 - 59 who children of Percentage receivedat least one high dose last the within A supplement vitamin 6 months 0-59 aged children of Percentage an under sleeping months insecticide treated mosquitonet (ITN) women pregnant of Percentage insecticide an under sleeping (ITN) net mosquito treated ¥ (1) A [

IIP-JHU | Retrospective evaluation of ACSD in Ghana A51

95% CI 95%

%) Miss( ACT 2006 MICS Comparison area ¥ ¥ area Comparison (weighted) n%

HID HID ACT 2006 MICS/ 2007 MICS suppl. suppl. MICS 2007 MICS/ 2006 suppl

136 51 ³ 146 35 ³ 7 27 - 43 2007 MICS 95% CI n % SHANTI REGIONS A %) Miss( No Data Data No chloroquine and comparison areas, Ghana CCRA AND A Comparison area ¥ area Comparison n% 2003 DHS REATER G HID No Data 44 66 518 64 6 - 68 60 554 9 600 3 2.8 - 1 5 chloroquine igh impact” districts igh impact” districts 95% CI95% n % AND URBAN AREAS OF %) Miss( UER .D) No Data chloroquine SEE APP NCLUDED TREATMENT WITH APPROPRIATE ANTIMALARIAL ACCORDING TO NATIONAL POLICY ( I

n% DHS EXCLUDING THE , 1998/1999 DHS 1998/1999 HID ¥ area Comparison No Data No n% 35 39 430 23 1 18 - 28 43 32 396 38 0.7 33 - 43 357 28 405 30 1.2 25 - 35 32 54 33632 22 0 2 17 - 26 318 29 20 [66] 5 276 15 - 25 40 1.1 33 - 47 145 50 ³ 158 36 ³ 0 28 - 44 60 78 637 60 0 55 - 64 44 71 518 67 6 62 - 72 554 53 602 61 2.5 55 - 68 chloroquine (2) NATIONALLEVEL – n i y HANA G ever f

ith Case management indicators over time in “h Case management indicators m w

59 - 0 DEFINITION OF PNEUMONIA DIFFERENT FROM ren ren

OMPARISON AREA IS NY ANTIMALARIAL MEDICATION C

A MICS

IMCI case management management case IMCI indicators Percentage of children aged 0-59 0-59 aged of children Percentage months diarrhoea with receiving ORS, RHF or increased fluids feeding continued and Child 0-59 aged of children Percentage months suspected with an taken to pneumonia provider health appropriate Percentage of children aged 0-59 0-59 aged of children Percentage months suspected with with treated pneumonia antibiotics Percentage of children aged 0-59 0-59 aged of children Percentage fever receiving months with (program)¹ drugs antimalarial Ghana antimalarial polic previous 2 weeks, rec'd rec'd weeks, 2 previous antimalarial appropriate treatment (effective)²

Table H2. ¥ (1) (3)

A52 IIP-JHU | Retrospective evaluation of ACSD in Ghana

95% CI

7 %) Miss( 29 26 - 33 2006 MICS Comparison area ¥ n% 4027

2006 MICS/ 2007 suppl. ¹ MICS n% HID 2007 3222 12 MICS suppl MICS UMASI K CI 95% 95% CCRA AND %) A Miss(

.D) SEE APP ( n%

2003 DHS DHS

] HID Comparison area ¥ n% impact” districts and comparison areas, Ghana (weighted) areas, impact” districts and comparison 45 85 56628 [43] 41 240 018 37 - 45 49 n/a 484 0.6 42 - 56 188 5221 258 64 n/a 527 55 0.3 131 55 - 72 31 297 159 71 1.1 51 53 26 - 35 1 62 - 80 0.2 180 44 - 58 125 57 84 0.7 49 - 65 183 60 0 51 - 68 259 6 4123 23 9 20 - 25 273 6 4441 21 1.4 19 - 23 3314 12 4311 27 0.8 24 - 31 CASES >24 CI 30 27 36 74 68 27 - 27 19 - 19 - 23 - 60 - 54 95% 95% AND THE MAJOR METROPOLITAN AREAS OF 2 %) Miss( UER BASED ON ; No data No 28 n% EXCLUDING THE , 1998/1999 DHS 16 RESULTS NOT SHOWN : A HID Comparison area ¥ / UNABLE TO CALCULATE BREASTFEEDING INDICATORS AS No data N n% ; [ 39 1139 509 23 [28] 0.1 23519 29 n/a 0 19013 67 n/a 1 165 61 0

270 5606 NATIONALLEVEL 25-49] – HANA G Feeding behaviour indicators over time in “high Feeding behaviour NO FULL BIRTH HISTORY BASED ON CASES : 2006:

OMPARISON AREA IS MICS

C

UNWEIGHTED IMCI feeding behavior IMCI behavior feeding indicators Percentage of newborns put to put ofnewborns Percentage ofbirth hour one breast within the Percentage of infants aged 0-5 months are exclusively who breastfed Percentage of infants aged 6-9 months are breastfed who and complementaryreceive food 20- aged ofPercentage children 23 months are currently who breastfeeding ofPercentage households salt consuming iodized (>=15ppm) ofPercentage households salt consuming iodized (>=15ppm) no salt) (exclude HH with (1) , Table H3. [ ¥

IIP-JHU | Retrospective evaluation of ACSD in Ghana A53

37 26 - %) CI 95% Miss( No data No data No 2006 MICS MICS 2006 Comparison area ¥ area Comparison 2006 MICS/ 2007 MICS suppl. 2007 MICS 2006 MICS/ No data No data No HID 2007 HID 2007

482 78 529 75 1 70 - 79 MICS suppl UMASI K , nurse/midwife midwifeor auxilliary %) CI 95% n % n % CCRA AND Miss( A No data No 2003 DHS “high impact” districts and comparison areas, Ghana (weighted) and “high impact” districts HID Comparison area ¥ No data No 45 7845 552 64 76 552 2 58 72 - 80 478 2 54 - 63 89 478 53245 81 80 33 532 0.3 562 76 - 84 66 47 0.3 0.6 61 - 71 43 - 5145 484 18 6345 566 20 531 3545 60 664 0 58 35 30 - 40 0.5 55 - 65 485 0.4 566 31 - 40 36 40 0 533 32 - 40 42 481 57 0 36 - 48 531 48 0.5 43 - 53 45 445 542 1 4 4 542 0 - 2 1 480 444 82 0 - 1 52 479 500 506 67 43 33 11 6 500 29 - 38 37 - 48 31 6 CI n % n % 70 59 51 43 42 27 61 - 61 - 50 - 41 - 33 - 33 - 19 95% %) AND THE MAJOR METROPOLITAN AREAS OF Miss( postnatal care (2) Skilled health worker defined as Doctor as defined worker health Skilled (2) postnatal care UER No data No data No data No data No Comparison area ¥ Comparison 1998/1999 DHS EXCLUDING THE , HID n% n% No data No No data No data No data No 39 7439 505 64 66 505 1 55 1 39 63 505 46 0.7 39 1739 508 19 3839 0.2 508 72 38 0.1 509 23 0 NATIONALLEVEL eries assumed to have appropriate appropriate to have eries assumed – HANA G Antenatal delivery and postnatal care indicators over time in Antenatal delivery and postnatal care

OMPARISON AREA IS C

ANC, assisted delivery and delivery assisted ANC, postnatal care indicators Percentage of pregnant women who report report women who of pregnant Percentage skilled a health to visits prenatal 3 least at worker ² report women who of pregnant Percentage skilled a health to visits prenatal 4 least at worker ² women receiving of pregnant Percentage treatment for preventative intermittent year previous in pregnancy during malaria SP) (any women receiving of pregnant Percentage treatment for preventative intermittent year previous in pregnancy during malaria SP) (2+ against protected of newborns Percentage doses TT pregnancy) (2+ tetanus during protected fully of newborns Percentage tetanus against 3 women receiving of pregnant Percentage months of iron supplementation. skilled by attended of births Percentage ² worker health a receiving of newborns Percentage worker a skilled health by visit postnatal ² ofdelivery¹ 3 days within vitamin A of receiving women Percentage months 2 of birth within supplementation Table H4: ¥ (1)Women with institutional deliv

A54 IIP-JHU | Retrospective evaluation of ACSD in Ghana

103 272 591 206 377 420 226 295 199 2268 5% 9% 9% 51% 1146 49% 1122 81% 1848 12% 26% 17% 19% 10% 13% Percent Total Children lete interviews p Table I1c: Under five Children with five Children Under Table I1c: com Total Female Male 3 Rural 2 6 7 8 Urban 5 1 Sex 4 Residence 3 702 674 229 171 798 294 437 317 487 771 378 406 914 488 3288 1911 2517 the HIDs by socio-demographic characteristics of 28 15 7% 5% 9% 21% 21% 58% 24% 13% 77% 10% 15% 23% 12% 11% 0.1% Percent Total Women Table I1b: Eligible women with Eligible Table I1b: complete interviews Primary Primary Secondary + Total Education None Currently pregnant Currently Yes Not sure <24 <12 6 7 8 Months since last birth 5 3 Rural 2 4 Urban 1 Residence 712 344 412 319 184 395 744 567 390 2580 3324 6% 10% 78% 22% 17% 12% 10% 12% Percent Total Households Description of households, eligible women and children under five supplemental MICS 2007-8 in the “high impact” districts, MICS 2007-8 in the “high impact” districts, five supplemental and children under households, eligible women of Description

Ghana Tables presenting 2007-8 results for key indicators in population Table I1: Table I1a: Households interviewed Households Table I1a: Total Bawku Municipality 21% Bawku West Bolgatanga Municipality 12% Talensi-Nabdam Garu-Tempane Residence Bongo Rural Urban Districts Districts Districts Districts Districts Builsa Kasena-Nankana Appendix I

IIP-JHU | Retrospective evaluation of ACSD in Ghana A55

- 87 77 72 84 79 85 188 310 209 397 23m¹ Number of children 12 ² rd, by the total total the by rd, 98 95 94 89 94 92 99 95 99 95 (%) ACSD indicator indicator (%) vaccine vaccine Innoculated against DPT before 12m 12m before (%) Rec'd vaccine vaccine - 87 99% 99% 75 92% 97% 72 92% 100% 85 99% 100% 7985 98% 99% 98% 100% 187 97% 98% 309 95% 99% 208 95% 99% 396 96% 99% 23m¹ Number of children 12 ² (%) 77% 82% 81% 79% 79% 71% 86% 86% 78% 80% ACSD indicator indicator (%) vaccine before 12m Innoculated against measles n/a n/a n/a n/a n/a n/a n/a n/a (%) Rec'd vaccine vaccine 83% 99% 77% 84% 98% 84% 84% 98% 83% 88% 85% 99% 80% 84% 99% 80% 83% 97% 73% 86% 99% 87% 83% 99% 78% 82% 97% 89% 84% 98% 82% Seen % with % with EPI card- 2 3 4 Male Rural Urban Builsa 88% n/a n/a n/a n/a n/a n/a n/a n/a Bongo 87% n/a n/a n/a n/a n/a n/a n/a n/a Female Poorest Least Poor Bawku West 80% n/a n/a n/a n/a n/a n/a n/a n/a Garu-Tempane Talensi-Nabdam 84% n/a n/a n/a n/a n/a n/a n/a n/a Total Kasena-Nankana 87% n/a n/a n/a n/a n/a n/a n/a n/a Vaccination by socio-demographic characteristics in “high-impact” districts, Ghana 2007-8 “high-impact” districts, in by socio-demographic characteristics Vaccination

Bawku Municipality 77% n/a n/a n/a n/a n/a n/a n/a n/a Bolgatanga Municipality 85% n/a n/a n/a n/a n/a n/a n/a n/a ACSDindicator: multiply the percent of children thatreceived vaccination before first birthday, according toca immunization Children 12-23 months Children 12-23 of age, still alivewith non-missing data for indicator calculation:weighted Children 12-23 months of age vaccinated against mealses and DPT Districts ² ¹ Residence percentage of children vaccinated, according to card or mother’s report. Sex Wealth index quintiles Table I2:

A56 IIP-JHU | Retrospective evaluation of ACSD in Ghana

socio-demographic characteristics in “high-impact” districts, “high-impact” districts, in socio-demographic characteristics 87 377 417 417 389 375 221 395 460 460 439 991 984 368 332 176 198 515 232 175 260 age¹ 1975 1607 6-59 months6-59 of Number of children 90% 93% 88% 91% 88% 90% 72% 90% 92% 91% 94% 90% 90% 89% 93% 83% 89% 96% 93% 96% 74% 91% 88% Vitamin A A Vitamin supplementation (%)supplementation 4 3 2 6-11 Male Rural 12-23 24-35 36-47 48-59 Urban Builsa Bongo Female Poorest Least Poor Bawku West Bawku Garu-Tempane Vitamin A supplementation (one-dose) in previous 6 months by

Talensi-Nabdam Total Kasena-Nankana Bawku MunicipalityBawku Bolgatanga Municipality Children 6-59 months of age, still6-59 alive Children withfor non-missing data Children 6-59 months of age receiving one dose months of age receiving 6-59 Children 6 months in the previous supplementation A vitamin ¹ indicator calculation: weighted Wealth index quintiles Age in months Age Sex Residence Districts Districts Table I3: Ghana 2007-8

IIP-JHU | Retrospective evaluation of ACSD in Ghana A57

102 196 268 225 589 205 376 294 418 483 399 463 507 405 434 452 473 469 428 age² 1838 1119 1137 2256 months of Number of children 0-59 0-59 children (%) 76% 65% 54% 56% 45% 72% 53% 72% 59% 53% 57% 62% 67% 63% 49% 48% 58% 56% 58% 58% 57% 59% 58% ITN* last night Slept under an an under Slept 76% 58% 54% 62% 55% 60% 47% 76% 69% 64% 61% 71% 76% 60% 61% 61% 61% 60% 58% 61% 66% 53% 50% last night (%) Sleptever- under an treated mosquito net Percentage of children who: children of Percentage 62% 55% 61% 70% 65% 61% 61% 61% 62% 61% 61% 61% night (%) Slept under mosquito net last last net mosquito 2 3 4 0-11 Male Rural 12-23 24-3536-4748-59 67% 54% 51% Urban Builsa 78% Bongo 76% Female Poorest Least poor Bawku West 58% Utilization of bednets by children aged 0-59 months by socio-demographic characteristics in “high-impact” districts, Ghana districts, in “high-impact” characteristics 0-59 months by socio-demographic by children aged bednets Utilization of Garu-Tempane 55% Talensi-Nabdam 77% Kasena-Nankana 59% Bawku Municipality 47% Bolgatanga Municipality 71% ITN=Mosquito net treated with insecticide in the previous 12 months,or along-lasting net Total children under five who slept inlast HH night,with non-missing dataindicator for calculation: Childrenaged 0-59 months sleeping undera mosquito net, treated a net or an ITN¹ thenight preceding the survey Residence Districts Sex ¹ Age in months months in Age quintiles index Wealth Total ² weighted 2007-8

Table I4:

A58 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2007 MICS West regions) regions) West Ghana Statistical Upper East Region East Upper supplementary supplementary Northern and Upper Upper and Northern Services and UNICEF; TA (Macro and JHSPH) (also data available in (also data available Sampling methods / size; Sampling frame/ selection/weights; Revised questionnaire Training manual; Interviewer manual; Supervisory field report; Datafile for analysis frame/ manual; Revised analyses analyses Supervisor Interviewer for analysis; Services and Commission) questionnaire; Report of data MICS 2006 & Ghana AIDS / size; Sampling Training manual; manual; Datafile Ghana Statistical (PEPFAR, Macro Ministry of Health of Ministry selection/weights; with UNICEF; TA with Sampling methods

Revised analyses analyses National National National National Ghana DHS 2003 questionnaire; Report of data Training manual; Services & DHS, Ghana Statistical selection/weights; Macro International Supervisor manual; Interviewer manual; Sampling methods / Datafile for analysis; Datafile for analysis; size; Sampling frame/ hana 1998 to 2008 presented in the ACSD evaluation report report in the ACSD evaluation to 2008 presented 1998 hana

English Region Region UNICEF) Upper East East Upper questionnaires; ACSD 2003 Navrongo Health (CDC – Atlanta & Datafile for analysis Datafile for analysis , TA Research Centre Yes Yes Yes Yes Yes Yes Yes Yes Revised analyses analyses National National questionnaire; Report of data Training manual; Services & DHS, Ghana Statistical selection/weights; DHS 1998-99 Macro International Supervisor manual; Interviewer manual; Sampling methods / Datafile for analysis; Datafile for analysis; size; Sampling frame/ Geographic Geographic Coverage Implementing (& TA) Agency Datafile for available reanalysis Survey document- ation available Methodology and implementation of household surveys in G surveys in of household implementation and Methodology Survey Component Survey Table F1: General

IIP-JHU | Retrospective evaluation of ACSD in Ghana A59

20 2007 MICS HID: 173 full listing urban/rural household listing fieldsurvey work 2 stage sampling Done by GSS from Done by only; technical team throughout period of stratified by district & stratified by supplementary supplementary Listing of selected HH recommended standard, 300 UE, UW & early 2006 listing before Done by GSS by Done some re-listed 25 in rural UE, MICS 2006 from household clusters chosen UW & Northern; and urban/rural; survey fieldsurvey work Northern regions May – July 2005;May clusters selection 2 stage sampling, Standards Survey 5; oversampling in from Ghana Living stratified by region stratified by Complete listing in 20 in all other HHs

412 Ghana work point) work other regions 2000 census; DHS 2003 UW, Northern & and urban/rural; household listing clusters selection – JuneMay 2003 2 stage sampling, before survey field stratified by region stratified by Brong Ahafo; 16 in 20 in the UE, UW & oversampling in UE, Done by Macro from from Macro by Done Brong Ahafo regions Complete HH listing; clusters chosen from Northern and 15 in all

83 Unknown Unknown Unknown stratification ACSD 2003 2 stage sampling design; Unknown 400 work regions October 1998 EAs. August – household listing 2 stage sampling chosen from 1984 DHS 1998-99 before survey field Complete listing in in all other regions urban/rural; clusters EAs with <500 HHs; EAs with Done by Macro from from Macro by Done Northern regions; 15 stratified by region & stratified by partial listing in larger 20 in the UE, UW and census; oversampling in UE, UW & Northern selection selection Household Household Stratification Stratification & sampling of clusters of Number clusters of Number households per cluster Mapping/ listing Survey Component Survey Sampling & enumeration

A60 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2007 Kumasi MICS Immunization; Anthropometry Learning; Child National Health Illness; Malaria; Insurance; Birth Newborn Health; Under-five [Men's] Security of Tenure; of Security added to HH quest. Violence; HIV/AIDS; Marriage and Union; Full birth history was Toxoid; Maternal and supplementary supplementary Education; Vitamin A; Registration and Early Breastfeeding; Care of Pretested in peri-urban child mortality; Tetanus & Flooding module was Sociodemographic Info ; Contraception; Domestic added to quest. women’s Household, women's and Household, women's English English Tenure; Malaria; Domestic and Early Vitamin A; June 2006 Security of of Security and Union; Household, Pretested in women's and and Newborn Development; Greater Accra Immunization; Breastfeeding; Contraception; Anthropometry MICS 2006 Care of Illness; Learning; Child SES Info ; child HIV Knowledge; Health; Marriage region in 2 urban Birth Registration Violence; Female and 2 rural EAs in Genital Mutilation; Under-five [Men's] Sexual Behaviour; mortality; Maternal [Men's] module Dagbani Ghana Husband’s Pretest of all DHS 2003 breastfeeding; AIDS & STDs; AIDS & STDs; sexual activity; sexual activity; Contraception; Women’s Work; Work; Women’s local languages; Background and questionnaires in Immunization and Ewe, Nzema, Ewe, and Sociodemographic Akan, Ga, English, Info; Reproduction; urban & rural areas Also pretested AIDS Fertility Preferences;Fertility Household, women's 5-7 May 5-7 May 2003 in all 5 Health; Marriage and Pregnancies, ANC, &

women English women's Unknown Household, status, work of status, work Contraception; ACSD 2003 Hygiene, marital illness and care; & breastfeeding; Sociodemographic Info; Reproduction; Pregnancies, ANC, Immunzation; Child [Men's] Dagbani pretested. Pretest of all breastfeeding; Contraception; languages were Background and 1998; the 5 local Ewe, Hausa,Ewe, and Health; Marriage; Immunization and DHS 1998-99 Height and Weight Sociodemographic Akan, Ga, English, Husband/Partner’s Info; Reproduction; Fertility Preferences;Fertility Household, women's Pregnancies, ANC, & questionnaires in Sept Women’s Work; AIDS; Work; Women’s Modules aires used Questionn- included in Language of Language questionnaire questionnaire questionnaire questionnaire Pre-test / pilot women's/child aires Survey Component Survey Questionn-

IIP-JHU | Retrospective evaluation of ACSD in Ghana A61

driver 2007 JHSPH MICS trainees March 2008 urban Kumasi 4 teams in HIDs discussion of the Sept – Dec 2007 interviews among Interviewer manual 1 supervisor; 1 field Follow-up a few with 2007; TA by Macro & 2007; TA by supplementary supplementary editor; 4 interviewers; 1 editor; 4 interviewers; Two inTwo Aug-Sept weeks Interviewing techniques, 2 days conducted in peri- questionnaires, and mock additional clusters in Feb-

t teams driver guides trainees 9 and mock July, 2006 July, techniques, Interviewing field editor; 4 interviewers; 1 interviewers; 1 supervisor; 1 MICS 2006 3 month period 80 interviewers questionnaires, Standard MICS interviews in 16 discussion of the operators: 17–31 interviews among and 10 data entry urban & rural EAs 3 days conducting 2003 August 2006 September, 2008

guides training. 15 teams Ghana 1 nurse; 4 DHS 2003 October, 2003 Standard DHS Standard DHS Standard DHS 6-27 July 2003 Late July – lateLate July entry operators;entry training including Nurses trained in nurses & 12 data anemia and AIDS anthro. measures. blood collection for 102 interviewers, 23 interviewers; 1 interviewers; driver 1 supervisor; 1 editor;

Unknown Unknown Unknown Unknown Unknown ACSD 2003

training. training. 14 teams Nov 1998. 1 supervisor 1 field editor 3 interviewers 3 interviewers measurement. 1 driver (male) Standard DHS Standard DHS anthropometric Nov – Feb 1999 July – Sept 2003 (male or female) (male or female) (13/14 were male) (13/14 were DHS 1998-99 3 week period, Oct- 3 week Two day training on on training day Two Mid NovemberMid 1998 2003 July July Late Standard DHS guides Unknown up work work Manuals Survey start- Survey Period of field Period of field Logistics & timing Training content Practice admin survey in field team Survey composition of Number teams Survey Component Survey Training Field organization / work

A62 IIP-JHU | Retrospective evaluation of ACSD in Ghana

2008 2007 Jones) MICS Upper East region JHU team for 1 week Similar to MICS 2006 Similar to MICS 2006 Similar to MICS 2006 supplementary supplementary completed early June by GSS, Macro, & IIP- by (Trevor Gareth Croft & Trevor Croft, consultant Similar to MICS 2006, TA Start-up supervision done Data GSS internal by GSS by checking Unknown fieldwork; fieldwork; field teams Completed consistency Field editors consistency by MICS 2006 standard MICS entry operatorsentry concurrent with concurrent with concurrent with concurrent with supervisors & 4 Office editors at November 2006 Data processing Data processing feedback sent to with 2 data entry double and entry edited/cleaned for GSS using CSPro Done according to fieldwork; 10 data fieldwork; secondary editors; GSS Ghana for internal 10 regional Field editors using CSPro DHS 2003 DHS standard concurrent with concurrent with concurrent with concurrent with entry operators;entry Completed mid- Data processing Data processing coordinators and double and entry fieldwork; 12 data sent to field teams Done according to December 2003 by by 2003 December fieldwork; feedback Data edited/cleaned consistency by GSS byconsistency supervised fieldwork statisticians acted as Office editors at GSS consistency checking GSS coordinated and

analysis analysis Unknown Unknown Unknown imputed to “6” ACSD 2003 SPSS & Stata for Files transferred to Missing birth monthMissing birth Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Yes Unknown Unknown Unknown Unknown developed for internal entered and edited using Field editors 1999 by GSS 1999 by DHS standard for DHS surveys. DHS 1998-99 Done according to (ISSA) programme for Survey Analysis Analysis for Survey The data were then Thedata were Data edited/cleaned consistency by GSS byconsistency microcomputers and Office editors at GSS the Integrated System the Integrated System Completed mid-March data team aires Quality Quality Technical Technical Editing of Data entry Data entry questionn- birth dates procedures procedures supervision supervision Data editing control loop control Observation Observation of interviews of interviews Imputation of Imputation of Finalization of Survey Component Survey Supervision Data processing

IIP-JHU | Retrospective evaluation of ACSD in Ghana A63

2007 MICS informants Field visits; key supplementary supplementary from UNICEfF MICS 2006 report; available Ghana MICS 2006 38 ) Ghana DHS 2003 DHS 2003 report(

from UNICEF; discussion with ACSD 2003 Files transferred Howard Goldberg Howard 37 ) DHS 1998-99 DHS 1998 report(

Sources: Survey Component Survey

A64 IIP-JHU | Retrospective evaluation of ACSD in Ghana

CI 92 61 82 97 l. 54 - 79 - 93 - 95% 95% .88 - .88 ) pp %) miss( MICS 2007 Su n% 396 80 0.9 397 95 0.7 1975 90 2 2256 58 0.5 CI 94 52 79 98 85 87 - 34 - 53 - 88 - 78 - 95% 95% vious 12 m not available in data %) miss( No Data Data No MICS 2006 n% 30 66 2 in card s applied to all children reported as vaccinated). 7 131 91 2 6 144 430 1 %) miss( 4 4 3 in previous 6 months only (pre 84 26 27 N/A n% 276 73 17 30 82 0 656 323 66 7 30 93 1.4 1394 CI 30 97 77 87 16 - 75 - 59 - 65 - 95% 95% %) miss( N/A same as DPT of children reported vaccination before 12m 2003 DHS 2003 ACSD-CDC time for ACSD “high impact” districts time 31 [32] 0 n/a 166 39 68 0 39 76 0 185 86 1 ³ 230 23 0.4 1 1 5 78 59 64 N/A 2002 2002 IHNS n% n% CI 71 70 78 59 - 50 - 58 - 95% 95% %) miss( ed on MICS protocols (where distribution N/A tors from IHNS 2002 survey report ( 2) ITN = Insecticide treated net defined as treated within 12 before months the survey or long-lasting net. No Data No Data 1998/1999 DHS 1998/1999 n% 38 60 1 38 68 0 155 65 2 G EPI+ and ITN coverage indicators over time in the “high impact” districts, Ghana (weighted) time in the “high impact” districts, EPI+ and ITN coverage indicators over

Tables presenting priority indicators over PPENDIX 2 (1) IHNS data not available; indica (3) Only available for children 6-32 months of age (4) Includes bednets treated A Table G1. Indicators* of children Percentage months who 12-23 aged DPT of 3 doses received vaccine of children Percentage months who 12-23 aged Hib against immunized are - 6 children of Percentage least at received who 59 A high dose vitamin one last the within supplement 6 months ITNs of children Percentage sleeping months 0-59 aged insecticide an under treated mosquito net (ITN) of pregnant Percentage an under sleeping women insecticide treated mosquito net (ITN) EPI+ of children Percentage months who 12-23 aged against immunized are measles *All vaccination indicators calculated bas *All vaccination indicators calculated

IIP-JHU | Retrospective evaluation of ACSD in Ghana A65

CI 33 58 60 58 24 - 41 - 48 - 95% 95% 41 - 41 0 6 %) Miss( 4 4 ACT % 50 51 .D) .D) n 2007 MICS suppl 2007 MICS 357 28 0.9 554 53 1.6 SEE APP

(

CI 40 76 20 -20 58 -58 95% 95% DHS %) n/a n/a 145 n/a n/a 136 Miss( 4 4 % n/a n/a 2006 MICS n

20 %) Miss( % CDC No data n 431 37 1 31 30 1 367 61 0.3 38 67 3.6 367 59 0.3 38 0.3 3.6 0 - 0.8 554 9 1.6 6 - 12 2003 ACSD- DEFINITION OF PNEUMONIA DIFFERENT FROM NY ANTIMALARIAL MEDICATION CI 46 87 84 18 - 54 - 47 - 95% 95% A MICS

(2) (4) %) Miss(

the “high impact” districts, Ghana (weighted) the “high impact” districts,

] No data No % chloroquine chloroquine ACT (since 2004) 2003 DHS CASES n 43 32 0 44 71 0 44 66 0 29 [66] 0 n/a 206 35 0 20

>24 1 % 24 39 53 BASED ON ; 2002 IHNS n URVEY REPORT S

2002

CI 49 89 72 28 - 68 - 36 - 95% 95% IHNS %) Miss( RESULTS NOT SHOWN : A / No data % N [ chloroquine chloroquine

1998/1999 DHS n 32 2 035 0 - 5 39 2 60 78 0 32 54 2 25-49] INDICATORS FROM FROM INDICATORS ; Illness case management indicators over time in management indicators Illness case 3

BASED ON CASES : DATA NOT AVAILABLE IMCI case case IMCI indicators management 2 Table G2. NCLUDED TREATMENT WITH APPROPRIATE ANTIMALARIAL ACCORDING TO NATIONAL POLICY IHNS I

(1) UNWEIGHTED Percentage of children aged 0-59 months with suspected pneumonia treatedantibiotics with Percentage of children aged 0-59 months with diarrhoea receiving ORS, RHF or increased fluids and continued feeding Percentage of children aged 0-59 months with fever receivingappropriate antimalarial drugs Percentage of children aged 0-59 months with antimalarial receiving fever drugs Ghana antimalarial policy Percentage of children aged 0-59 months with suspected pneumonia taken toan appropriate health provider (3) [

A66 IIP-JHU | Retrospective evaluation of ACSD in Ghana

CI 63 92 56 64 95% 95% 42 - 77 - 47 - 46 - %) miss( n%

2002

CI 55 484 52 0.8 29- 95% IHNS %) miss( 2006 MICS² 2006 suppl² MICS 2007 n% INDICATORSFROM ; %) miss(

.D) CDC n% 2003 ACSD- SEE APP (

DATA NOT AVAILABLE DATANOT DHS IHNS

CI 97 328 45 2 28 42 0 95% 95% 73 - (1)

] %) miss( CASES >24 2003 DHS 2003 21 n/a n/a n/a 93 82 8 25 [92] 0 n/a 125 84 0 28 [43] 0 n/a 168 39 3 32 [56] 0 n/a 258 55 0 18 n/a n/a n/a 94 50 5 30 [53] 0 n/a 159 53 0 BASED ON ; 33 2002 2002 IHNS¹ 95% CI n % n % RESULTS NOT SHOWN : A / UNABLE TO CALCULATE BREASTFEEDING INDICATORS AS N %) ; [

miss( 25-49] 1998/1999 DHS 1998/1999 n% NO FULL BIRTH HISTORY Feeding behaviour indicators over time in the “high impact” districts, Ghana (weighted) impact” districts, Ghana over time in the “high indicators Feeding behaviour

BASED ON CASES : 2006:

MICS

UNWEIGHTED Indicators Percentage of Percentage 20-23 aged children months who are currently breastfeeding 13 n/a n/a n/a Percentage of Percentage 0-5 infants aged months who are exclusively breastfed 39 of Percentage 6-9 infants aged [28]months who are 3 and breastfed receive n/a complementary food 19 n/a n/a n/a Percentage of Percentage the to put newborns one breast within of hour birth 39 11 0 3-19 45 85 0 [ (2) Table G3.

IIP-JHU | Retrospective evaluation of ACSD in Ghana A67

85 82 77 - 73 - 95% CI 95% %) miss( No data No data n% CI 61 485 40 1 35-46 68 480 8257 2 479 78-86 74 67 484 2 61-72 63 1 57-68 62 481 57 1 51-62 98 47895 89 478 2 86-92 81 2 86 482 78 1 33- 40- 48- 41- 85- 95% 95% 37 - 77 - 64 - 0 1 %) miss( 47 52 No data No data 2006 MICS 2007suppl MICS n% 28 28 23 47 16 %) miss( 27 <1 51 <1 CDC No data No data birth in the previous 12 month for “high impact” districts, birth in the previous 12 month for “high impact” districts, n% 2003 ACSD- 2003 332 322 74 4 28 75 0 332 CI 78 326 74 2 28 86 0 65 281 10 16 47 320 56 4 28 61 1 98 326 82 2 28 91 0 40- 19- 58- 95% 95% 50 - 5-34 10-30 47-70 2 0 0 0 %) miss(

52 18 20 No data WOMEN WITH INSTITUTIONAL DELIVERIES ASSUMED TO HAVE APPROPRIATE POSTNATAL CARE 2003 DHS (2) 45 58 45 33 0 44 45 45 64 0 45 45 4 0 0-10 45 4 0 0-10 318 5 5 23 54 16

45 81 47 2002 2002 IHNS¹ URVEY REPORT S

75 MIDWIFE ORAUXILLIARY MIDWIFE 52 - 52 / 95% CI n % n % 2002

%) miss( NURSE IHNS , No data No data 72 0 61-83 O CTOR 1998/1999 DHS 1998/1999 D No data on # of days n% 39 17 1 7-27 39 19 0 10-27 39 63 0 46-79 39 64 1 INDICATORS FROM FROM INDICATORS ; ³

Antenatal and postnatal care over time among womenAntenatal and with a live

DATA NOT AVAILABLE KILLED HEALTH WORKER DEFINED AS IHNS S

Indicators Percentage ofnewborns receivinga postnatal a visitskilled by health worker of² within3days delivery Percentage ofbirths skilled attended by health worker³ Percentage of pregnant women receiving intermittentpreventative treatment for malaria during pregnancy SP) in previous(any year Percentage of pregnant women receiving intermittentpreventative treatment for malaria during pregnancy in previous(2+ doses) year Percentage ofnewborns protected againsttetanus (2+ doses TTduring pregnancy) Percentage ofnewborns protected fully against tetanus Percentage of pregnant women receiving 3 months of iron supplementation. Percentage ofwomen receiving vitamin A supplementation within 2 months of 39 birth Percentage of pregnant women who reportat least 4 prenatalvisits to a skilled healthworker³ Percentage of pregnant women who reportat least 3 prenatalvisits to a skilled healthworker³ 39 74 1 66-82 45 78 0 Ghana (weighted) Table G4: (1) (3)

A68 IIP-JHU | Retrospective evaluation of ACSD in Ghana

95% CI %) Miss( (weighted)

No Data 2006 MICS 2006 MICS Comparison area ¥ Comparison n% 555 50 0 44 - 56

2006 MICS/ 2007 MICS suppl. suppl. MICS 2007 MICS/ 2006 No Data HID 2007 MICS suppl MICS same as DPT U E R 2 SD “high impact” districts and the SHANTI REGIONS

95% CI95% n % A

. (%) Miss Miss CCRA AND No Data A LASTING NET - REATER REATER n% 2003 DHS 2003 G

] HID area ¥ Comparison No Data No 31 [32] 328 2 0.1 0.5 - 4 230 23 2765 3 0.4 2 - 4 2256 58 2668 24 0.9 21 - 26 CASES >24 95% CI n % rity indicators over time in AC S AND URBAN AREAS OF THE (%) MONTHS BEFORE THE SURVEY OR LONG Miss HID BASED ON ; 12 No Data No Data No Data n% EXCLUDING THE , 1998/1999 DHS 1998/1999 RESULTS NOT SHOWN :

A / N HID area ¥ Comparison [

No Data No Data No Data n% 38 6038 511 68 60 515 1 55 - 64 65 39 1 68 60 - 70 536 39 67 76 0.5 537 64 - 71 76 396 0.3 80 73 - 80 397 549 95 78 1.1 545 76 - 79 79 1.8 78 - 81 155 65 2099 22 3 17 - 26 185 86 2271 79 2 77 - 81 1975 90 2368 96 1 95 - 97 NATIONALLEVEL 25-49]

– HANA G EPI+ and ITN coverage indicators over time in “high impact” districts and comparison areas, Ghana time in “high impact” districts and EPI+ and ITN coverage indicators over H

BASED ON CASES : NSECTICIDE TREATED NET DEFINED AS TREATED WITHIN I

=

Tables presenting comparisons of prio Table H1. comparison area OMPARISON AREA IS ITN

PPENDIX C

UNWEIGHTED Indicators ITN ¹ EPI+ * Percentage of children aged 12-23 12-23 aged children of Percentage against immunized are who months measles 12-23 aged children of Percentage of 3 doses received who months vaccineDPT 12-23 aged children of Percentage against immunized are who months Hib 6 - 59 who children of Percentage receivedat least one high dose last the within A supplement vitamin 6 months 0-59 aged children of Percentage an under sleeping months insecticide treated mosquitonet (ITN) women pregnant of Percentage insecticide an under sleeping (ITN) net mosquito treated ¥ (1) A [

IIP-JHU | Retrospective evaluation of ACSD in Ghana A69

95% CI 95%

%) Miss( ACT 2006 MICS Comparison area ¥ ¥ area Comparison (weighted) n%

HID HID ACT 2006 MICS/ 2007 MICS suppl. suppl. MICS 2007 MICS/ 2006 suppl

136 51 ³ 146 35 ³ 7 27 - 43 2007 MICS 95% CI n % SHANTI REGIONS A %) Miss( No Data Data No chloroquine and comparison areas, Ghana CCRA AND A Comparison area ¥ area Comparison n% 2003 DHS REATER G HID No Data 44 66 518 64 6 - 68 60 554 9 600 3 2.8 - 1 5 chloroquine igh impact” districts igh impact” districts 95% CI95% n % AND URBAN AREAS OF %) Miss( UER .D) No Data chloroquine SEE APP NCLUDED TREATMENT WITH APPROPRIATE ANTIMALARIAL ACCORDING TO NATIONAL POLICY ( I

n% DHS EXCLUDING THE , 1998/1999 DHS 1998/1999 HID ¥ area Comparison No Data No n% 35 39 430 23 1 18 - 28 43 32 396 38 0.7 33 - 43 357 28 405 30 1.2 25 - 35 32 54 33632 22 0 2 17 - 26 318 29 20 [66] 5 276 15 - 25 40 1.1 33 - 47 145 50 ³ 158 36 ³ 0 28 - 44 60 78 637 60 0 55 - 64 44 71 518 67 6 62 - 72 554 53 602 61 2.5 55 - 68 chloroquine (2) NATIONALLEVEL – n i y HANA G ever f

ith Case management indicators over time in “h Case management indicators m w

59 - 0 DEFINITION OF PNEUMONIA DIFFERENT FROM ren ren

OMPARISON AREA IS NY ANTIMALARIAL MEDICATION C

A MICS

IMCI case management management case IMCI indicators Percentage of children aged 0-59 0-59 aged of children Percentage months diarrhoea with receiving ORS, RHF or increased fluids feeding continued and Child 0-59 aged of children Percentage months suspected with an taken to pneumonia provider health appropriate Percentage of children aged 0-59 0-59 aged of children Percentage months suspected with with treated pneumonia antibiotics Percentage of children aged 0-59 0-59 aged of children Percentage fever receiving months with (program)¹ drugs antimalarial Ghana antimalarial polic previous 2 weeks, rec'd rec'd weeks, 2 previous antimalarial appropriate treatment (effective)²

Table H2. ¥ (1) (3)

A70 IIP-JHU | Retrospective evaluation of ACSD in Ghana

95% CI

7 %) Miss( 29 26 - 33 2006 MICS Comparison area ¥ n% 4027

2006 MICS/ 2007 suppl. ¹ MICS n% HID 2007 3222 12 MICS suppl MICS UMASI K CI 95% 95% CCRA AND %) A Miss(

.D) SEE APP ( n%

2003 DHS DHS

] HID Comparison area ¥ n% impact” districts and comparison areas, Ghana (weighted) areas, impact” districts and comparison 45 85 56628 [43] 41 240 018 37 - 45 49 n/a 484 0.6 42 - 56 188 5221 258 64 n/a 527 55 0.3 131 55 - 72 31 297 159 71 1.1 51 53 26 - 35 1 62 - 80 0.2 180 44 - 58 125 57 84 0.7 49 - 65 183 60 0 51 - 68 259 6 4123 23 9 20 - 25 273 6 4441 21 1.4 19 - 23 3314 12 4311 27 0.8 24 - 31 CASES >24 CI 30 27 36 74 68 27 - 27 19 - 19 - 23 - 60 - 54 95% 95% AND THE MAJOR METROPOLITAN AREAS OF 2 %) Miss( UER BASED ON ; No data No 28 n% EXCLUDING THE , 1998/1999 DHS 16 RESULTS NOT SHOWN : A HID Comparison area ¥ / UNABLE TO CALCULATE BREASTFEEDING INDICATORS AS No data N n% ; [ 39 1139 509 23 [28] 0.1 23519 29 n/a 0 19013 67 n/a 1 165 61 0

270 5606 NATIONALLEVEL 25-49] – HANA G Feeding behaviour indicators over time in “high Feeding behaviour NO FULL BIRTH HISTORY BASED ON CASES : 2006:

OMPARISON AREA IS MICS

C

UNWEIGHTED IMCI feeding behavior IMCI behavior feeding indicators Percentage of newborns put to put ofnewborns Percentage ofbirth hour one breast within the Percentage of infants aged 0-5 months are exclusively who breastfed Percentage of infants aged 6-9 months are breastfed who and complementaryreceive food 20- aged ofPercentage children 23 months are currently who breastfeeding ofPercentage households salt consuming iodized (>=15ppm) ofPercentage households salt consuming iodized (>=15ppm) no salt) (exclude HH with (1) , Table H3. [ ¥

IIP-JHU | Retrospective evaluation of ACSD in Ghana A71

37 26 - %) CI 95% Miss( No data No data No 2006 MICS MICS 2006 Comparison area ¥ area Comparison 2006 MICS/ 2007 MICS suppl. 2007 MICS 2006 MICS/ No data No data No HID 2007 HID 2007

482 78 529 75 1 70 - 79 MICS suppl UMASI K , nurse/midwife midwifeor auxilliary %) CI 95% n % n % CCRA AND Miss( A No data No 2003 DHS “high impact” districts and comparison areas, Ghana (weighted) and “high impact” districts HID Comparison area ¥ No data No 45 7845 552 64 76 552 2 58 72 - 80 478 2 54 - 63 89 478 53245 81 80 33 532 0.3 562 76 - 84 66 47 0.3 0.6 61 - 71 43 - 5145 484 18 6345 566 20 531 3545 60 664 0 58 35 30 - 40 0.5 55 - 65 485 0.4 566 31 - 40 36 40 0 533 32 - 40 42 481 57 0 36 - 48 531 48 0.5 43 - 53 45 445 542 1 4 4 542 0 - 2 1 480 444 82 0 - 1 52 479 500 506 67 43 33 11 6 500 29 - 38 37 - 48 31 6 CI n % n % 70 59 51 43 42 27 61 - 61 - 50 - 41 - 33 - 33 - 19 95% %) AND THE MAJOR METROPOLITAN AREAS OF Miss( postnatal care (2) Skilled health worker defined as Doctor as defined worker health Skilled (2) postnatal care UER No data No data No data No data No Comparison area ¥ Comparison 1998/1999 DHS EXCLUDING THE , HID n% n% No data No No data No data No data No 39 7439 505 64 66 505 1 55 1 39 63 505 46 0.7 39 1739 508 19 3839 0.2 508 72 38 0.1 509 23 0 NATIONALLEVEL eries assumed to have appropriate appropriate to have eries assumed – HANA G Antenatal delivery and postnatal care indicators over time in Antenatal delivery and postnatal care

OMPARISON AREA IS C

ANC, assisted delivery and delivery assisted ANC, postnatal care indicators Percentage of pregnant women who report report women who of pregnant Percentage skilled a health to visits prenatal 3 least at worker ² report women who of pregnant Percentage skilled a health to visits prenatal 4 least at worker ² women receiving of pregnant Percentage treatment for preventative intermittent year previous in pregnancy during malaria SP) (any women receiving of pregnant Percentage treatment for preventative intermittent year previous in pregnancy during malaria SP) (2+ against protected of newborns Percentage doses TT pregnancy) (2+ tetanus during protected fully of newborns Percentage tetanus against 3 women receiving of pregnant Percentage months of iron supplementation. skilled by attended of births Percentage ² worker health a receiving of newborns Percentage worker a skilled health by visit postnatal ² ofdelivery¹ 3 days within vitamin A of receiving women Percentage months 2 of birth within supplementation Table H4: ¥ (1)Women with institutional deliv

A72 IIP-JHU | Retrospective evaluation of ACSD in Ghana

103 272 591 206 377 420 226 295 199 2268 5% 9% 9% 51% 1146 49% 1122 81% 1848 12% 26% 17% 19% 10% 13% Percent Total Children lete interviews p Table I1c: Under five Children with five Children Under Table I1c: com Total Female Male 3 Rural 2 6 7 8 Urban 5 1 Sex 4 Residence 3 702 674 229 171 798 294 437 317 487 771 378 406 914 488 3288 1911 2517 the HIDs by socio-demographic characteristics of 28 15 7% 5% 9% 21% 21% 58% 24% 13% 77% 10% 15% 23% 12% 11% 0.1% Percent Total Women Table I1b: Eligible women with Eligible Table I1b: complete interviews Primary Primary Secondary + Total Education None Currently pregnant Currently Yes Not sure <24 <12 6 7 8 Months since last birth 5 3 Rural 2 4 Urban 1 Residence 712 344 412 319 184 395 744 567 390 2580 3324 6% 10% 78% 22% 17% 12% 10% 12% Percent Total Households Description of households, eligible women and children under five supplemental MICS 2007-8 in the “high impact” districts, MICS 2007-8 in the “high impact” districts, five supplemental and children under households, eligible women of Description

Ghana Tables presenting 2007-8 results for key indicators in population Table I1: Table I1a: Households interviewed Households Table I1a: Total Bawku Municipality 21% Bawku West Bolgatanga Municipality 12% Talensi-Nabdam Garu-Tempane Residence Bongo Rural Urban Districts Districts Districts Districts Districts Builsa Kasena-Nankana Appendix I

IIP-JHU | Retrospective evaluation of ACSD in Ghana A73

- 87 77 72 84 79 85 188 310 209 397 23m¹ Number of children 12 ² rd, by the total total the by rd, 98 95 94 89 94 92 99 95 99 95 (%) ACSD indicator indicator (%) vaccine vaccine Innoculated against DPT before 12m 12m before (%) Rec'd vaccine vaccine - 87 99% 99% 75 92% 97% 72 92% 100% 85 99% 100% 7985 98% 99% 98% 100% 187 97% 98% 309 95% 99% 208 95% 99% 396 96% 99% 23m¹ Number of children 12 ² (%) 77% 82% 81% 79% 79% 71% 86% 86% 78% 80% ACSD indicator indicator (%) vaccine before 12m Innoculated against measles n/a n/a n/a n/a n/a n/a n/a n/a (%) Rec'd vaccine vaccine 83% 99% 77% 84% 98% 84% 84% 98% 83% 88% 85% 99% 80% 84% 99% 80% 83% 97% 73% 86% 99% 87% 83% 99% 78% 82% 97% 89% 84% 98% 82% Seen % with % with EPI card- 2 3 4 Male Rural Urban Builsa 88% n/a n/a n/a n/a n/a n/a n/a n/a Bongo 87% n/a n/a n/a n/a n/a n/a n/a n/a Female Poorest Least Poor Bawku West 80% n/a n/a n/a n/a n/a n/a n/a n/a Garu-Tempane Talensi-Nabdam 84% n/a n/a n/a n/a n/a n/a n/a n/a Total Kasena-Nankana 87% n/a n/a n/a n/a n/a n/a n/a n/a Vaccination by socio-demographic characteristics in “high-impact” districts, Ghana 2007-8 “high-impact” districts, in by socio-demographic characteristics Vaccination

Bawku Municipality 77% n/a n/a n/a n/a n/a n/a n/a n/a Bolgatanga Municipality 85% n/a n/a n/a n/a n/a n/a n/a n/a ACSDindicator: multiply the percent of children thatreceived vaccination before first birthday, according toca immunization Children 12-23 months Children 12-23 of age, still alivewith non-missing data for indicator calculation:weighted Children 12-23 months of age vaccinated against mealses and DPT Districts ² ¹ Residence percentage of children vaccinated, according to card or mother’s report. Sex Wealth index quintiles Table I2:

A74 IIP-JHU | Retrospective evaluation of ACSD in Ghana

socio-demographic characteristics in “high-impact” districts, “high-impact” districts, in socio-demographic characteristics 87 377 417 417 389 375 221 395 460 460 439 991 984 368 332 176 198 515 232 175 260 age¹ 1975 1607 6-59 months6-59 of Number of children 90% 93% 88% 91% 88% 90% 72% 90% 92% 91% 94% 90% 90% 89% 93% 83% 89% 96% 93% 96% 74% 91% 88% Vitamin A A Vitamin supplementation (%)supplementation 4 3 2 6-11 Male Rural 12-23 24-35 36-47 48-59 Urban Builsa Bongo Female Poorest Least Poor Bawku West Bawku Garu-Tempane Vitamin A supplementation (one-dose) in previous 6 months by

Talensi-Nabdam Total Kasena-Nankana Bawku MunicipalityBawku Bolgatanga Municipality Children 6-59 months of age, still6-59 alive Children withfor non-missing data Children 6-59 months of age receiving one dose months of age receiving 6-59 Children 6 months in the previous supplementation A vitamin ¹ indicator calculation: weighted Wealth index quintiles Age in months Age Sex Residence Districts Districts Table I3: Ghana 2007-8

IIP-JHU | Retrospective evaluation of ACSD in Ghana A75

102 196 268 225 589 205 376 294 418 483 399 463 507 405 434 452 473 469 428 age² 1838 1119 1137 2256 months of Number of children 0-59 0-59 children (%) 76% 65% 54% 56% 45% 72% 53% 72% 59% 53% 57% 62% 67% 63% 49% 48% 58% 56% 58% 58% 57% 59% 58% ITN* last night Slept under an an under Slept 76% 58% 54% 62% 55% 60% 47% 76% 69% 64% 61% 71% 76% 60% 61% 61% 61% 60% 58% 61% 66% 53% 50% last night (%) Sleptever- under an treated mosquito net Percentage of children who: children of Percentage 62% 55% 61% 70% 65% 61% 61% 61% 62% 61% 61% 61% night (%) Slept under mosquito net last last net mosquito 2 3 4 0-11 Male Rural 12-23 24-3536-4748-59 67% 54% 51% Urban Builsa 78% Bongo 76% Female Poorest Least poor Bawku West 58% Utilization of bednets by children aged 0-59 months by socio-demographic characteristics in “high-impact” districts, Ghana districts, in “high-impact” characteristics 0-59 months by socio-demographic by children aged bednets Utilization of Garu-Tempane 55% Talensi-Nabdam 77% Kasena-Nankana 59% Bawku Municipality 47% Bolgatanga Municipality 71% ITN=Mosquito net treated with insecticide in the previous 12 months,or along-lasting net Total children under five who slept inlast HH night,with non-missing dataindicator for calculation: Childrenaged 0-59 months sleeping undera mosquito net, treated a net or an ITN¹ thenight preceding the survey Residence Districts Sex ¹ Age in months months in Age quintiles index Wealth Total ² weighted 2007-8

Table I4:

A76 IIP-JHU | Retrospective evaluation of ACSD in Ghana 67 58 29 66 39 79 93 74 71 n/a n/a n/a n/a n/a n/a n/a n/a 357 299 195 183 174 59m with diarrhoea¹ children 0- children Number of

n/a n/a n/a n/a n/a n/a n/a n/a 28% 32% 32% 27% 32% 30% 28% 23% 34% 19% [14%] [28%] [26%] [32%] ORS/ RHF/ feeding (%) with continued continued with increased fluids fluids (%) fluids increased increased ORS/ RHF/ oea inprevious weeks 2 ORS + ORS RHF (%) (%) ORS ORS Children 0-59 with diarrh 59m* children 0- Number of 9% 302 [14%] [21%] [50%] 17% 377 n/a n/a n/a 16% 429 42% 43% 73% 23% 295 n/a n/a n/a 14% 420 33% 37% 69% 12% 272 n/a n/a n/a 16% 1848 42% 48% 75% 10% 103 n/a n/a n/a 16% 199 n/a n/a n/a 22% 185 [41%] [41%] [69%] 14% 226 n/a n/a n/a 14% 474 39% 51% 74% 22% 399 [41%] [46%] [74%] 17% 473 56% 62% 79% 15% 1382 [47%] [54%] [81%] 14% 591 n/a n/a n/a 16% 457 29% 36% 77% 17%15% 1122 1146 43% 39% 48% 45% 74% 74% 19% 206 n/a n/a n/a 16% 435 37% 40% 68% % with % with diarrhea /a 28 31 31 64 31 27 80 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n 145 16% 2268 41% 47% 74% igh-impact” districts, Ghana 2007-8 districts, Ghana igh-impact” 59m with children 0- children Number of pneumonia¹ /a n with 48% % treated % treated anitiobics n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a 50% 51% 71% 65% 55% 60% 45% 48% 46% 54% 39% 50% 41% 35% 53% health health facility % taken to to % taken hs with illness in the previous 2 weekshs within the previous illness 59m* children 0- Number of 6% 2266 7% 295 6% 420 7% 429 6% 271 6% 1846 6% 103 8% 199 5% 302 9% 183 4% 226 7% 399 6% 473 6% 1381 5% 591 7% 473 6% 1121 7% 455 5% 435 7% 1145 8% 376 demographic characteristics in “h characteristics demographic 10% 206 % with % with suspected suspected pneumonia 87 89 45 15 49 29 55 48 98 61 554 465 112 119 358 126 116 280 119 103 274 123 fever¹ 59m with children 0- Number of 53% 53% 67% 74% 51% 66% 85% 34% 55% 37% 63% 53% 57% 55% 47% 53% 59% 47% 47% 48% 59% 42% anti- onia, for children 0-59mont and diarrhea malarial % given any 59m* children 0- Number of 9% 302 25% 2265 22% 419 26% 1846 23% 428 25% 473 24% 473 27% 455 24% 435 fever % with % with Children 0-59 with fever in previous 2 weeks Children 0-59 with suspected pneumonia in previous 2 weeks 4 3 2 Illness case management by socio- Illness case 0-5 6-11 31% 184 male 25% 1120 Rural 12-23 29% 399 24-59 26% 1380 Urban Builsa 30% 295 Bongo 16% 102 female 24% 1145 Poorest ive with non-missing data for indicator calculation: weighted; n/a - small sample >25 cases;size [unweighted] - >50 cases in at least one cell f Least Poor Bawku West 22% 226 Garu-Tempane 33% 376 Talensi-Nabdam 30% 206 Table I5: Kasena-Nankana 17% 272 Bawku Municipality 22% 591 Bolgatanga Municipality 25% 198 Care management of fever, suspected pneum Total ¹ Children under Districts Residence Age in months in months Age Gender Wealth index quintiles

IIP-JHU | Retrospective evaluation of ACSD in Ghana A77

a

44 518 602 554 weeks² fever in last two No. of children with children of No. & 2007-8) 61% 53% ny AM AM ny A treatment 3% 9% AM¹ 64% 67% 66% 71% Appropriate Appropriate CT/ CT/ 0% 9% 0% 3% A CoArtem 1% 1% 3% 6% 2% 2% 9% 15% Data not available not Data available not Data preceding the survey in “high-impact” districts and comparison areas over time, the survey in “high-impact” districts Ghan preceding 64% 66% 43% 28% Children with a fever in the last two weeks who were treated with: 1% 0% 0% 1% SP/ SP/ Fansidar Chloroquine Amodiaquine Quinine Treatments given for fever in the 2 weeks Children under five with non-missing data for indicator calculation: weighted DHS 1998/1999 National comparison High Impact districts DHS 2003 National comparison High Impact districts MICS 2007/2008 High Impact districts NOTE: Anti-malarial treatment columns are not mutually exclusive ² ¹ Appropriate antimalarial treatment defined as Mali policy for first line malaria treatment (CQ in 1998-9 & 2003; ACT in 2006 MICS 2006 National comparison

Table I6: 2007-8

A78 IIP-JHU | Retrospective evaluation of ACSD in Ghana

n/a n/a 279 145 336 158 Number of pneumonia¹ 59 months with months 59 children aged 0- n/a n/a 55% 31% 44% 36% / neighbors Not / treated treated at home n/a n/a 4% 1% 1% 2% the 2 weeks preceding the survey in “high-impact” districts survey in “high-impact” districts preceding the the 2 weeks n/a n/a 0% 0% 20% 23% n/a n/a 0% 0% 28% 12% vendorworker Health Village Other Private sector drug n/a n/a 5% 6% 5% 3% Children with suspected pneumonia in last the two weeks who were to: taken Private Health center /center facility n/a n/a 16% 34% 28% 46% Public health health Public center / facility center Locations where care was sought for suspected pneumonia in sought for suspected pneumonia care was Locations where Children under five with non-missing data for indicator calculation: weighted DHS 1998/1999 National comparison High Impact districts DHS 2003 National comparison High Impact districts MICS 2006 National comparison MICS 2007/2008 High Impact districts Note: Mutally exclusive of in order table ¹ Table I7: and comparison areas over time, Ghana 2007-8 and comparison areas over

IIP-JHU | Retrospective evaluation of ACSD in Ghana A79

73 47 27 93 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a 125 23m² 23m² Number of Number children 20- children teristics in “high-impact” in “high-impact” teristics ] n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a 84% [85% [83%] [74%] [87%] Continued Continued breastfeeding breastfeeding 33 79 34 80 34 26 27 29 n/a n/a n/a n/a n/a n/a n/a n/a 131 159 9m² 9m² Number of Number children 6------n/a n/a n/a n/a n/a n/a n/a n/a 53% 52% 53% [49%] [62%] [50%] [58%] [52%] [67%] [38%] feeding feeding mentary mentary Comple- thers by socio-demographic charac 58 62 46 31 30 32 24 37 38 32 36 44 52 42 144 115 216 120 258 138 0-5m² 0-5m² children children Number of Number among children 0-5 months, complementary feeding among % n/a 55% 54% 80% 33 55% [55%] [45%] [52%] [58%] [73%] [72%] [49%] [37%] [53%] [53%] [75%] [54%] [62%] [79%] breastfeed Exclusively Exclusively 93 95 60 46 61 42 62 73 56 82 94 91 97 108 390 228 256 484 12m¹ previous previous Birth within Birth within -- -- 50% 44% 57% 42% 54% 47% 60% 56% 47% 52% [57%] [50%] [66%] [43%] [45%] [53%] [41%] [55%] Timely initiation of initiation of breastfeeding 2 3 4 0-2 3-5 male Rural Urban female Poorest Least Poor Prevalence of infant feeding behaviours as reported by mo behaviours as reported Prevalence of infant feeding ealth index quintiles Women with a live birth in previous 12 months with non-missing data for indicator analysis: weighted Children with non-missing data for indicator analysis: weighted Timelyinitiation of breastfeeding,exclusive breastfeeding W Districts Builsa Kasena-Nankana Bongo Bolgatanga Municipality West Bawku Bawku MunicipalityBawku Talensi-Nabdam Garu-Tempane Residence Sex Sex Age in months in months Age Total ¹ ² n/a>25 cases; - small size sample [unweighted] - >50 cases in at least one cell districts, Ghana 2007-8 districts, Table I8:

A80 IIP-JHU | Retrospective evaluation of ACSD in Ghana

61 60 42 62 83 75 73 46 94 62 92 92 92 96 478 385 224 254 313 103 107 Birth Birth 12m¹ within previous 85% 81% 80% 84% 79% 79% 83% 89% 75% 75% 80% 83% 93% [74%] [77%] [83%] [89%] [83%] [75%] [73%] [94%] worker 3+ 4+ [84%] [87%] [86%] [94%] [93%] [84%] [83%] [100%] trained health care care health trained Prenatal visits witha the previous 12 months 61 47 61 42 62 72 56 81 12m¹ previous Birth within within Birth 83% 94 94% 78% 484 89% 76% 390 88% 75% 228 92% 80% 256 87% 74% 104 92% 77% 318 88% 84%71% 62 93 94% 84% 70% 107 84% 85% 93 92% 80% 93 91% 83% 97 96% [66%] [89%] [91%] [84%] [90%] [68%] [75%] [51%] protection TT2 TT Full Neonatal tetanus 68% 63% 62% 65% 61% 61% 63% 64% 53% 56% 70% 68% 66% [59%] [72%] [76%] [68%] [74%] [51%] [67%] [34%] n who have given birth in birth given n who have 54 46 41 62 72 81 92 60 61 60 91 92 93 96 487 387 227 252 103 316 107 Birth Birth 12m¹ within previous 70% 67% 66% 63% 70% 63% 67% 70% 56% 67% 69% 69% 71% [59%] [74%] [73%] [58%] [68%] [67%] [53%] [69%] % IPT during IPT pregnancy pregnancy 1+ 2+ 84% 81% 76% 83% 84% 73% 81% 89% 86% 82 79% [82%] [85%] [69%[ [82%] [80%] [89%] [60%] n previous 12 months with non-missing data for indicator analysis: weighted 2 3 4 h i 0-5 79% 6-11 84% Rural None Urban Poorest Least Poor Primary School Antenatal care indicators among women giving birth within the in previous 12 months by indicators soci o-demographic characteristics Antenatal care

Secondary school+ Secondary Antenatal care (including IPT, TT, IPT, (including wome among Fe) care Antenatal Kasena-Nankana MunicipalityBolgatanga WestBawku Bawku Municipality [93%] Talensi-Nabdam Garu-Tempane Residence Bongo Districts Districts Builsa n/a - small sample >25 cases;size [unweighted] - >50 cases in at least one cell ¹ Women with a live birt Months since birth Months since Mother's education level Mother's education Wealth index quintiles Total “high-impact” districts, Ghana 2007-8 “high-impact” districts, Table I9:

IIP-JHU | Retrospective evaluation of ACSD in Ghana A81

birth in the previous 12 months by socio-demographic birth in the previous

95 94 92 60 93 92 82 60 72 41 56 61 481 108 104 389 318 226 255 12m² previous Birth within within Birth 60% 57% 62% 58% 63% 52% 51% 60% 66% 54% 54% 57% 56% [51%] [60%] [58%] [63%] [43%] [63%] Postnatal Postnatal with Vitamin A supplementation 12m² previous Birth within within Birth ¹ 77% 97 40% 485 29% 95 46% 93 65% 62 23% 93 27% 107 38% 104 71%33% 94 392 36% 319 attendant Skilled birth Skilled 3 4 2 0-5 46% 230 6-11 35% 256 Rural None Urban Builsa [36%] 61 [55%] 60 Bongo [46%] 61 Poorest Least Poor Assisted delivery and post-natal among womenAssisted delivery care giving

Bawku West [37%] 62 Garu-Tempane [33%] 81 Primary School Talensi-Nabdam [34%] 56 Kasena-Nankana [45%] 47 [55%] 47 Bawku MunicipalityBawku [43%] 73 Secondary school+ Bolgatanga Municipality [55%] 42 Women with a live birth in previous 12 months with non-missing data for indicator analysis: Trained health care worker: doctor, nurse/midwife or auxilliary midwife Delivery and postnatal care indicators among womenDelivery who have birth in given Total ² weighted n/a - small sample>25 cases; size [unweighted] - >50 cases in at least one cell Wealth indexquintiles Months since birth Mother's education level Residence Districts ¹ Table I10: characteristics in “high-impact” districts, Ghana 2007-8 districts, Ghana in “high-impact” characteristics

A82 IIP-JHU | Retrospective evaluation of ACSD in Ghana

45 39 566 485 533 508 Birth within previous 12m³ previous No 63% 35% 25% 21% 12% 15% assistance² 40% 20% 37% 24% 47% 71% attendant¹ Traditional birth Traditional

y 0% 0% 0% 1% 0% 0% health worker Communit Delivery assistedDelivery by: 1% 2% 4% 0% 3% 17% Aux. midwife 29% 36% 11% 37% 17% 13% Doctor Nurse/midwife Health providers assisting deliveries in “high-impact” districts and comparison areas over time, Ghana 2007-8 and comparison areas over in “high-impact” districts deliveries Health providers assisting

Chart mutually exclusive in order of doctor to no assistance Trainied oruntrained TBA assisted assistanceby or friend/relative No Women withbirtha in previouswith12m non-missing datafor indicatorcalculation: weighted DHS 2003 comparisonNational MICS 2006 4% comparisonNational 4% Note: ¹ ² ³ High Impact districts Impact High 2% MICS 2007/2008 districts Impact High 2% DHS 1998/1999 comparisonNational 4% High Impact districts Impact High 1% Table I11:

IIP-JHU | Retrospective evaluation of ACSD in Ghana A83

APPENDIX J Additional tables for nutrition

Figure J1: Protocol for inclusion and exclusion of cases for nutrition analyses in “high-impact” districts and national comparison as measured in 1998 DHS, Ghana

Excluded: 1998 Excluded: National High impact comparison Total Number of children districts under five

n=2612 n=199

Incomplete 11% 13% result

Wt/Ht Wt/Ht 4% outliers 3% outliers

5% Unknown DOB 0.6%

Ht/age Ht/age 4% outliers 5% outliers

Wt/age Wt/age 1% outliers 2% outliers

A84 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Figure J2: Protocol for inclusion and exclusion of cases for nutrition analyses in “high- impact”districts and national comparison as measured in 2003 DHS, Ghana 2003 Excluded: Excluded: National High impact comparison Total Number of children districts under five

n=2834 n=241

Did not sleep in household last 4% 3% night

Non-biological 12% children 15%

4% Incomplete 25% result

Wt/Ht Wt/Ht 4% outliers 7% outliers

<0.1% Unknown DOB 0%

Ht/age Ht/age 3% outliers 3% outliers Wt/age Wt/age 1% outliers 1% outliers

IIP-JHU | Retrospective evaluation of ACSD in Ghana A85

Figure J3: Protocol for inclusion and exclusion of cases for nutrition analyses in “high-impact” districts and national comparison as measured in 2006 & 2007 MICS, Ghana

Excluded: Excluded: National 2006/2007 High impact comparison Total Number of children districts under five n= 2,606 n=2,268

1% Incomplete 1% result

Wt/Ht Wt/Ht 2% outliers 1% outliers

Unknown DOB 5% 0.5%

Ht/age Ht/age 3% outliers 2% outliers

Wt/age Wt/age 0.5% outliers 0.4% outliers

A86 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table J4: Prevalence of stunting among children 0-59 months of age by sub-groups of the population in the “high impact” districts and comparison in the 2006 and 2007 MICS, Ghana 2006/2007 MICS High Impact Districts Geographic comparison % severely % severely % stunting (< stunting % stunting stunting (< - -2 SD) (< -3 SD) n (< -2 SD) 3 SD) n Region Upper East 29% 9% 2192 Western 28% 9% 324 Central 34% 9% 262 Greater Accra 15% 3% 60 Volta 23% 8% 225 Eastern 30% 11% 425 Ashanti 38% 15% 232 Brong Ahafo 30% 9% 285 Northern 36% 15% 523 Upper West 28% 9% 94 Residence Urban 26% 7% 403 19% 6% 550 Rural 30% 10% 1789 35% 12% 1879 Sex Male 33% 11% 1089 33% 12% 1255 Female 26% 8% 1103 29% 10% 1175 Age 0-11 11% 6% 469 11% 3% 518 12-23 31% 10% 387 37% 12% 527 24-35 37% 10% 455 42% 16% 478 36-47 35% 12% 454 37% 13% 474 48-59 32% 9% 427 30% 10% 432 Wealth index quintiles Poorest 32% 11% 719 44% 19% 219 2 28% 10% 640 37% 17% 266 3 30% 9% 449 34% 13% 584 4 31% 7% 190 34% 10% 747 Least Poor 19% 7% 194 17% 5% 614

IIP-JHU | Retrospective evaluation of ACSD in Ghana A87

Table J5: Prevalence of wasting among children 0-59 months of age by sub-groups of the population in the “high impact” districts and comparison in the 2006 and 2007 MICS, Ghana 2006/2007 MICS High Impact Districts Geographic comparison % severely % severely % wasting (< wasting (< - % wasting (< wasting (< - -2 SD) 3 SD) n -2 SD) 3 SD) n Region Upper East 8% 2% 2226 Western 6% 1% 342 Central 4% 1% 294 Greater Accra 0% 0% 60 Volta 6% 2% 244 Eastern 4% 1% 447 Ashanti 4% 1% 243 Brong Ahafo 3% 1% 306 Northern 9% 3% 566 Upper West 8% 2% 103 Residence Urban 9% 1% 409 5% 1% 579 Rural 8% 2% 1817 6% 2% 2023 Sex Male 9% 3% 1103 6% 1% 1335 Female 8% 2% 1123 5% 2% 1268 Age 0-11 14% 4% 470 11% 3% 539 12-23 14% 4% 395 7% 1% 536 24-35 9% 2% 462 5% 2% 497 36-47 1% 0% 461 2% 0% 525 48-59 4% 0% 439 3% 1% 505 Wealth index quintiles Poorest 11% 3% 729 8% 3% 250 2 8% 2% 653 9% 3% 292 3 7% 3% 452 6% 1% 625 4 5% 1% 193 5% 1% 793 Least Poor 7% 0% 199 3% 1% 642

A88 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table J6: Prevalence of underweight among children 0-59 months of age by sub-groups of the population in the “high impact” districts and comparison in the 2006 and 2007 MICS, Ghana 2006/2007 MICS

Geographic comparison High Impact Districts area % % severely % severely underweight underweight % underweight underweight (< -2 SD) (< -3 SD) n (< -2 SD) (< -3 SD) n Region Upper East 21% 5% 2230 Western 11% 1% 330 Central 13% 2% 270 Greater Accra 5% 0% 61 Volta 14% 6% 235 Eastern 13% 4% 440 Ashanti 19% 5% 241 Brong Ahafo 11% 3% 291 Northern 21% 8% 539 Upper West 16% 4% 97 Residence Urban 17% 2% 407 10% 2% 564 Rural 21% 5% 1823 17% 5% 1940 Sex Male 23% 6% 1102 17% 5% 1292 Female 19% 4% 1128 13% 4% 1213 Age 0-11 16% 4% 478 13% 4% 552 12-23 24% 7% 396 18% 4% 531 24-35 26% 7% 460 18% 6% 489 36-47 18% 4% 459 13% 4% 487 48-59 19% 3% 437 13% 3% 445 Wealth index quintiles Poorest 22% 6% 734 24% 10% 229 2 23% 6% 653 20% 8% 275 3 20% 4% 453 16% 4% 605 4 15% 2% 191 15% 4% 768 Least Poor 13% 2% 200 9% 1% 627

IIP-JHU | Retrospective evaluation of ACSD in Ghana A89

APPENDIX K Methodological challenges

This section discusses the methodological challenges of the evaluation design. Many of the weaknesses are due to the retrospective nature of the evaluation, which necessitates relying on existing—even if imperfect—data and information. The drawbacks of retrospective evaluations have been explained elsewhere.(39) We first discuss general methodological considerations, and then describe challenges in measuring levels of coverage for each ACSD implementation package. Complementing this section, appendix F provides descriptions of surveys included in the evaluation and appendix E provides a side-by side comparison of the questions utilized for indicator calculation for each survey.

Challenges in documentation. Due to the retrospective nature of the evaluation, it was difficult to collect complete and standardized information on ACSD implementation activities and other health activities in the HIDs. The collaborative nature of ACSD makes it difficult to distinguish which activities were: 1) carried out as part of the ACSD program, 2) carried out with only partial technical and/or financial support from the ACSD program, or 3) carried out by ACSD partners, but independent of the ACSD program. This was especially difficult in Ghana due to the large number of NGOs and governmental programs in the Upper East region. Primary data sources pertaining to ACSD activities taking place in 2003 and earlier were less available than more recent documents; where necessary, we relied on summative reports and presentations for this information. Sometimes, although not often, information in one document conflicted with information found in other sources. In these cases, we present the information found in the most primary source. The evaluation team has collected and reviewed the available information pertaining to ACSD implementation and contextual factors to ensure the most complete documentation and interpretation possible. However, some uncertainty and gaps in information will be inevitable.

Challenges in utilization of existing surveys. One challenge was to establish a baseline using preexisting data. The 1998 DHS occurred several years before ACSD implementation began. It is difficult to know if any differences in the 1998 DHS compared to the endline surveys are due to changes during the ACSD period or before. The 2003 DHS survey occurred towards the beginning of ACSD implementation. Many packages were rolled out after the 2003 survey but several began before such as logistic EPI+ support and ITN campaigns. Knowing the limitations, we focused on the 1998 DHS while also examining the 2003 DHS to get a full picture. The 2003 ACSD survey estimates were also considered but given less importance due to data quality issues. We were unable to obtain accompanying documentation for the 2002 IHNS survey in the Upper East region and could not perform the analysis with confidence.

The 1998 and 2003 DHS had limited sample sizes for calculation of baseline coverage indicators in the HIDs, especially those indicators measured among small subgroups of the sample such as exclusive breastfeeding or careseeking for pneumonia. These small sample sizes affect the precision of point estimates and the statistical power to detect small differences over time.

The second major challenge was comparing the baseline DHS surveys to the endline MICS surveys. The DHS and MICS use slightly different methodologies to collect data. DHS ask only biological mothers of young children about intervention coverage, while MICS questions caretakers of children, even if not biologically related, about intervention coverage. Also differences in the conduct of the survey, the DHS and MICS questionnaires and interviewers’ style of asking questions may have introduced some bias into the comparison of coverage levels between 1998-9, 2003 and 2006-7. Appendices D and E note differences in the DHS and MICS questions used for indicator calculations; appendices F review the differences between the surveys. The major differences were in breastfeeding indicators and definition of pneumonia cases. For infant feeding, the DHS (2003) only collects data on a woman’s youngest children whereas the MICS collects data on all under-five children. The DHS defines suspected pneumonia cases as “cough” plus “difficulty breathing” whereas the MICS also includes “difficulty breathing due to problem

A90 IIP-JHU | Retrospective evaluation of ACSD in Ghana

with the chest”. However, these differences were minimal and we would not expect them to affect the findings.

The 2006 MICS, used for endline coverage of the comparison area, occurred almost one year before the 2007 MICS in the Upper East region. We compared estimates of coverage between 2006 and 2007 in the HIDs to assess if the one-year time lag could have influenced our results. Most coverage indicators remained relatively stable in the HIDs between 2006 and 2007, and were not statistically significant. ITNs for children and IPTp were significantly greater in 2007 as compared to 2006 in the HIDs; coverage with any antimalarial for fever was significantly less in 2007. For these three indicators, we reran statistical tests using the 2006 MICS as our endline estimate to identify any possible bias introduced by using the 2007 MICS survey only in the HIDs. Statistical inferences were the same for trends over time and differences in changes over time in the HIDs and comparison area.

During the MICS 2007 supplemental survey, there was extensive flooding throughout the northern regions of Ghana, including the HIDs. In order to assess the impact of the flood, we prepared an additional module to the Household questionnaire. The module aimed to measure the severity of the flood on the household focusing on damage and migration. Twenty-eight percent of the households in the HIDs reported being affected by the flood and 24 percent reported damage. We selected several coverage indicators that might be differentially affect by the flood: diarrhea, pneumonia and fever case management and ITN use in underfive children. There is a significantly higher proportion of children receiving an antimalarial for fever in the households affected by the flood. (50% vs. 61%; P=0.02). For the remaining indicators tested, there is no significant difference between households affected by the flood and those not affected.

Challenges in measuring vaccination and vitamin A coverage. Baseline vaccination coverage estimates are based on very small numbers which may affect precision. In the MICS 2007, there was some confusion about the timing of vitamin A campaigns and how it was presented to the survey respondents. We discussed the issue with UNICEF-Ghana and the survey teams; they recommended that a positive response to any campaign where vitamin A was distributed be counted as the child receiving vitamin A in the previous six months.

Challenges in measuring ITN coverage. The 1998 DHS did not collect data on ITN use. The 2003 DHS collected data on ITN use from the head of household in the Household questionnaire. While the MICS 2006 and 2007 collected ITN use data from the caretaker in the under-five questionnaire. Comparability could be an issue because the caretaker might have more accurate information on childcare than the head of household. However we expect the effect to be very small. The MICS 2006 and 2007 did not collect data on whether pregnant women slept under an ITN last night.

Challenges in measuring case management and feeding practices. The 1998 and 2003 DHS surveys contained inadequate sample size (less than 25 cases) in the HIDs to determine coverage of complementary and continued feeding. We analyzed exclusive breastfeeding at baseline but the sample size is very small, less than 50 cases. As previously mentioned the DHS and MICS collect infant feeding data on slightly dissimilar populations: youngest children versus all children. However, we do not believe this difference impacts the inference.

Again, the two baseline DHS surveys had a small sample size for the illness case management indicators. The 1998 DHS only collected data on the type of fever treatment, not specific anti-malaria drugs. Therefore, we also included the proxy indicator of “any anti-malarial treatment of fever” for all surveys. Specific anti-malarial treatment was available in the other surveys. The 2003 DHS did not collect data on antibiotic treatment of suspected pneumonia. As previously mentioned the MICS and DHS questionnaires use different pneumonia definitions. The MICS defines suspected pneumonia as “cough” plus “difficulty breathing” plus “problem in the chest” and if the child does not fit all criteria then the subsequent pneumonia questions are skipped.

IIP-JHU | Retrospective evaluation of ACSD in Ghana A91

Challenges in measuring antenatal, delivery and postnatal care.

The 1998 DHS collected limited data on antenatal and postnatal care compared to the other surveys. There was no data available for IPTp, full tetanus protection and length of iron supplementation. The 2003 DHS did not collect data on full tetanus protection and both baseline surveys had small sample sizes in the HIDs. The MICS 2006 and 2007 did not collect data on iron supplementation or postnatal care.

Challenges in measuring nutrition.

Both baseline surveys had very low sample sizes for anthropomorphic measures in the HIDs. The 1998 DHS and the MICS surveys collected nutritional data in the womens and under-five questionnaires, respectively. The 2003 DHS collected data in the household questionnaire. Following DHS protocol, we excluded children whose biological mothers were not in the household listing and those who did not sleep in the household last night for 2003. This insures that only the children from selected households are measured. We could not follow this protocol for the 1998 DHS and the MICS surveys due to questionnaire structure. However, inclusion of the excluded children in the 2003 DHS did not greatly affect the nutritional estimates, changing them by less than a percentage point.

Challenges in measuring mortality and data quality assessment.

The aim of this section is to provide more detail on child mortality data in Ghana “high-impact” areas, particularly as to the data quality and its likely impact on the estimates documented in the main report.

Figure 22 in the main part of the Ghana report shows mortality change by year for the “high-impact” (HID) and national comparison area. While mortality appears to have declined in the HIDs from 1997 to 2006, the comparison area is flat over the period 1997 to 2003 and projected as flat for the period since 2003. With this large degree of uncertainty in mortality change in the national comparison area during ACSD implementation, comparison between the two areas is problematic, and made more so by the large 95% confidence limits around the HIDs yearly estimates.

Nevertheless, it is still necessary to assess mortality data quality for the HIDs. Is the nearly 20% U5MR decline an actual decline or is it the result of non-sampling errors. A first step is to focus on the elements included in table K1. This table is extensively used in DHS final reports to provide an assessment of data quality (see for example the Ghana DHS 2003 report, page 284). The table naturally divides into three parts.

The first part, on number of births, is used to identify any unexpected peaks or dips in the number of living, dead or total births, and the right-most set of three columns in the table, headed Calendar year ratio helps more easily identify these variations. If the number of births changed in the same direction by the same amount each year, the value in these last three columns would be 100. The wider the divergence from this smooth change in the number of births, the larger the divergence from 100. Table K1 shows a wide variation around 100 – from 76 to 152.

Despite the detail provided by these data, a chart can provide a clearer picture. Figure K1 shows the number of births by year from table A1 and highlights an issue that has become a common occurrence in DHS – the shift of births from the 5-year period immediately before the survey data collection, to the previous 5-year period.

A92 IIP-JHU | Retrospective evaluation of ACSD in Ghana

Table K1: Births by calendar year in Ghana "high impact" districts. Upper East region 2007

Number of births Percentage with complete birth date¹ Sex ratio at birth² Calendar year ratio³

Calendar year Living Dead Total Living Dead Total Living Dead Total Living Dead Total 2007 423 10 433 100 100 100 89 94.3 89.1 - - - 2006 397 29 426 100 100 100 109.3 127 110.4 - - - 2005 443 32 475 100 92.2 99.5 95.7 70.3 93.7 104.6 100.6 104.3 2004 451 35 485 99.8 92.2 99.3 93.2 98.6 93.6 106.2 75.9 103.3 2003 405 59 464 99.1 86.6 97.5 115.1 121 115.8 100.4 152.1 104.9 2002 357 43 400 98.9 95 98.5 116.4 124 117.2 81 77.1 80.6 2001 477 52 529 98.8 98.7 98.8 125.6 181.9 130.2 122.2 112.8 121.2 2000 423 50 473 98.4 84.3 96.9 96.2 131 99.3 93.3 91.1 93.1 1999 430 57 487 98.2 88.4 97 116.6 203 124.1 101.2 115.1 102.7 1998 426 50 476 96.9 85.6 95.7 122 179.6 126.9 103.6 90.9 102.2 2003-2007 2,120 165 2,284 99.8 92 99.2 99.6 103.7 99.9 - - - 1998-2002 2,113 252 2,365 98.2 90.3 97.4 115.1 162.6 119.3 - - - 1993-1997 1,765 285 2,050 96.8 86.5 95.4 93.4 118 96.5 - - - 1988-1992 1,346 270 1,617 95.3 83.4 93.3 113.2 111.6 112.9 - - - <= 1991 1,060 345 1,404 92 80.2 89.1 109.5 121.7 112.3 - - - All 8,403 1,317 9,720 97.1 85.6 95.5 105.2 122.9 107.5 - - - 1 Both year and month of birth given 2 (Bm/Bf)x100, where Bm and Bf are the numbers of male and female births, respectively 3 [2Bx/(Bx-1+Bx+1)]x100, where Bx is the number of births in calendar year x

Figure K1: Births by calendar year in “high-impact” districts, Upper East region 2007

600

500

400

300 Births

Asked health questions for all children with birth in 2002 or later 200

Living Dead Total 100

0 Jul 1998 - Dec 2001 Jan. 2004 - Jul. 2007 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Year

The primary cause of this shift of births has been ascribed to interviewers pushing births outside a period where they have to ask many detailed questions about a child. For the MICS 2007 supplemental, this period applied to any child born after 1 January 2002. The dip in births for 2002 is evident in figure A1, as is the peak in 2001. The result of this can be a shift in mortality between the two 5-year periods used for reporting U5MR by DHS. In general this appears to lead to a decrease in mortality for the 5-year period immediately before data collection, and an increase in mortality for the preceding 5-year period – leading to an estimated faster decline in mortality than is actually occurring.

IIP-JHU | Retrospective evaluation of ACSD in Ghana A93

Unfortunately, while the endline period (January 2004 to July 2007) generates a smoothed average of the mortality, the baseline period (July 1998 to December 2001) does not. The reason for this latter situation is that the baseline period includes the peak of births in 2001, but not the dip in 2002.

However, the impact on child mortality of this unbalanced baseline period can be assessed by including the annual estimates of U5MR. This is done in figure K2, where it can be seen that there is no significant impact on the U5MR estimates by year. Furthermore, the baseline period provides a balanced average across the small humps and dips of the annual U5MR.

The second part of table K1 contains the three sets of columns headed Percentage with complete birth date. This shows that births with a complete birth date vary from 100% down to 84% over the ten-year period from 1998 to 2007. Not having a complete birth date (month and year) increases the uncertainty of the mortality estimates and hence one would like to have close to 100% of births with complete birth dates. Respondents in Ghana and other countries in West Africa have difficulty in providing complete birth dates, as can be see from a review of the comparable table in DHS reports in Benin and Senegal. At the same, Ghana is not the worst of countries in West Africa in providing a complete birth date.

Figure K2: Births and U5MR by year in “high-impact” districts, Upper East region 2007

600 200

180 500 160

140 400 120

300 100 Births 80 200 60 Births

Living Dead Total U5MR 40 U5MR (deaths per 1000 live births) 100 20

0 Jul 1998 - Dec 2001 Jan. 2004 - Jul. 2007 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Year

However, month is the major missing part of the birth date. The implication is that mortality estimates for multiple year periods should reduce the impact of missing month in the birth date.

The third part of table K1 is the three columns headed Sex ratio at birth. These are used to check, using the last row of table, that the sex ratio of total births is around 105, as generally more males than females are born. The sex ratio for those that have died should also be larger than the sex ratio for total births since in general more males die than females. In addition, the table is used to assess variability by year. In the latter case, there is a dip around 2005, to 70, and three noticeable peaks at 2001, 1999 and 1998 of 180 to 200. This suggests that the shift in births noted in figure K1 may also be associated with a differential shift with respect to sex, and particularly in terms of deaths.

A94 IIP-JHU | Retrospective evaluation of ACSD in Ghana

However, table K2 shows that the periods used for calculating mortality (as delineated in figure A1) in general provide an averaging of births and deaths data across the low and high sex-ratios. At the same time the value of 171 for the ratio for those who have died in the baseline period is surprising high and suggests that some female deaths may have been missed. Other than for this one exception, the periods used in the ACSD evaluation in Ghana for estimating endline and baseline mortality, reduce the impact of these sex-ratio variations.

Table K2: Sex ratios at birth by year in “high-impact” districts, Upper East region 2007

Sex ratio at birth*

Calendar year Living Dead Total Sex ratio, multi-years 2007 89 94.3 89.1 LDT 2006 109.3 127 110.4 96.3 95.1 96.2 2005 95.7 70.3 93.7 2004 93.2 98.6 93.6 2003 115.1 121 115.8 2002 116.4 124 117.2 2001 125.6 181.9 130.2 2000 96.2 131 99.3 113.9 171.2 118.8 1999 116.6 203 124.1 1998 122 179.6 126.9 All 105.2 122.9 107.5 * (Bm/Bf)x100, where Bm and Bf are the numbers of male and female births, respectively

A conclusion from this section is that there are quality concerns with the mortality data from the high- impact areas, but that they are reduced by the selection of baseline and endline periods for calculation of U5MR.

IIP-JHU | Retrospective evaluation of ACSD in Ghana A95

APPENDIX L References for Appendices

1. FAO. Gateway to Land and Water Resources, Ghana obtained from: http://www.fao.org/ag/agl/swlwpnr/reports/y_sf/z_gh/gh.htm#overview, accessed 10 August 2008., 2004.

2. U.S. Census Bureau. International Data Base - Country Summary: Ghana, 2008.

3. Efem, I.J.-A., Caroline; Anemana, Sylvester; Addai, Edward; Awittor, Evelyn; Ankrah, Victor. Report of the Review of the Accelerated Child Survival and Development Programme in the Upper East Region of Ghana, Nov. 2004, 2004.

4. Germer, J.S., Joachim. Climate at Valley View University. Stuttgart, Germany, University of Hohenheim, 2008.

5. Ghana Statistical Service (GSS) and Macro International Inc (MI). Ghana Demographic and Health Survey 2003. Calverton, Maryland, GSS and MI, 2003.

6. Aventis. K-O Tab Net Treatment Kit, ND.

7. Republic of Ghana. Districts of Ghana, Republic of Ghana.

8. Adjasi, C.D.K.O., K.A. Poverty Profile and Correlates of Poverty in Ghana. International Journal of Social Economics. 34 (7): 22 (2007).

9. United Nations Children's Fund. The State of the World's Children - Child Survival. New York, UNICEF, 2008.

10. Otupiri, E.O.-A., Rose. Health and Development Programs and Policy Mapping Exercise in Upper East Region and the Rest of Ghana. Kumasi, School of Medical Sciences, Kwame Nkrumah University of Science and Technology, 2007.

11. UNICEF. Report of the Review of the Accelerated Child Survival and Development Programme in the Upper East Region of Ghana, November 2004. Upper East Region, 2004.

12. Abaseka, V. Annual Performance Review, 2004. Upper East Region, 2004.

13. NA. A Brief Resume on ACSD Acitivies In Upper East Region. Upper East Region, 2006.

14. UNICEF. In-House Annual Review (2004), 2004.

15. NA. ASCD Annual Report: January-December 2005. Upper East Region, 2005.

16. Abaseka, V. Annual Report of Activities of ACSD, January-December 2005. Upper East Region, 2005.

17. UNICEF. In-House Annual Review (2005), 2005.

18. Abaseka, V. Abaseka, V./2006. Upper East Region, 2006.

19. NA. Report on Accelerated Child Survival and Development Meeting (ACSD) in Upper East Region-3rd August 2006. Upper East Region, 2006.

A96 IIP-JHU | Retrospective evaluation of ACSD in Ghana

20. Abaseka, V. & Nsiire, A. Overview & Update on ACSD in UER. Upper East Region, 2004.

21. Service, G.H. Annual Report 2005, Upper East Regional Health Administration 2006

22. Service, G.H. Annual Report, 2004; Upper East Regional Health Administration (2005).

23. Office, U.R.H. None (EPI Report). Upper East Region, 2004?

24. GHS. Expanded Programme on Immunization, Ghana Health service EPI Annual Report - 2006 [Health Children Happy Nation!], 2007.

25. NA. Achievements 2004. Upper East Region, 2004.

26. Administration(?), D.H. Executive Summary (Annual Report). Upper East Region, 2004.

27. Administration(?), D.H. None (Annual Report). Upper East Region, 2005.

28. UNICEF. Final Progress/Financial Report; A Grant For Better Utilization of Immunization Services; SC/2000/0329. Upper East Region, Northern Region, 2003.

29. NA. A Report on Monitoring Visits of CIMI Agents, ND.

30. NA. 2nd Report of KNUST Team on Implementation of Region-wide C-IMCI, UER. Upper East Region, 2004.

31. NA. Report on 2nd CBA Training in Half of Upper East Region. Upper East Region, 2004.

32. NA. 5th Report of KNUST Team on Region-wide Impementation of C-IMCI, UER. Upper East Region, 2004.

33. NA. Report on Extension Staff Training Workshop Held at Bawku on 26-29 Oct. 2004, 2004.

34. NA. IMCI Case Management Training, 2006.

35. NA. Database For Community Based Volunteers and Midwives Upper East Region-2004, 2004.

36. WHO. WHO Child Growth Standards: Methods and development: Length/height-for-age, weight- for-age, weight-for-length, weight-for-height and body mass index-for-age. Geneva, World Health Organization, 2006.

37. Ghana Statistical Service and Macro International, I. Ghana Demographic and Health Survey (October 1999).

38. Ghana Statistical Service, N.M.I.f.M.R.a.M.I., Inc. Ghana Demographic and Health Survey, 2003 September 2004.

39. Bryce, J., Gilroy, K., Black, R.E., Jones, G. & Victora, C.G. A Retrospective Evaluation of the Accelerated Child Survival and Development Project in West Africa; Inception Report. Baltimore, MD, Johns Hopkins University Institute for International Programs, 2007

IIP-JHU | Retrospective evaluation of ACSD in Ghana A97

APPENDIX M Mapping of partners; activities in “High-impact” districts (Upper East region) and nationally

A98 IIP-JHU | Retrospective evaluation of ACSD in Ghana

HEALTH AND DEVELOPMENT PROGRAMS AND POLICY MAPPING EXERCISE IN UPPER EAST REGION AND THE REST OF GHANA

School of Medical Sciences Kwame Nkrumah University of Science and Technology Kumasi, Ghana December, 2007

IIP-JHU | Retrospective evaluation of ACSD in Ghana A99 Health and development programs and policy mapping 1996­2007

This document is prepared by the School of Medical Sciences-Kwame Nkrumah University of Science and Technology, Kumasi for the sole purpose of The Institute for International Programs, based at Johns Hopkins University Bloomberg School of Public Health, internal use. All information contained in this document may not be disclosed, distributed or reproduced in whole or in part to any third party without the express written permission of The Institute for International Programs.

Authors: Dr. Easmon Otupiri and Ms Rose Odotei-Adjei SMS-KNUST

2

A100 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

CONTENTS

CONTENTS 3 ACKNOWLEDGEMENTS 5 LIST OF ACRONYMS AND ABBREVIATIONS 6 LIST OF TABLES AND FIGURES 10 EXECUTIVE SUMMARY 11

1. INTRODUCTION 1.1 Background Information 12 1.2 Country Profile 13 1.3 Objectives and Framework of Mapping Exercise 14

2. METHODOLOGY 2.1 Study Methods and Design 15 2.1.1 Desk Review 15 2.1.2 Key Informant Interviews 15 2.1.3 Field Work 15 2.1.4 Period of Mapping Exercise 15 2.1.5 Organization of Report 16

3. HEALTH AND DEVELOPMENT PROGRAMS UPPER EAST REGION 3.1 Profile of Upper East Region 17 3.2 Diocesan Health Services 19 3.3 Ghana Red Cross Society 22 3.4 World Vision International 23 3.5 Widows and Orphans Movement 26 3.6 Community Water and Sanitation Agency 26

4. HEALTH AND DEVELOPMENT PROGRAMS GHANA 4.1 Ghana Sustainable Change Project 28 4.2 Japanese International Cooperation Agency 30 4.3 United States Agency for International Development 31 4.4 Donor Support 34

3

IIP-JHU | Retrospective evaluation of ACSD in Ghana A101 Health and development programs and policy mapping 1996­2007

4.5 World Health Organization 35 4.6 Opportunities Industrialization Centres 37 4.7 Department for International Development 38 4.8 Danish International Development Agency 38 4.9 Engender Health 40

5. CHANGES IN HEALTH POLICIES IN GHANA 5.1 Global and Regional Policies 41 5.2 National Policies 42 5.3 Health Sector Policies 43 5.4 Health Interventions and Programs 44

6. CONCLUSION 51

APPENDICES List of Selected Reviewed Documents 52 Summary of Focus Group Discussions Results 54 Interview Schedule for Regional and National Level Agencies 55 Focus Group Discussion Guide for Groups in Upper East Region 56 Dummy Table for Mapping Exercise 58

4

A102 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

ACKNOWLEDGEMENTS

SMS-KNUST would like to acknowledge the following individuals and organizations for their support in the production of this report.

Non-Governmental Organizations and Communities – Upper East Region Mr. Joseph Ayembilla Diocesan Health Service, Bolgatanga Mr. James Tobiga Diocesan Health Service, Bolgatanga Mr. Joseph Abarike Azumah Ghana Red Cross Society Ms. Benedicta Pielore World Vision International Ms. Betty Ayagiba Widows and Orphans Ministry Mr. Suleman Alhassan Action Aid Communities in Bongo district Health Partners – Accra Mr. Jacob Larbi GSCP Ms Comfort Yankson GSCP Mr. George Graves Woode JICA Ms. Julia A. Pwamang USAID Ms. Gregoria Dawson-Amoah World Bank Dr. Nana Ama Brantuo WHO Mr. Stanley Diamenu WHO Dr. Atubrah OICI Ms. Yvonne Agbesi DFID Ms. Helen Dzikunu DANIDA Mr. Marius DeJong Netherlands Embassy Ms Loretta Benton EU Ghana Health Service Dr. K.O Antwi-Agyei EPI Dr. Isabella Sagoe-Moses Child Health Dr. Henrietta Odoi-Agyarko Reproductive Health Unit Ms. Esi Amoafo Vitamin A programme Ms Vida Abaseka RHA – Upper East Region Dr. Joseph Amankwah Regional Director of Health – Upper East Dr. K. Marfo DDHS – Bongo District Dr. Dodoo DDHS – Bawku Municipality Ms. Naa Kokor Allotey NMCP KNUST Ms. Janet Asihene Department of Community Health – SMS Ms. Bibi Kaleem Department of Community Health – SMS Mr. Samuel Boateng Department of Community Health – SMS Dr. E.N.L. Browne Department of Community Health – SMS

5

IIP-JHU | Retrospective evaluation of ACSD in Ghana A103 Health and development programs and policy mapping 1996­2007

LIST OF ACRONYMS AND ABBREVIATIONS

ACSD Accelerated Child Survival and Development ACT Atermisinin Combination Therapy ADB African Development Bank ADP Area Development Programme AED Academy for Educational Development AIDS Acquired Immune-deficiency Virus AMCROSS American Red Cross ANC Ante-natal Clinic ARI Acute Respiratory Infection BCC Behaviour Change Communication BCG Bacillus Chalmette Guerin CBD Community-based Distributor CBGP Community-based Growth Promotion CD4 Cluster of Differentiation 4 CDD Control of Diarrhoeal Diseases CEDEP Centre for the Development of People CHPS Community-based Health Planning and Services CHPW Child Health Promotion Week CIDA Canadian International Development Agency C-IMCI Community-Integrated Management of Childhood Illness CND Canadian Dollar CRS Catholic Relief Services CSM Cerebro-spinal Meningitis CSO Civil Society Organization CWC Child Welfare Clinic CWSA Community Water and Sanitation Agency DANIDA Danish International Development Agency DDHS District Director of Health Services DFID Department for International Development DIS Daily Immunization Services DHMT District Health Management Team

6

A104 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

EMD Epidemic Meningococcal Disease EPI Expanded Programme on Immunizations EU European Union FBO Faith-based Organization GAIN Global Alliance for Improved Nutrition GAVI Global Alliance for Vaccine and Immunization GBP Great Britain Pound GDP Gross Domestic Product GFATM Global Fund to fight AIDS, Tuberculosis and Malaria GHS Ghana Health Service GPRS Ghana Poverty Reduction Strategy GRCS Ghana Red Cross Society GSCP Ghana Sustainable Change Project GSK Glaxo SmithKline GOG Government of Ghana HC Health Centre HIRD High Impact Rapid Delivery HIV Human Immune-deficiency Virus HRAP Human Rights-based Approach to Planning HSPS Health Sector Programme Support IEC Information Education and Communication ILO International Labour Organization IDSR Integrated Disease Surveillance and Response IMCI Integrated Management of Childhood Illness IMR Infant Mortality Rate

IPTP Intermittent Preventive Treatment in Pregnant Women

IPTI Intermittent Preventive Treatment in Infants ITN Insecticide Treated Net JHPIEGO Johns Hopkins International Education Programme in Gyn & Obst JICA Japanese International Cooperation Agency K-N Kassena-Nankana LSS Life Saving Skills MCH Maternal and Child Health MDG Millennium Development Goal

7

IIP-JHU | Retrospective evaluation of ACSD in Ghana A105 Health and development programs and policy mapping 1996­2007

MH Maternal Health MNH Maternal and Neonatal Health MoH Ministry of Health MSHAP Multi-sector HIV/AIDS Project MTHS Medium Term Health Strategy NEPAD New Partnership for Africa’s Development NGO Non-Governmental Organization NHIS National Health Insurance Scheme NID National Immunization Day NMCP National Malaria Control Programme OICI Opportunities Industrialization Centres International OPV Oral Polio Vaccine ORS Oral Rehydration Salt PLWHA People Living with HIV/AIDS PMTCT Prevention of Mother-to-Child Transmission POW Programme of Work RBM Roll Back Malaria RCH Reproductive and Child Health RH Reproductive Health RHA Regional Health Administration RHI Rural Help Integrated RED Reaching Every District SHARP Strengthening HIV/AIDS Response SMI Safe Motherhood Initiative

SWAP Sector-wide Approach TBA Traditional Birth Attendant TT Tetanus Toxoid U5MR Under-five mortality rate UK United Kingdom UN United Nations UNAIDS Joint United Nations Programme on HIV/AIDS UNDP United Nations Development Programme UNFPA United Nations Population Fund UNICEF United Nations Children’s Fund

8

A106 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

US United States USAID United States Agency for International Development USD United States Dollar VCT Voluntary Counselling and Testing VIP Ventilated Improved Pit VLOM Village Level Operated and Maintained WATSAN Water and Sanitation WB World Bank WFP World Food Programme WHO World Health Organization WVI World Vision International

9

IIP-JHU | Retrospective evaluation of ACSD in Ghana A107 Health and development programs and policy mapping 1996­2007

LIST OF TABLES AND FIGURES

Table 2.1 Mapping exercise steps Table 3.1 Trend in growth of health facilities and health centres in UER Table 3.2 Diocesan health and development programmes (1998-06) Table 3.3 GRCS health and development programmes (1999-06) Table 3.4 World Vision International health and development programmes (1996-06) Table 3.5 Community Water and Sanitation Agency Projects, UER (1973-2005) Table 4.1 GSCP health and development programmes (1995-09) Table 4.2 JICA health and development programmes (2003-07) Table 4.3 USAID support for HIV/AIDS, child survival and nutrition (1998-07) Table 4.4 USAID funded interventions in 30 target districts (2003-07) Table 4.5 Donor support for health including HIV/AIDS (2003-07) Table 4.6 Donor support for water and sanitation (2003-07) Table 4.7 WHO health and development programmes (2003-07) Table 4.8 OICI health and development programmes (2003-06) Table 4.9 DANIDA health support (2003-07) Table 4.10 DANIDA health funds for HIRD (2006) Table 5.1 Immunization coverage in Ghana (1997-2006) Table 5.2 Integrated Measles/Polo/Vitamin A/ITN Distribution Campaign 2006 Table 5.3 National Immunization Days 2005 Table 5.4 GAVI Immunization financing 2006-2010 Table 5.5 Major child nutrition-related projects in Ghana (1988-2010) Table 5.6 Health policies and programmes in Ghana Figure 3.1 Spot map of health institutions in Upper East Region, 2006

10

A108 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

EXECUTIVE SUMMARY

This report presents the results of the Health and Development Programmes and Policy Mapping Exercise in Upper East Region and the rest of Ghana which was conducted in July- August 2007. The Report covers the period 1996-2007. Its aim was to provide data for an external retrospective evaluation of the Accelerated Child Survival and Development (ACSD) Programme in four countries of West Africa. The ASCD aims to reduce child mortality using cost effective interventions, namely the Integrated Management of Childhood Illness (IMCI+), Antenatal Care (ANC+) and Expanded Programme on Immunizations (EPI+) interventions.

The Mapping Exercise covered non-Governmental Organizations (NGOs) and Civil Society Organizations (CSOs) involved in maternal and child health interventions in Upper East Region and Health Development Partners and the Ghana Health Service at the national level. This report presents data on the specific child health, reproductive health, maternal health, micro credit and, water and sanitation interventions by the various agencies in the Upper East Region. At the national level the Report provides data on policies and programmes by the Ghana Health Service with nationwide coverage. It also includes a description of the activities and funding provided by the health development partners to Ghana over the period 1996-2007. The main data collection methods used were; interviews, focus group discussions and a review of secondary data in the form of reports, newsletters and presentations.

In Upper East Region various agencies have implemented maternal and child health interventions with varying degrees of coverage in terms of districts, communities and population. Even though quite a number of such agencies were identified data were collected from the top- five performing agencies in terms of programme relevance and coverage. The Diocesan Health Service provides static and outreach maternal and child health services through a hospital, seven (7) health centres and many outreach points. The Ghana Red Cross Society which is uniquely positioned as an auxiliary of the Ghana Health Service has implemented maternal and child health interventions including a child survival project in three districts from 2000-2002. The World Vision International focused its health and development programme in one district and since 1996 has supported and or implemented maternal and child health interventions in selected communities. The Community Water and Sanitation Agency has since 1994 provided more than 2000 water points and nearly 600 latrines region-wide.

At the national level the Expanded Programme on Immunization, the Child Health Programme, the Vitamin A Programme and the National Malaria Control Programme among others have implemented various interventions designed to reduce the morbidity and mortality burden of children under-five. Health development partners such as the United Nations Agencies, the World Bank, and bi-lateral and multi-lateral agencies have all supported Ghana’s Health Sector Programme. Some agencies provided support at the national level only while others supported at the national level and provided support directly to some districts.

Ghana has been implementing almost the full range of cost-effective evidence-based maternal, neonatal and child health interventions and this combined with an increasing expenditure on health should have resulted in improved maternal and child health indicators for the country. If there is evidence to demonstrate that the child health indicators for Upper East region are better than the rest of the country and that the difference is significantly attributable to the ACSD intervention then the health systems for delivery of the interventions evidenced to reduce the morbidity and mortality burden in the rest of Ghana should be revised

11

IIP-JHU | Retrospective evaluation of ACSD in Ghana A109 Health and development programs and policy mapping 1996­2007

1.0 INTRODUCTION

Even though Ghana has achieved commendable economic growth in recent years (GDP growth rate at 6% in 2004 and 2005) the same cannot be said about the health gains in the country. Health indicators that showed a steady improvement in the 1990s have stagnated and in some instances have worsened in spite of the increased expenditure on health.

The 2006 Ghana Multiple Indicator Cluster Survey (preliminary report) reveals worsening of infant rate (IMR) and stagnation of the hitherto worsening under-five mortality rate (U5MR) in Ghana over the past eight years, even though these rates have been decreasing consistently over the past two decades. The infant mortality rate has increased from 57/1000 live births (LB) in 1998 to 64/1000 LB in 2003 to 71/1000 LB in 2006. Within the same period the under-five mortality (U5MR) has increased from 108/1000 live births to 111/1000 LB. This presents a major challenge to achieving the country’s targets for the 4th Millennium Development Goal (MDG 4).

1.1 Background In 1998 the health status indicators for the Upper East Region were the worst in Ghana; the infant and under-5 mortality rates in the Region were 82 and 155 deaths per 1,000 live births respectively, while the corresponding national figures were 57 and 108 deaths per 1,000 live births. The regional figures were marked by important district disparities. Malaria, diarrhoea and acute respiratory infections (ARI) with malnutrition as an underlying cause are responsible for most of the deaths. Since 1995, the United Nations Children’s Fund (UNICEF) in partnership with the Ministry of Health/Ghana Health Service (MOH/GHS) has been working to reduce child morbidity and mortality in Upper East Region in two target districts; Bawku East and Builsa. The partnership provided child survival interventions such as immunization campaigns, promotion of exclusive breastfeeding, vitamin A supplementation and, iron and folic acid supplementation at antenatal clinics. The national traditional birth attendant (TBA) programme trained TBAs to conduct safe delivery in the communities. Rural Help Integrated (RHI), a non-governmental organization (NGO) based in the Region trained community-based distributors (CBDs), to distribute family planning devices and also treat minor ailments in children and adults using chloroquine for malaria, and oral rehydration salts (ORS) for diarrhoea, while referring serious cases to health centres.

In spite of all these interventions the health status of children under-five remained poor and access to health care was limited; the interventions were not reaching enough of those who needed them most; the poor and vulnerable. Coverage of the key child survival interventions remained critically low. There was need to introduce a more rationale-based integrated approach that would use the tenets of human rights-based approach to programming (HRAP). This would ensure active community involvement by using the triple ‘A’ construct (Assessment, Analysis and Action) to get the interventions to reach the neediest.

West Africa is the region of the world with the highest maternal, neonatal and child mortality rates. Large scale collaboration across 100 districts within 11 countries in West and Central Africa began in 2002 with the aim of a phased approach to scaling up essential child health interventions. Partnership was key – funded by the Canadian Government and initiated by UNICEF, Accelerated Child Survival and Development (ACSD) involves the expertise and partnership of multiple players, including governments and health ministries, the World Health Organization (WHO), the World Bank, non-governmental organizations, NGOs and local community leaders and members.

12

A110 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Through ACSD, effective interventions for children and pregnant women are bundled in an integrated, cost-effective package including immunization of children and pregnant women, micronutrient supplementation, breastfeeding promotion, supply of oral rehydration solution for diarrhoea and insecticide-treated bed nets for protecting children and women from malaria. The approach focuses on extending health coverage to underserved communities and using community outreach efforts to deliver services and commodities closer to families. Outreach services are also accompanied by programmes to educate families on home-based healthcare practices for their children.

In 2000 UNICEF started implementing aspects of the Upper East Region ACSD project in collaboration with the Ghana Government and with support from the Canadian International Development Agency (CIDA), the MOH/GHS, the Ghana Red Cross (GRC) and the American Red Cross (AMCROSS) in three districts (Bolgatanga, Bawku East and Bawku West). The main objective was to use results-based planning techniques and evidence from other interventions within Africa and in-country experience to increase coverage with three packages of high-impact rapid delivery interventions known to reduce the morbidity and mortality burden in children- under-five and pregnant women. The focus was on EPI+ (vaccinations, Vitamin A supplementation, ITN usage and deworming); ANC+ (IPTp, tetanus vaccination, Iron/folic acid supplementation, ITNs usage and PMTCT) and IMCI+ (clinical and home management of malaria, diarrhoea and ARI, community-based growth promotion and iodated salt usage). The initial selected implementation was scaled-up to assume a region-wide dimension in 2002.

1.2 Country Profile The Republic of Ghana located in West Africa is bordered on the north and north-west by Burkina Faso, on the east by Togo, on the south by the Gulf of Guinea, and on the west by La Côte d’Ivoire.

Formerly a British colony known as the Gold Coast, Ghana was the first majority-ruled nation in sub-Saharan Africa to achieve independence, in 1957. The population of the country, according to the 2000 Population census was 18,800,000. However current estimates in the year 2005, put the population of Ghana at 21,946,000. The total area of Ghana is 238,500 km2 (92,090 miles2). The capital is Accra.

Ghana’s overall long-term vision for growth and development is detailed in the GHANA VISION 2020 document. The Medium Term Health Strategy (MTHS) Towards Vision 2020 articulates the national health plan which has been made operational in three programmes of work (POW) spanning five years each; POW I (1996-2001), POW II (2002-2006) and POW III (2007-2011). The Ghana Poverty Reduction Strategy (GPRS I and II) provides broad policy directions to guide the implementation of POW II and III in three key areas; bridging the equity gap, ensuing sustainable financial arrangements for the poor and enhancing efficiency in the health system. The policy thrust of each annual POW is informed by an assessment of the previous year’s POW by joint independent Ministry of Health/Ghana Health Service/Health partner reviews with external assistance. A number of sector-wide indicators have been developed to measure performance. Total per capita health expenditure grew in 2005-6, by 40% in nominal and 26% in real terms. Source of funding include public and donor sources, as well as user fees paid through public facilities. Total health as a proportion of total government expenditure increased from 12% in 2002 to 14% in 2005.

13

IIP-JHU | Retrospective evaluation of ACSD in Ghana A111 Health and development programs and policy mapping 1996­2007

1.3 Objectives and Framework of the Mapping Exercise 1. To collect information about other health and development programs in districts in the Upper East Region from 1999 to present, including the overall effort and geographical coverage of these projects.

2. To collect information about other large-scale health and development initiatives in the rest of Ghana from 1999 to present, including the overall effort and geographical coverage of these projects.

3. To document changes in national and local health policies in Ghana from 1999 to present that may have impacted child health and survival.

14

A112 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

2.0 Methodology

2.1 Study Methods and Design In general, the mapping exercise employed literature/secondary resources review, semi- structured interviews, focus group discussions, observations and visual techniques such as village mapping and transect walks. A sample interview guide for semi-structured interviews with NGOs, programme managers of GHS and health development partners is included at the end of this document.

2.1.1 Desk review Information was gathered and analysed on child health, child health strategies and programmes, IMCI, and the context in which reproductive and child health interventions are implemented. The full list of documents reviewed is in the reference. The data gathered were entered into a dummy table

2.1.2 Key informant interviews Semi-structured interviews were conducted with programme managers, directors and coordinators at headquarters, health partners and their collaborating agencies and the private sector. The interviews were recorded either manually or with a digital voice recorder.

2.1.3 Field work The Upper East region was visited. At the regional level key informant interviews were undertaken with NGOs in reproductive and child health. At the community level, key informant interviews involved community health officers, volunteers and community leaders, and focus group discussions were done with men and women’s group separately with the view validating the information gathered at the regional level. At the national level unstructed interviews were held with health development partners, bi-lateral and multi-lateral donor agencies and program managers of relevant units within the Ghana Health Service such as EPI, vitamin A, child health and RBM.

All data collected were manually analyzed.

2.1.4 Period of mapping exercise The major part of the mapping exercise took place from July 2 – August 13, 2007. Due to the incomplete and sometimes outright paucity of information gathered a series of follow-ups are on-going. Information gathering at the national level has been very challenging. Agencies, health development partners and programme managers of the GHS were quite uncooperative in many instances.

2.1.5 Organization of Report The report starts with an introduction that captures the background to the report and gives an account of the objectives for the mapping exercise. Chapter two states the methods used for the data collection. Chapter three gives an account of the health and development programs in Upper East Region dating 1996-2007. In Chapter four the health and development programs undertaken by health development partners in Ghana are presented. Chapter five looks at the health-related policies from the global angle to the national dimension. In Chapter six conclusions are drawn based on the data collected.

15

IIP-JHU | Retrospective evaluation of ACSD in Ghana A113 Health and development programs and policy mapping 1996­2007

Table 2.1: Mapping Exercise Steps Area of Inquiry/Topic Methodologies 1. Health and development projects in Upper • July 2007, discuss tools with study East region agencies • Study visit notes and reports a. Summarize interviews of personnel from the collected from agencies Ghana Health Service, Diocesan Health Service, • Transcribe audio recordings of Ghana Red Cross Society and World Vision interviews International. • Take digital pictures of some respondents b. Meet women’s groups and community-based agents in communities region-wide Interviews • Ghana Health Service c. Undertake transect walks through selected • Diocesan Health Service communities to observe evidence of interventions • Ghana Red Cross Society (bore holes, clinics, services) by agencies. • World Vision International

• CWSA

• AfriKids • Action Aid • Rural Aid • SYTO • Action Child Mobilization

Focus Group Discussions • Community-based volunteers and mothers

Observation • Evidence of interventions by NGOs

Analysis • Manual 2. National Policies and Programmes • Send out letters to request interviews with health development a. Meet and interview representatives of health partners and GHS programme development partners managers • Late-July undertake interviews and b. Meet and interview GHS programme collect reports for study managers • Transcribe interviews • Initiate report writing • Undertake follow-up to complete gaps in information collected 3. Report Writing • Early-August prepare initial draft report • Submit final draft report by end of August 2007

16

A114 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

3.0 Health and Development Programmes in Upper East Region (1996-2007)

3.1 Profile of Upper East Region The Upper East Region is located in the north-eastern corner of the country between longitude 0° and 1° West and latitudes 10° 30″N and 11°N. It has two international boundaries; namely Burkina Faso to the north and the Republic of Togo to the East. Peoples of these three countries share so much in common: language, socio-cultural and belief systems. There is intense cross border movement of people, goods and services at these borders. The challenges of disease surveillance and control in particular and health service delivery in general arising out of this geo-physical and social cultural associations are enormous and often overwhelming.

The Region lies within the meningitis belt of Africa. This belt is made up of 21 countries with a population of 250 million in the age group 2- 29 years. This age group is the most vulnerable to Epidemic Meningococcal Disease (EMD/CSM) epidemics. Focal outbreaks and sometimes very widespread and devastating epidemics are commonplace events in the region each year. The Region also lies within the savannah blinding onchocerciasis belt of West Africa. Before the inception of Onchocerciasis Control Programme, blinding rates from onchocerciasis were as high as 10% in some communities. Even though the disease is practically controlled, the flies still pose serious nuisance to farming communities along the fertile river basins. In addition to mass distribution of ivermectin to communities with residual infections, active epidemiological surveillance is on-going for early detection of any recrudescence of the disease.

The other major characteristic features are: • Surface area of the region is 8,842 sq.km (about 3.7% of the country), with: • A short and scanty rainfall of about 800-900mm per annum followed by a long dry season with dry harmattan winds and hot periods – 40o C. • Population from 2000 census is 920,089 (this is about 4.8% of total population of country) • Growth rate 1.1% • Projected Population for 2006 is 982,510 • Density 110 people/sq.km, range 36 - 175 as compared to national average of 91 • Population is largely rural (87%). • Settlement pattern is highly dispersed in 911 communities • Five main languages are spoken in the region (Gurune, Kusal, Kasem, Buili and Bisa)

Road network The Region has 1017 kilometres of feeder roads. Of this, 700 km representing 69% are motorable and 317 km, representing 31% are certified as non-motorable. It has a total of 54.8 km of trunk roads. Of these 31.0 km is national road, 63.6 km is inter-regional and 173.3 km is regional roads.

Safe water coverage • In Bawku East - 55.62% • Bawku West - 96.49% • Bolgatanga - 39.25% • Bongo - 59.40% • Builsa - 74.02% • Kassena-Nankana 17

IIP-JHU | Retrospective evaluation of ACSD in Ghana A115 Health and development programs and policy mapping 1996­2007

Fig 3.1: Spot Map of Health Institutions, Upper East Region, 2006

&d d d

Ñd dÑ Ñ & & ÑÑ & & d d &d d d &d & & & & Ñ Ñ d & Ñ d & d Ñ & Ñ & & ddÑ & Ñ & && Ñ & Ñ & & d d d & Ñ & & d & Ñ Ñ & d & & Ñ & Ñ d &d Ñ d & & & d & &d d d d d & Ñ & & Ñ d d d & d d Ñ & &d &d Ñd Ñ & Ñ d &d & Ñd Ñ d Ñ d d Ñ d & Ñ d Ñ

Ñ & &

Ñ &

Ñ

d Clinics Ñ District Hospita ls & Ñ Health Centres & Completed CHPS compounds Ñ Reg Hospital New distsam .s hp Bawku East N Bawku West 10 0 10 20 Kilometers Bolgatanga Bongo Builsa W E Garu-Tempane Kassena-Nankana Talensi-Nabdam S

Source: Regional Annual Health Report, 2006

Table 3.1: Trend in Growth of Health Facilities and Health Centres in UER Institution/Year 2002 2006 % Increase Hospital 6 6 0.0 Health Centres 26 32 23.1 Clinics 46 47 2.2 Maternity Home Private) 2 2 0.0 CHPS 7 68 871.4 Training Institutions 4 5 25.0 Total 91 160 75.8 Source: Regional Annual Health Report, 2006

This mapping exercise identified six (6) leading CSOs/NGOs regionally but was able to capture information from four (4). Of the two (2) from which information was not available, one (Rural Help Integrated) which was very active in sexual and reproductive issues had folded up and the other (Action Aid) had experienced such a high staff turnover that it was impossible to get their records straight.

Rural Help Integrated was active in Bolgatanga, Bongo, and Builsa districts. It distributed condoms, promoted family planning and provided home-based management of uncomplicated cases of malaria and diarrhoea (in children and adults) with referral for severe cases through a network of community-based distributors. The districts in which it operated recorded remarkable increases in contraceptive acceptance and use. Rural Help Integrated handed its assets over to

18

A116 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

the GHS and AfriKids (an NGO that supports street children). Action Aid has been involved in reproductive health interventions region-wide. It has also provided hand dug wells in Bawku West district and has funded interventions through other NGOs and agencies such the Widows and Orphans Movement, the Ghana Red Cross Society and the Diocesan Health Service. Action Aid funded the training and equipping of over 50 TBAs in various communities.

Information was collected from the Navrongo—Bolgatanga Diocesan Health Service of the Roman Catholic Church, the Ghana Red Cross Society (Upper East Region), the Widows and Orphans Movement and the World Vision International (Bongo Area Development Programme).

A quasi-governmental agency responsible for water and to some extent sanitation in small towns and communities; Community Water and Sanitation Agency, was included in the mapping exercise for Upper East region in view of the impact water and sanitation has on child survival.

3.2 Diocesan Health Service The Navrongo-Bolgatanga Diocese of the Roman Catholic Church is one of the 18 arch- dioceses in Ghana. It covers 11 districts with a total population of 1.6 million spread over 885 communities in Upper East and Northern regions. The Diocesan Health Office provides static and outreach services through one (1) hospital and seven (7) health centres and many outreach points. The Health Office works in conjunction with the Catholic Relief Services.

Table 3.2: Diocesan Health and Development Programmes (1998-2006)

Year Intervention Activity Indicator District Community 1998 RH (MNH) Skilled attendant at 110 deliveries Builsa, K-N, Wiaga, Sirigu, birth Bongo Biu 1999 RH (MNH) Skilled attendant at 123 deliveries Bongo, Builsa Kongo, birth Wiaga

RH (MH) ANC 4550 women Bongo, Builsa, K-N Kongo, Wiaga, Nakolo, Biu, CH CWC 6998 children Bongo, Zorko Builsa, K-N Kongo, Nakolo Wiaga, Zorko 2000 RH (HIV/AIDS) Care and support for PLWHA and orphans X X X and vulnerable children RH (MNH) 110 deliveries Bongo, Builsa Nakolo, Zorko Skilled attendant at birth

RH (MH) 5693 Bongo, Builsa Nakolo, ANC attendance Zorko, Kongo, Biu CH Bongo,Builsa

19

IIP-JHU | Retrospective evaluation of ACSD in Ghana A117 Health and development programs and policy mapping 1996­2007

CWC 8612 Nakolo, attendance Zorko, Kongo 2001 RH (HIV/AIDS) Health education, training of HIV/AIDS X X X educators

RH (MNH) Skilled attendant at 541 deliveries Bongo, Nakolo, birth Builsa, K-N Sirigu, Kongo, Wiaga, Zorko, RH (MH) ANC 9697 Bongo, Biu attendances Builsa, K-N Nakolo, Kongo, RH (MNH) TT2 Immunization 2871 Bongo, Wiaga, Biu Builsa, K-N Sirigu, Kongo, CH CWC 15252 Bongo, Wiaga, Zorko attendances Builsa, K-N

CH EPI BCG (2594), Bongo, Sirigu, OPV 3 Builsa, K-N Kongo, (2765), DPT Wiaga, Zorko 3 (2719), Measles Sirigu, (3917) Kongo, Wiaga, Zorko 2002 RH/CH Free services for pregnant women and children under-five

RH (MNH) Skilled attendant at 292 deliveries Bongo, Kongo, birth Builsa, Nakolo, K-N Sirigu, Wiaga, Zorko RH (MH) ANC 9089 Bongo, Kongo, Builsa, K-N Nakolo, Sirigu, Wiaga, Biu, Zorko CH CWC 14059 Bongo, Kongo, Builsa, K-N Sirigu, Wiaga, Zorko 2003 NOT AVAILABLE 2004 RH (HIV/AIDS) Support for orphans 500 Bongo, and vulnerable Builsa, children Bawku East, Bawku West,

20

A118 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Bolgatanga RH (MNH) 857 Kongo, Skilled attendance at Bongo, Nakolo, birth Builsa, K-N Sirigu, Wiaga, Zorko RH (MH) 7304 ANC Bongo, Kongo, Builsa, K-N Nakolo, Sirigu, Wiaga, CH 19480 Zorko CWC Bongo, Builsa, K-N Nakolo, Sirihu, Wiaga, Zorko 2005 RH (MNH) Skilled attendance at 1053 Bongo, Biu, Kongo, birth Builsa, K-N Nakolo, Sirigu, Wiaga, Zorko RH (MH) ANC 8204 Bongo, Builsa, K-N Biu, Kongo, Nakolo, Sirigu, Wiaga, CH CWC 35390 Bongo, Zorko Builsa, K-N Biu, Kongo, Nakolo, CH C-IMCI training 20 staff Bongo, Sirigu, Wiaga, Builsa, K-N Zorko

WATSAN Mechanized borehole 1 Bongo

Zorko 2006 RH(HIV/AIDS) Support for PLWHA and orphans and vulnerable children

HIV awareness creation RH (MNH) Skilled attendance at 1133 Bongo, Biu, Kongo, birth Builsa, K-N Nakolo, Sirigu, Wiaga, Zorko WATSAN Mechanized borehole 1 Bongo Nakolo Health systems Vehicle/ambulance 1/1 Bongo strengthening Zorko/Kongo Source: Field data, 2007

21

IIP-JHU | Retrospective evaluation of ACSD in Ghana A119 Health and development programs and policy mapping 1996­2007

3.3 Ghana Red Cross Society, Upper East Region The Ghana Red Cross Society (GRCS) seeks to serve humans and promote healthy living in more deprived areas. The GRCS launched its first primary health care programme in 1974 at Nsuopun (Western Region) when through women volunteers called mothers clubs it sought to reduce morbidity and mortality in vulnerable children and mothers. Currently more than 500 mothers clubs are registered nationwide. In Upper East Region the GRCS is uniquely positioned as an auxiliary of the Ghana Health Service. The American Red Cross (AMCROSS) supported 60 GRCS mothers clubs from 1999-2002 to implement a child survival project in three districts (Bawku East, Bolgatanga and Bawku West that focused on; Social mobilization for immunization Nutrition and breastfeeding education Acute respiratory infection recognition and prompt referral Diarrhoea case management with ORS Malaria prevention and home based care.

Table 3.3: Ghana Red Cross Society Health and Development programmes (1999-2006) Year Intervention Activity Indicator District Community 1999 CH Child survival intervention; 15000 children Bawku East, 47 health education Bawku West, communities home-based management of Bolgatanga malaria home-based management of diarrhea referrals for ARI 2000 CH Child survival intervention 18097 children Bawku East, 60 (AMCROSS) through 60 Bawku West, communities mothers’ clubs (total Bolgatanga membership of 120 mothers)

WATSAN Wells and Mozambican toilets 3 wells Bawku East 3 Mozambican and West, toilets Bolgatanga

CH Child survival (EU) through 180 Not available Bawku East 145 mothers’ clubs and West, communities Bolgatanga 2001 MCH Training in child survival 400 members Bawku West 200 interventions of mothers’ and East, communities clubs Bolgatanga 2002 MCH Child survival project ends 2003 MCH Basic care for women and Bawku East All (over children project in rural and West, 900)communit communities (EU)-training for Bolgatanga, ies within women Builsa, K-N, Upper East Bongo Region ACSD training 1802 women 2004 MCH ACSD 1802 women Bawku East

22

A120 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

and West, Bolgatanga, Builsa, K-N, Bongo

2005 MCH ACSD retraining 140 women Bawku East and West, Bolgatanga, Builsa, K-N, Bongo

Health systems strengthening Bicycles and Bawku East medicines and West, Bolgatanga, Builsa, K-N, Bongo

WATSAN Well provision 1 Bolgatanga

2006 MCH ACSD 9750 mothers Bawku East and West, Bolgatanga, Builsa, K-N, Garu- Tempane, Talensi- Nabdam, Bongo Source: Field data, 2007

3.4 World Vision International – Bongo Area Development Programme The World Vision International (WVI) developed the Area Development Programme (ADP) as a strategy to implement a total development agenda for areas in greatest need. The Bongo ADP began in 1996 with funding from World Vision Switzerland. The ADP has operated district wide but with emphasis on three zones (sub-districts); Bongo-Soe, Beo-Adaboya and Bongo Central.

Table 3.4: World Vision International Health and Development Programs (1996-2007) Year Intervention Activity Indicator District Community 1996- CH Construction of nutrition 3 centres Bongo Bongo-Soe, 98 rehabilitation centres Bongo Central, Beo

Equipment supply for 600 pre-school Bongo Adaboya supplementary feeding children

Logistics to DHMT for 10 weighing scales CWC 10 hanging scales, 5 cradle scales 1438 children dewormed, 2850 children

23

IIP-JHU | Retrospective evaluation of ACSD in Ghana A121 Health and development programs and policy mapping 1996­2007

vaccinated 1999 CH Training and logistic Bilhazia survey in Bongo support to DHMT school children

Construction of 1 Bongo Adaboya rehabilitation centre

Training on nutritional 150 lactating mothers Bongo Bongo Soe, management Adaboya 3 gas refrigerators, 8 Bongo Bongo Equipment supply sphygmomanometers Central

Supplementary feeding ( a 80 mini bags of Bongo Bongo Soe, lunch a day for school beans, 131 mini bags Beo, children) of rice, 370 mini-bags Adaboya of maize, 320L of cooking oil to cover 5791 children 2000 CH Supplementary feeding 2000 children Bongo (daily lunch for 10 nursery and primary schools)

Training in child nutrition 1200 lactating Bongo mothers and pregnant women WATSAN VIP toilet 26 Bongo Gowire Nayie, Kunkwa 2002 CH Logistic support for 2 17822 children Bongo District-wide rounds of polio, BCG and measles immunization and vitamin A supplementation

Training in child nutrition 1200 mothers Bongo

Deworming 4420 children Bongo

Supplementary feeding 2000 Bongo (daily lunch February- April)

RH HIV/AIDS education 998 youths Bongo

MH TBA skills training 10 Bongo District-wide Basic health training 1614 women Bongo

Micro credit Support for women 450 women Bongo 2003 CH Logistic support for 3 NIDs 76712 vaccinated Bongo District-wide

24

A122 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

for polio immunization and against polio in 3 vitamin A supplementation rounds and vitamin A given to 18952

Daily lunch (April-July) in 1325 children Bongo 10 schools MH Training in maternal 600 pregnant and Bongo nutrition lactating mothers

Micro credit Financial support 1800 women 2004 CH Logistic support for NIDs 64175 children Bongo District-wide vaccinated against polio

Deworming 17903 Bongo District-wide

RH HIV/AIDS educational Bongo 20 campaign communities 2005 MCH Best practices training 160 volunteers Bongo District-wide 70 TBAs

CH Training in child nutrition 640 mothers Bongo 6 sub- districts

Logistic support for 3 87,724 children Bongo District-wide rounds of NIDs vaccinated 22905 dewormed

Malaria control 1449 ITNs to children Bongo District-wide under-five

WATSAN Water pump management 122 people Bongo training 2007 CH Training in child nutrition 250 mothers Bongo Bongo Central, Bongo Soe, Beo, Adaboya

Supplementary feeding 934 pre-schoolers in Bongo 8 nurseries

Deworming 40000 children Bongo

RH Training in safe delivery 55 TBAs Bongo 6 zones

Training in maternal 1200 pregnant Bongo 7 zones health women

Micro-credit $ 1429 support for women 100 women Bongo

25

IIP-JHU | Retrospective evaluation of ACSD in Ghana A123 Health and development programs and policy mapping 1996­2007

Source: Field data, 2007

3.5 Widows and Orphans Movement – Upper East Region The Movement started out as the Widows Ministries, then Widows and Orphans Ministries before becoming a movement. The Movement promotes women’s and children’s rights through advocacy and capacity building. The Movement provides training and financial support to widows all over the Upper East Region. Currently 6865 widows are registered with the Movement. The movement is present in other regions in Ghana

3.6 Community Water and Sanitation Agency The Community Water and Sanitation Agency (CWSA) in Upper East region has been facilitating the region-wide provision of water supply and sanitation to communities. CWSA was established as a division of the Ghana Water and Sewerage Corporation in 1994. Water and sanitation coverage for Region stands at 51.27% and 0.81% respectively.

Over the period 1994 to 2007 a total of 2 067 water point sources (boreholes and hand dug wells) and eight (8) pipe-borne schemes have been provided by the CWSA with support from various partners. A total of 75 institutional VIP latrines and 498 household VIP latrines were provided throughout the Region during the same period.

Table 3.5: Community Water and Sanitation Projects, UER (1973-2005) Year Intervention Activity Indicator District Community 1973- Water Boreholes fitted with hand 1 860 Region-wide 1981 provision pumps (CIDA) 1979- Training Education on maintenance and 1 000 1992 (CIDA) management of water pumps communities 1982- Training Maintenance , repairs and 1 000 1988 (CIDA) hygiene promotion communities 1988- Training Ownership and management 50 Bolgatanga 1992 (UNDP) of pumps communities

Installation of village level 50 Bolgatanga operated and maintained pumps

Hand pump mechanics 100 Bolgatanga 1993- Community Animation of pump 1 647 Region-wide 2000 water communities project (CIDA) Borehole conversion to VLOM 1 602

Training of pump mechanics 3 204 2004- WATSAN Boreholes 68 Region-wide 2005 (GOG) VIP 4 2000- Water Boreholes 500 Region-wide 500 2004 provision communities (WB) Hand dug well with hand pump 4

26

A124 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Rehabilitation of water system 4 Bawku, Talensi- Nabdam K-N, Builsa Source: Field data, 2007

27

IIP-JHU | Retrospective evaluation of ACSD in Ghana A125 Health and development programs and policy mapping 1996­2007

4.0 Health and Development Programmes in Ghana

4.1 Ghana Sustainable Change Project (GSCP) The Ghana Sustainable Change Project (GSCP) is one of the leading NGOs which undertake USAID funded projects. GSCP is active mainly in 30 districts spread across the seven (7) southern regions in Ghana; Brong-Ahafo, Ashanti, Volta, Eastern, Western, Central and Greater Accra.

Table 4.1: Ghana Sustainable Change Project Health and Development Programs (2005- 2009) Year Intervention Activity Indicator Region District 2005- RCH Family planning, HIV/AIDS Western 4 districts; 09 stigmatization and behaviour (4 districts) Ahanta West change communication and Bibiani- malaria control Ahwiaso- Bekwai, Juabeso, Bia Central 13 districts; (13 districts) Abua-Asebu- Kwamankese, Agona, Ajumako- Enyan- Essiam, Asikuma- Odoben- Brakwa, Komenda- Edina- Eguafo- Abirem, Mfantsiman Twifo-Heman- Lower Denkyira, Upper Denkyira, Assin North, Assin South, Awutu-Efutu- Senya, Cape Coast, Gomoa, Greater Accra (1) Dangbe West

Volta (5) Kajebi, Akatsi, North Tongu,

28

A126 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

South Tongu, Nkwanta Eastern (2) Kwahu North, Birim North Ashanti (3) Ahafo-Ano South, Bosomtwe- Atwima- Kwanwoma, Amansie West Brong-Ahafo (2) Asutifi, Sene 2005 CH Training in malaria 100 health Upper West, communication staff Upper East, Northern, Greater Accra, Eastern, Brong-Ahafo

RH Training in HIV/AIDS BCC 19 health personnel 2006 CH Training in malaria 46 health Upper East, communication strategy personnel Upper West, Northern

MCH Training 750 7 southern 30 districts personnel regions from NGOs, CSOs

RH FP methods distribution 4 176 560 7 southern 30 target condoms regions districts 249 324 oral contracept ives

Training in HIV/AIDS 23 health communication personnel

Training in HIV/AIDS stigma 268 reduction members of FBOs and

29

IIP-JHU | Retrospective evaluation of ACSD in Ghana A127 Health and development programs and policy mapping 1996­2007

NGOs

Training in social mobilization 23 2007 RH Training on HIV/AIDS stigma 268 reduction members of FBOs and NGOs

Training in counselling skills 60 health personnel Source: Field data, 2007

4.2 Japanese International Cooperation Agency (JICA) JICA is one of the leading health development partner agencies in Ghana.

Table 4.2 Japanese International Cooperation Agency Health and Development (2003-07) Year Intervention Activity Indicator Region District 2003- RCH Financial support to GHS US $ 377 095.65 Eastern, Birim North, 06 PPAG to run static and Central Amuano outreach services –FP, Praso deliveries, child welfare clinics 2004- CH Funding for Parasite Greater Dangbe East 08 control project Accra 2004 RH Equipment supply 2 CD4 counter Greater 1 district per machines, Accra, region 2 haemoglobin Eastern analyzers + reagents 2005- RH HIV/AIDS control Eastern, 6 districts, 09 Ashanti 4 districts 2005 CH Financial support to GHS US $ 170 000 National National for EPI

Funding ITN retreatment US $ 29 905 Upper West Sissala East, Sissala West, Lawra 2006 RH Equipment supply 2 CD4 counter Brong-Ahafo Wenchi, machines, 2 Dormaa haemoglobin analyzers, 2 chemistry analyzers + reagents

CH EPI funding US $ 49 000 National National 2007 CH Funding for EPI US $ 170 000 National National Source: Field data, 2007

30

A128 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

4.3 United States Agency for International Development (USAID) USAID is currently committed to assist the Ministry of Health and the Ghana Health Service to improve the health status of Ghanaians through the Strategic Objective Grant Agreement Number Seven which covers the period 2003-2007. Over this period the USAID has provided nearly US $ 20 000 000 annually in the form of technical assistance and other support to Ghana.

USAID funded maternal and child health interventions through implementing partners in 30 target districts in the seven regions of southern Ghana (Brong-Ahafo, Eastern, Volta, Western Central Ashanti and Greater Accra). The specific interventions include; ¾ family planning; training of providers, improving private sector marketing of contraceptives and BCC at the community level ¾ newborn and neonatal care; immunizations ¾ hygiene improvement ¾ safe motherhood; strengthening health systems quality care interventions ¾ BCC interventions ¾ Scaling-up of proven cost effective clinical practices suitable for low-resource settings ¾ Nutrition of mothers infants and young children

Table 4.3: USAID support for HIV/AIDS, Child Survival and Nutrition (1998-2007) Year Intervention Activity Indicator Region District 1998 RH HIV/AIDS US $ 2995000

CH Child survival US $ 2412000

Nutrition Micro nutrient US $ 200000

1999 RH HIV/AIDS US $ 3925000

CH Child survival US $ 3350000

Nutrition Micro nutrient US $ 500000 2000 RH HIV/AIDS US $ 4025000

CH Child survival US $ 4350000

Nutrition Micro nutrient US $ 1000000 2001 RH HIV/AIDS US $ 4950000

CH Child survival US $ 4010000 2002 RH HIV/AIDS US $ 5500000

CH Child survival US $ 4300000

2003 RH HIV/AIDS US $ 8000000

CH Child survival US $ 3600000 2004 RH HIV/AIDS US $ 6300000

31

IIP-JHU | Retrospective evaluation of ACSD in Ghana A129 Health and development programs and policy mapping 1996­2007

CH Child survival US $ 3200000 2005 RH HIV/AIDS US $ 9135000

CH Child survival US $ 3200000 2006 RH HIV/AIDS US $ 6255000

CH Child survival US $ 2835000 2007 RH HIV/AIDS US $ 6750000

CH Child survival US $ 2986000 Source: Field data, 2007

USAID’s target regions and districts are; Ashanti Region (Bosomtwe-Atwima-Kwanwoma, Ahafo-Ano South and Amansie West districts), Brong-Ahafo Region (Sene and Asutifi districts), Central Region (Abura-Asebu-Kwamankese, Agona, Ajumako-Enyan-Essiam, Asikuma- Odoben-Brakwa, Assin North, Assin South, Awutu-Efutu-Senya, Cape Coast, Gomoa, Komenda-Edina-Eguafo-Abirem, Mfantsiman, Twifi-Heman-Lower Denkyira and Upper Denkyira districts), Eastern Region ( Kwahu North and Birim North districts), Greater Accra Region (Dangbe West district), Volta Region (Kadjebi, Akatsi, North Tongu, South Tongu and Nkwanta districts), Western Region (Juabeso, Ahanta West, Bibiani-Anhwiaso-Bekwai and Bia districts).

Table 4.4: USAID-funded Interventions in 30 target districts (2003-2007) Project Implementing partner Sub-grantees Region District Community- Population Council American College Seven 30 target districts based health of Nurse target planning and Midwives (ACNM) regions services Engender Health

Centre for the Development of People (CEDEP) Quality of care Quality Health Abt Associates Seven 30 target districts Partners/Engender target Health JHPIEGO regions

Initiatives Inc Strengthening Academy for Catholic Relief Ashanti Adansi East, HIV/AIDS Educational Services (CRS) (9 ) Adansi West, Response Development (AED) Adansi South, (SHARP) Futures Group Amansie East, Amansie West, Bosomtwe-Atwima- Kwanwoma, Kumasi Obuasi Offinso Eastern Asuogyaman, (9) Fanteakwa,

32

A130 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Kwaebibirim, Kwahu South, Yilo Krobo, Manya Krobo, Koforidua, Akwapim South, Suhum-Kraboa- Coaltar

Greater Accra, Accra (3) Tema, Ashiedu-Keteke

Western Takoradi, (10) Sekondi, Jomoro, Amanfi East, Amanfi West, Nzema West, Shama-Ahanta East, Shama-Ahanta West, Wassa West, Mpohor-Wassa East

Volta (1) Ketu Ghana AED Exp Momentum/ Seven 30 target districts plus Sustainable Group Africa target 29 SHARP districts Change Project Manoff Group regions CARE Social Netmark Volta, All districts marketing of Ashanti, ITNs Eastern, Brong-Ahafo

Technical DELIVER John Snow Inc National assistance in procurement and logistics of health commodities Demographic ORC Macro Ghana Statistical National and Health Service Surveys Hygiene and 60000 sanitation beneficiaries in Upper East, Upper West and Brong-Ahafo Source: Field data, 2007

33

IIP-JHU | Retrospective evaluation of ACSD in Ghana A131 Health and development programs and policy mapping 1996­2007

4.4 Donor support (bi- and multi-lateral) including The World Bank The World Bank (WB) provides funding for health through the Sector-Wide Approach (SWAp) which is a pooled funding mechanism that operates at the national level. Under the Health Sector Support Project I the World Bank provided funding for health to the tune of US $ 35m from 1998 to 2002. For the period 2005-2011 the Bank will support the national HIV/AIDS programme with US $ 20m through the multi-sector HIV/AIDS Program (MSHAP). Specific child health interventions by the Bank in the form of Community-based Growth Promotion can be found in four (4) districts in as many regions; Komenda-Edina-Eguafo-Abirem district (Central Region), Sewfi-Wiawso district (Western Region), Kadjebi district (Volta Region) and Bongo district (Upper East Region). A total of 180 community-based growth promoters have been trained in these districts.

Table 4.5: Donor-support for Health including HIV/AIDS (2003-2007) 2003 2004 2005 2006 2007 Total (US $m) 104.31 174.10 180.18 165.12 150.61 Credit (US $m) 14.47 38.06 43.96 8.82 17.78 Grant (US $m) 89.84 136.03 136.21 156.30 132.83 World Bank 18.06 55.42 63.38 6.36 12.16 ADB 0 0 0 3.87 4.02 EU 1.04 0.64 5.77 0.15 0 Denmark 10.29 15.23 10.43 9.14 7.49 Japan 2.00 4.12 7.72 6.53 5.80 Netherlands 11.33 16.18 4.81 28.65 25.34 Nordic Devp 0 0 1.04 0.11 1.60 Fund Spain 0 15.00 0 0 3.02 UK 22.95 23.48 25.74 36.22 20.68 US 20.46 21.12 21.06 16.68 32.39 ILO 0 0.20 0.20 0.10 0 IOM 0 0 0.01 0.01 0 UNAIDS 0.57 0.41 0.28 0.26 0.26 UNFPA 3.50 3.50 3.50 4.50 2.96 UNICEF 4.14 7.57 8.18 18.30 6.40 UNDP 0.24 0.44 0.30 0.53 3.05 WFP 0.28 1.63 0.90 0.85 1.22 WHO 6.32 5.08 5.68 6.43 6.43 Global Fund 3.15 4.08 21.15 26.16 17.80 Source: Field data, 2007

Table 4.6: Donor support for Water and Sanitation (2003-2007) 2003 2004 2005 2006 2007 Total (US $m) 53.07 44.81 59.21 75.15 84.62 Credit (US $m) 13.55 11.83 10.89 16.34 29.63 Grant (US $m) 39.52 32.99 48.33 58.81 54.99 World Bank 12.66 11.31 20.50 9.26 25.64 ADB 0 0 0.52 6.22 8.47 EU 1.05 0.75 8.90 0 16.00 Canada 1.31 1.30 1.73 2.95 1.58

34

A132 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Denmark 15.57 14.01 11.60 11.94 12.61 France 0.36 1.68 2.16 6.60 0.01 Netherlands 6.15 8.24 7.86 20.73 0 Nordic Devp 0.80 0.08 0.40 1.53 6.13 Fund Spain 10.00 0 0 0 3.02 UK 0.22 0.84 0.73 10.90 3.66 US 0.92 1.47 1.34 1.65 1.30 Source: Field data, 2007

4.5 World Health Organization The World Health Organization (WHO) provides funding through the pooled funding mechanism at the national level. Over the period 2003-2007 in addition to the funding at national level, the WHO provided funding for some district-specific interventions; EPI, Outreach (child welfare clinics) and Health System strengthening (review meetings, supervision and training) in selected districts.

Table 4.7: WHO Health and Development Programmes (2003-2007) Year Intervention Activity Indicator Region District 2003 CH EPI, Child Welfare Clinic Brong-Ahafo Sene, (1) Health Training, Monitoring, system supervision, micro planning Volta (4) Krachi, strengthening North Tongu, Kpando, Jasikan

Ashanti (1) Sekyere East

Western (1) Mpohor- Wassa West

Eastern (1) Kwahu North 2004 CH EPI, Child Welfare Clinic Volta (4) Krachi, North Tongu, Health Training, Monitoring, Jasikan, system supervision, micro planning Kpando strengthening Brong-Ahafo Nkoranza (1)

Northern (2) Bole, Nanumba

Ashanti (2) Amansie East, Sekyere East

Eastern (1) Kwahu North 2005 CH EPI, Child Welfare Clinic Volta (4) Kpando,

35

IIP-JHU | Retrospective evaluation of ACSD in Ghana A133 Health and development programs and policy mapping 1996­2007

Krachi West Health Training, Monitoring, Jasikan, system supervision, micro planning Krachi East strengthening Ashanti (3) Kwabre Amansie East, Sekyere East

Northern (2) Bole, Sawla-Tuna- Kalba

Brong-Ahafo Techiman, (2) Sunyani 2006 CH EPI, Child Welfare Clinic Northern (2) Bole, Sawla-Tuna- Health Training, Monitoring, Kalba system supervision, micro planning strengthening Volta (4) Adaklu- Anyigbe South Dayi, Krachi East, Tain

Ashanti (1) Atwima- Nwabiagya 2007 CH EPI, Child Welfare Clinic Volta (3) South Dayi, Jaskan, Health Training, Monitoring, Kpando system supervision, micro planning strengthening Ashanti (2) Ahafo-Ano South

Eastern (1) Manya- Krobo

Greater Accra (3) Source: Field data, 2007

36

A134 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

4.6 Opportunities Industrialization Centres International (OICI) OICI is one of the major implementing agencies for USAID-Ghana. OICI with support from Counterpart International (US-based NGOs) undertook some reproductive and child health programs from 2004 to 2006.

Table 4.8: OICI Health and Development (2003-2006) Year Intervention Activity Indicator District Community 2004 MCH Nutrition, training, BCC Savelugu- Nanton, Tolon- Kumbungu

Micro credit Financial support to women in Savelugu- agriculture Nanton, Tolon- Kumbungu

CH Community-based growth Savelugu- promotion Nanton, Tolon- Kumbungu 2005 MCH Nutrition, training, BCC Savelugu- Nanton, Tolon- Kumbungu

Micro credit Financial support to women in Savelugu- agriculture Nanton, Tolon- Kumbungu

CH Community-based growth Savelugu- promotion Nanton, Tolon- Kumbungu 2006 MCH Nutrition, training, BCC Savelugu- Nanton, Tolon- Kumbungu

Micro credit Financial support to women in Savelugu- agriculture Nanton, Tolon- Kumbungu

CH Community-based growth Savelugu- promotion Nanton, Tolon- Kumbungu Source: Field data, 2007

37

IIP-JHU | Retrospective evaluation of ACSD in Ghana A135 Health and development programs and policy mapping 1996­2007

4.7 Department for International Development (DFID) DFID of the United Kingdom has supported interventions of the GHS at the national level; DFID provided UK £ 35 000 000 in the form of financial assistance and UK £ 5 000 000 through technical assistance from 2002 to 2006. In 2006 DFID supported the purchase and distribution of ITNs with UK £ 6 000 000. In 2007 DFID spent UK £ 2 500 000 to purchase and distribute ITNs. Over 2007-2011 DFID plans to support the health sector with UK £ 50 000 000. DFID funds health and development activities through the Netherlands embassy in Ghana.

4.8 Danish International Development Agency-Health Sector Support Office The Danish Government has supported the health sector in Ghana through the Health Sector Program Support run by the Danish International Development Agency (DANIDA). DANIDA support has been packaged into 5-year phases which began in 1993. The data available covers phase III (HSPS III) of DANIDA’s support to the health sector which spans the period 2003- 2007.

The earmarked funding under the Danida Health Sector Programme Support has continued to provide technical and financial assistance to areas of critical importance to the success of POW II but which are difficult to implement or are at risk of being side-lined in a resource constrained environment.

Right from the start of HSPS III, attempts were made to channel ear-marked funds through the so-called “aid pool account”. Serious delays in the transfer of funds had negative impact on the implementation of planned activities and subsequently most activities were funded directly from the HSSO.

Table 4.9: DANIDA health support in 000 DKK (2003-2007) Year Intervention Activity Indicator Region District 2003 Improving access Exemptions for the poor 3.52 National to health system Risk sharing arrangements

Strengthening Management capacity 3.52 district health system Quality of care

Enhancing partnerships

Incorporating key issues on district agenda

Central level Regulation 0.23 initiatives Financial management

Policy development 2004 Improving access Exemptions for the poor 5.17 National to health system Risk sharing arrangements

38

A136 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Strengthening Management capacity 5.83 district health system Quality of care

Enhancing partnerships

Incorporating key issues on district agenda

Central level Regulation 2.31 initiatives Financial management

Policy development 2005 Improving access Exemptions for the poor 5.17 National to health system Risk sharing arrangements

Strengthening Management capacity 6.60 district health system Quality of care

Enhancing partnerships

Incorporating key issues on district agenda

Central level Regulation 2.31 initiatives Financial management

Policy development 2006 Improving access Exemptions for the poor 1.76 National to health system Risk sharing arrangements

Strengthening Management capacity 5.50 district health system Quality of care

Enhancing partnerships

Incorporating key issues on district agenda

Central level Regulation 1.76 initiatives

39

IIP-JHU | Retrospective evaluation of ACSD in Ghana A137 Health and development programs and policy mapping 1996­2007

Financial management

Policy development 2007 Improving access Exemptions for the poor 1.762 National to health system Risk sharing arrangements

Strengthening Management capacity 3.19 district health system Quality of care

Enhancing partnerships

Incorporating key issues on district agenda

Central level Regulation 1.76 initiatives Financial management

Policy development Source: Field data, 2007

Table 4.10: DANIDA Health Funds for HIRD 2006 Year Intervention Activity Indicator Region District 2006 Funding for Implementation of HIRD US $1.360m Upper West Region-wide maternal and child interventions health

US $ 1.275m Northern Region-wide

US $ 0.8075m Upper East Region-wide

US $ 0.8075m Central Region-wide Source: Field data, 2007

4.9 Engender Health Engender Health, an implementing partner of USAID, has a five-year project (June, 2004-May, 2008) on child health targeting 28 most deprived districts (USAID target districts) in seven regions, excluding the three northern regions. The components of the programme are: • Child survival focusing on the three components of IMCI but the community component is integrated into CHPS; • National level support especially in the development of standards and protocols for quality improvement; and • Regional level support, including capacity building, monitoring and supervision, provision of equipment and minor renovation of buildings. The project is demand-driven and intended to be aligned to the needs of beneficiaries.

40

A138 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

5.0 CHANGES IN HEALTH POLICIES IN GHANA

This section includes data on global and regional policies (MDGs 4 and 5, NEPAD and Abuja declaration) national policies (Ghana Vision 2020, Ghana Poverty Reduction Strategy, Medium Term Health Strategy), heath sector policies (5-year programmes of work, child health policy, reproductive health policy, drug policy, National Health Insurance, community-based health planning and services), and interventions/programs (Roll Back Malaria, EPI, HIRD, SMI, IMCI, HIV/AIDS and others.

5.1 Global and Regional Policies

5.1.1 Millennium Development Goals In 2000, a millennium summit was held under the auspices of the United Nations (UN) at which representatives from 189 countries committed themselves to sustaining development and eliminating poverty. They set goals, targets and indicators to measure progress towards achieving these goals by 2015. These goals became known as the Millennium Development Goals (MDGs). Of the 48 indicators, 18 are directly related to health. The reference year is 1990 and Ghana is committed to achieving these goals.

To achieve the MDG 4 Ghana has to reduce U5MR from 111/1000 LB in 2003 to at least 40/1000 LB by 2015. More importantly, the rate of reduction should be at least the same as the rate of fall between 1985 and 1990. Similarly the MMR has to be reduced from 214/100 000 LB to 54/100 000 LB. At the current rate of reduction Ghana is unlikely to achieve MDGs 4 and 5.

5.1.2 International Conference on Population and Development (ICPD) This conference was held in September 1994 in Cairo, Egypt. It led to the finalization of a programme of action in the area of population and development for the following 20 years. The 20-year goals were spelt out in four related thematic areas; universal primary education before 2015, reduction of infant and child mortality below 35 per 1000 LB and 45 per 1000 LB respectively by 2015, reduction of maternal mortality to levels where they no longer constitute a public health problem and access to the complete range of sexual and reproductive health services through the primary health care system by 2015.

5.1.3 Bamako Initiative The Bamako Initiative was to commit nations to implement strategies designed to increase essential drugs’ availability and other health care services for sub-Sahara Africa.

5.1.4 Abuja Declaration In April 2000, an African Summit on Roll Back Malaria was held in Abuja, Nigeria. Forty-four of the fifty malaria-affected countries in Africa were present. The nations committed themselves to the principles and targets of the Harare Declaration of 1997 and to initiate appropriate and sustainable action to strengthen health systems to ensure the achievement of certain targets particularly related to malaria. Additionally they were to commit at least 15% of their GDP to health. In 2005 and 2006 Ghana committed 14% of GDP to health.

5.1.4 New Partnership for Africa’s Development African heads of states and presidents have pledged themselves to the duty of poverty eradication on the continent. This pledge is captured in the New Partnership for Africa’s Development (NEPAD). The leaders recognize the urgent need to place African countries on a path of sustainable growth and development while participating actively in the world economic 41

IIP-JHU | Retrospective evaluation of ACSD in Ghana A139 Health and development programs and policy mapping 1996­2007 and body politic. The goals outlined in NEPAD are the same as MDGs but add a goal to achieve and sustain an average GDP growth rate of over 7% per anum for the next 15 years. Ghana is signatory to NEPAD.

5.2 National Health Policies

5.2.1 Ghana Vision 2020 In 1995, Ghana launched a programme of economic and social development policies dubbed “Ghana -Vision 2020”. This 25-year programme had as its long term goal to transform the country’s economy from its present low-income status to that of a middle-income country by the year 2020. In order to realize this vision, Vision 2020 looked at the 1990s level of social and economic development as a benchmark against which future progress would be measured. In addition, a medium term (1996-2000) objective to consolidate the foundations for accelerated economic and social development in the long term was also launched. Vision 2020 had a health component that sought to improve the health status of all Ghanaians through well articulated strategies.

5.2.2 Medium Term Health Strategy Based on the Vision 2020 document, the Ministry of Health (MoH) Ghana, published a Medium Term Strategic document in September 1995, which detailed the development of the health sector (health sector reform programme) in the medium term. To operationalize the Medium Term Strategy Health Strategy (MTHS), MoH in consultation with the donor community, regional and district level health management, identified the key medium term objectives set out in the Strategy for their achievement. These formed the basis of the first in the series of Health Sector 5 Year Programme of Work (5YPOW I) which covered the period 1997-2001. Since then a POW II has been implemented and currently Ghana is into the third in the series; POW III which will span the period 2007-2011.

Each POW has set strategic objectives, targets to be achieved and sector-wide indicators to measure progress. At the end of each year an assessment of the overall performance of the health sector is undertaken and recommendations are put forward with the view to facilitate the achievement of the strategic objectives.

5.2.3 Ghana Poverty Reduction Strategy I and II The Ghana Poverty Reduction Strategy (GPRS I and II) is the Government of Ghana’s medium term strategy for national development. The GPRS is the tool to ensure sustainable and equitable growth, accelerated poverty reduction and protection of the vulnerable and marginalized within a decentralized and democratic milieu.

For child health the GPRS places emphasis on the delivery of cost effective and high impact interventions to reduce U5MR particularly in the four regions with the poorest indicators; Central, Northern, Upper West and Upper East regions.

The POW represents the health sector’s response to the GPRS and aims to bridge the inequalities in health in Ghana

42

A140 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

5.2.4 Ghana Health Service The establishment of the Ghana Health Service (GHS) as the operational wing of the MoH was one of the major pillars acknowledged in the health sector reform process as described in the Medium Term Health Strategy. The GHS was established under Act 525 of 1996 and it was envisaged that its establishment and subsequent operationalization would contribute to the delivery of a more equitable, accessible, efficient and responsive health system.

5.3 Health Sector Policies

5.3.1 Child Health Policy In 1999 the MoH developed a comprehensive child health policy that targeted the management of the six leading causes of mortality; malaria, pneumonia, measles, malnutrition, diarrhoea and neonatal deaths. The Policy envisaged the reduction of U5MR from 66/1000 LB in 1997 to 50/1000 LB in 2001. The Policy identified five priority areas of action; improving neonatal health care, prevention and control of growth and nutritional problems, prevention of and control of infectious diseases and injuries, clinical care of the sick and injured child, and health related interventions. The Policy was developed before the MDGs were set and is currently under review to meet the challenges of the MDGs.

5.3.2 Reproductive Health Policy This policy was first published in 1996. Its 2003 edition (2nd edition) has been reviewed but the broad contents remain the same.

The components of reproductive health care services in Ghana are; • Safe motherhood (antenatal, safe delivery, post-natal care including breast feeding and infant health) • Family planning • Prevention and management of unsafe abortion and post-abortion care • Prevention and management of reproductive tract infections including sexually transmitted diseases and HIV/AIDS • Prevention and management of infertility • Prevention and management of cancers of female and male reproductive system including the breast • Responding to concerns about meno and andropause • Discouragement of harmful traditional practices and gender-based violence that affect the reproductive health of men and women • Information and counselling on human sexuality, responsible sexual behaviour, responsible parenthood, pre-conceptual care and sexual health

5.3.3 National Drug Policy In 1992 Ghana operated a revolving drug fund (influenced by the Bamako Initiative) using capital that had accumulated in health facilities through fees retention during the previous year. A ‘cash and carry’ manual written in 1989 provided some guidelines on the operational aspect of the Fund. Ghana revises its national drug policy, essential drug list and standard treatment guidelines regularly to meet current demands. The latest revision was in 2004. Most of the drugs for the management of common childhood illnesses are found in the List. The Policy does permit the use of antibiotics by community-based agents to manage uncomplicated ARI in children under-five.

43

IIP-JHU | Retrospective evaluation of ACSD in Ghana A141 Health and development programs and policy mapping 1996­2007

An anti-malaria drug policy for Ghana was published with support from the Global Fund and National Malaria Control Programme (NMCP) in 2004. This Policy introduced the use of ACT and sulphadoxine-pyrimethamine for the management of uncomplicated malaria and IPTp respectively.

5.3.4 The National Health Insurance Scheme In the 60s a limited number of fees for specific hospital services were introduced. In 1985 GOG introduced user fees for all medical services except specific communicable diseases. This was complemented with a full cost recovery for drugs as a way of generating revenue to address drug shortages.

Initially (1998) in order to reduce the financial barrier to accessing health services Ghana operated an exemption policy for children under-five, pregnant women, the elderly and the poor. Inadequate and very slow reimbursement limited its effectiveness.

In 2003, Act 650 was passed to govern the establishment of the National Health Insurance Scheme (NHIS). Nearly 75 districts were supported to set up district-wide mutual health insurance schemes and to initiate activities to recruit and register clients. It is envisaged that by 2009 every resident of Ghana should belong to a health insurance scheme. The NHIS is based on equity, cross-subsidization, quality of care and community ownership. Children under-18 are automatically covered if parents have paid at least the minimum contribution. Ghana has experienced one of the fastest growing national insurance schemes worldwide in terms coverage.

5.3.5 Community-based Health Planning and Services Community-based Health Planning and Services (CHPS) is a strategy to provide cost-effective and adequate quality basic primary health services to individuals and households at the community level through engagement of the communities in the planning and delivery of services. In began as a research project in the Kassena-Nankana district of Upper East region which sought to address inequalities in the health system by mobilizing both community and health services resources. In the medium term, MoH plans to deploy 1570 community health officers (community health nurses) to various communities nation-wide by 2006. The idea was that 80% of districts in Ghana would have completed CHPS implementation by 2006. Nation- wide scale up began in 1998. The scale-up is far behind schedule.

5.4 Health Interventions and Programmes

5.4.1 Safe Motherhood Initiative Safe motherhood-making pregnancy safer was adopted by Ghana in 1987. More than two decades after launching SMI, maternal mortality is still a major public health problem in Ghana.

5.4.2 Traditional Birth Attendants The concept of TBAs has been in Ghana for ages. In 1977, fifty-seven (57) TBAs were identified and trained in a rural community near Accra. In 1989 USAID (five regions), UNICEF three regions) and UNFPA (one region) sponsored a TBA training programme intended to institutionalize national standardized training of TBAs in all regions of Ghana. TBA training has been implemented under a number of different health projects for decades.

44

A142 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

5.4.3 Roll Back Malaria In Ghana malaria is the most common cause of morbidity and mortality in children under-five. In 1999 Ghana adopted the Roll Back Malaria (RBM) strategy to control malaria. The main pillars of action were; use of insecticide treated nets (ITNs) and materials, intermittent preventive treatment (IPTp) of malaria in pregnant women, effective management of cases and home- based management of fevers. In 2003, Ghana received support from the Global Fund to fight AIDS, tuberculosis and malaria (GFATM) to energize her roll back malaria activities. Twenty (20) districts were selected nationwide to pilot the Global Fund initiative. An important component of the initiative was the training of community-based agents who promoted malaria control at the household level through health education, sale and retreatment of ITNs and referral facilitation. In some districts these agents were trained to monitor and report adverse effects in pregnant women who had received IPTp using sulphadoxine pyremethamine at the facility level. Currently Ghana uses ACT for the treatment of uncomplicated malaria.

Substantive ITN distribution began in 2003 gradually increasing from about 150 000 nets annually to over half a million in 2005. Late in 2006 the first large scale of long-lasting insecticide treated nets was undertaken with assistance from UNICEF and DFID; over 2.1 million nets were distributed free of charge as part of the integrated measles/polio/vitamin A/ITN distribution campaign. The Global Fund now covers all districts in Ghana.

5.4.4 Expanded Programme on Immunization The Expanded Programme on Immunization (EPI) was introduced in 1978. Since 1985 it is operational in all districts and focused on immunization against tuberculosis, diphtheria, neonatal tetanus, pertussis, acute poliomyelitis, measles and yellow fever. It took the form of mass immunization till 1999 when the weaknesses of this approach were observed. A mix strategies including; national immunization days (NID) daily immunization services (DIS), child health promotion week, outreach, mop up, visit to island and lake communities (extremely hard to reach), reaching every district (RED), monitoring for action, supportive supervision, support to districts and addressing system-wide barriers were used to improve coverage. Immunization against Haemophilus influenza type B (Hib) and hepatitis B (HepB) were introduced in 2002.

Table 5.1: Immunization coverage in Ghana 1997-2006 (%) Antigen 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 BCG 70 77 85 94 91 97 93 92 100 100 OPV3 X X X 82 80 79 76 76 85 84 Measles 57 67 71 84 82 85 80 78 83 85 Yellow fever 41 41 64 74 76 71 73 76 82 84 TT2+ X X X 73 61 68 66 62 71 69 DPT/Penta 3 56 68 73 84 76 79 76 76 85 84 Source: EPI Ghana, 2007

In 2004 three rounds of maternal and neonatal tetanus campaigns were undertaken in 13 districts, in 2005 a similar campaign was organized in 27 districts and in 2006 another 27 districts benefitted from the campaign.

In 2005 a supplementary measles immunization activity (a catch-up campaign) was carried out nationwide for children aged 9 months – 15 years. In 2006 a follow-up measles campaign was carried out nationwide as part of an integrated measles/polio/vitamin A/ITN distribution campaign for children 9 months – 15 years.

45

IIP-JHU | Retrospective evaluation of ACSD in Ghana A143 Health and development programs and policy mapping 1996­2007

Table 5.2 Integrated Measles/Polio/Vitamin A/ITN Distribution Campaign 2006 Antigen/service provided Coverage (%) Measles 79 Polio 84.6 Vitamin A 84.3 ITN distribution 92.3 Source: EPI Ghana, 2007

National Immunization Days have been observed in Ghana for years. In 2005, four (4) rounds of synchronized polio NIDs were organized nationwide in two phases. Vitamin A supplementation was integrated in two (2) rounds of the polio NIDs while deworming was added to the campaign in Northern and Upper East regions during the two phases.

Table 5.3: National Immunization Days 2005 PHASE I Round Antigen/service Coverage (%) Round 1 OPV 104.8 Mebendazole 101.7 Round 2 OPV 104.4 Vitamin A 94.5 PHASE II Round Antigen/service Coverage (%) Round 1 OPV 103.8 Vitamin A 101.3 Round 2 OPV 107.9 Mebendazole 104.9 Source: EPI Ghana, 2007

The Global Alliance for Vaccine Initiative (GAVI) has been supporting EPI in Ghana.

Table 5.4: GAVI Immunization Financing (2006-2010) US$ 2006 2007 2008 2009 2010 Finance 4 466 413 6 893 500 11 506 211 11 803 898 12 092 671 Source: EPI Ghana, 2007

5.4.5 High Impact Rapid Delivery Interventions The high-impact, rapid-delivery and sustainable approach is a strategy for scaling-up maternal and child survival interventions in Ghana. The approach is based on the, MoH’s CHPS model, the IMCI strategy, Safe Motherhood Initiative and the ACSD approach. These four approaches have several common elements and complement one another. Some of the common elements are a focus on primary level high-impact, cost-effective interventions that address major causes of childhood deaths, use of community development approaches to extend service delivery rapidly, broad partnerships, and extensive planning at the micro level. In line with recommendations from the Ghana Health Summit (2004) an inter-sectoral approach will be adopted for strengthening the capacity of communities by using sound communication strategies and involving other ministries departments and agencies (such as Department of Community Development, NGOs and CBOs) that have comparative advantage in this area.

46

A144 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

5.4.6 Integrated Management of Childhood Illness In 1998, Ghana adopted IMCI as a key strategy for reducing U5MR and developed a strategic plan (2002-2006) with a goal to reduce morbidity and mortality due to major causes of diseases in the under-five and ensure healthy growth and development of children. The objectives of the policy are: • 60% of first level health facilities have at least one staff trained in IMCI; • 80% of prescribers correctly prescribe anti-malaria drugs for U5 children; • 60% of health workers correctly assess children for danger signs; • 80% of health workers correctly assess children for three main symptoms of cough, diarrhoea and fever; and • 50% of mothers/care givers of U5 children reporting to health facilities know the three rules of home care: give extra fluids, continue feeding and when to return.

A national IMCI orientation meeting recommended that IMCI should provide technical and advocacy support. IMCI was initially piloted in four districts (Ga, Atwima-Nwabiagya, Manya- Krobo and Tolon-Kumbungu) and planned to be scaled up to 30 districts by 2004, 60 districts by 2005 and 90 districts by 2006. IMCI is one of the priority interventions identified under the second POW - 2002 – 2006.

Collaboration between IMCI and Roll Back Malaria (RBM) started at the African Regional level in 1996, and has since expanded to operations at country level. In Ghana, there has been collaboration between the two programmes in case management training, home-based care, and Information, Education and Communication (IEC) among others. In November 2001, the MoH/GHS took the RBM-IMCI partnership a step further by involving other programmes – Expanded Programme on Immunisation and Integrated Disease Surveillance and Response (IDSR). The MOH/GHS in collaboration with WHO developed a proposal to integrate service provision, monitoring and evaluation of these interventions in 10 selected districts. These districts have therefore been designated as the districts of focus for the programme interventions in the IMCI, Malaria (RBM), EPI and IDSR. At present 62 out of the 138 districts have at least one health staff trained in IMCI, which falls short of the World Health Organisation’s requirement that 60% of all prescribers from 80% of districts should be trained in order to make an impact.

5.4.7 Community-based Growth Promotion Community-based growth promotion was piloted in three districts (Tolon-Kumbungu, Atwima and Manya Krobo) in 2001. The World Bank, GSK, WVI and Plan Ghana have supported the implementation of CBGP in 40 other districts. The initiative is to be scaled up in 65 districts as part of the Nutrition and Malaria Control for Child Survival Project.

5.4.8 Community-based Surveillance The community-based surveillance system was piloted in the Northern region in 1988 as an expansion of the village volunteers surveillance system (of the 1970s) as part of the Guinea worm eradication programme. The System has seen many modifications in different regions of the country and some of the regional variations are observed in terms of actual coverage, quality surveillance, supervision of volunteers and the use of data generated by the volunteers.

47

IIP-JHU | Retrospective evaluation of ACSD in Ghana A145 Health and development programs and policy mapping 1996­2007

5.4.9 Nutrition and micronutrient deficiency Priority actions undertaken to control under- and malnutrition are increased awareness on nutrition to improve child feeding practices, promotion and management of the malnourished child and prevention of micronutrient deficiencies such as iron, vitamin A and iodine deficiency. A number of initiatives have been introduced: • Infant and young child feeding strategy; • Community based nutrition and food security project; • Supplementary feeding programme; • Iodine deficiency disorders control programme; and • Vitamin A deficiency control programme.

Table 5.5: Major Child Nutrition-related Projects, Ghana (1988-2010) Funding Source Project Title Amount Implementer World Bank Health Sector Support USD 35m GOG Project (1988-2002) Ghana AIDS Response USD 25m GOG Project (2002-2005) Community-based USD 1.8m GOG Poverty Reduction Project – Nutrition and Food Security Component (1999 – 2003) POW II (2003-2006) USD 57.6m (credit) GOG USD 32.4m (grant) MSHAP (2002-2010) USD 20m GOG Community-based Rural USD 60m GOG Development (2004- 2008) Multi-lateral Micro-nutrient Deficiency UNICEF: 5.85m UNICEF/WFP/WHO/ Agencies Control (2006-2010) GAIN High Impact Rapid GAIN: 1.80m Delivery (2006-2008) Child Survival Bi-lateral Agencies Community-driven CND 12m CIDA Initiatives in Food Security (2005-2010) District Capacity Building CND 5m CIDA Project (2001-2005) School Feeding and USD 8.426m WFP Nutrition Education CBGP GBP 0.221 USAID/GHS/GSK RBM Round 4: USD 38.8m GFATM Round 2: USD 8.8m Source: Field data, 2007

5.4.10 Integrated maternal and child health campaign In 2007, the GHS launched the nationwide Integrated maternal and child health campaign. The target population was pregnant women and children 1 year or below. The Campaign involved

48

A146 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

TT and long-lasting insecticide-treated nets (LLINs) for pregnant women while children 1 year or below received; OPV, dewormer (children 24-59 months), vitamin A (children 6-59 months), ITN and birth registration. Lactating mothers (up to eight weeks post-partum) received vitamin A supplementation if they had not already received it.

5.4.11 Child Health Promotion Week The Child Health Promotion Week Celebration (CHPW) was instituted in 2004 by the Ghana Health Service to create and sustain awareness of the many services available in the health system to promote the healthy growth and prevention of common childhood conditions in children under five years. It was also to sensitize the general public on the importance of Births and Deaths Registration. In view of this, the Ghana Health Service dedicated the second week of May every year for the celebration of CHPW. This celebration has contributed immensely in creating awareness of the many services available as well as improving access to these services. It contributed to improving routine EPI coverage.

Table 5.6: Health policies and programs in Ghana Year Health policy Target Implemented in Ghana? Global/Regional Health Policies 1987 Bamako Initiative Population-wide Yes 1994 International Conference on Population and Reproductive health Yes Development Cairo 2000 Abuja Declaration Population-wide Yes 2000 Millennium Development Goals 4 and 5 Maternal and child Yes health 2001 New Partnership for Africa’s Development Population-wide Yes (NEPAD) National Health Policies 1995 Ghana Vision 2020 Population-wide 1995 Medium Term Health Strategy Population-wide 1996 GHS Act 525 Population-wide 2002-2005 GPRS I Population-wide 2006-2009 GPRS II Population-wide Health-sector policies 1997-2001, 5-year programmes of work I, II, III Population-wide 2002-2006, 2007-2011 1998 CHPS Population-wide

49

IIP-JHU | Retrospective evaluation of ACSD in Ghana A147 Health and development programs and policy mapping 1996­2007

1996 Reproductive health Population-wide 1989 National Drug Population-wide 2003 National Health Insurance Scheme Population-wide Health Interventions/programmes 1987 Safe Motherhood Initiative Maternal health Traditional Birth Attendants Maternal health 1999 NMCP-Roll Back Malaria Maternal and child health 1978 Expanded Programme on Immunization Child health 2004 High Impact Rapid Delivery Maternal, neonatal and child health 1998 IMCI Child health 2000 CBGP Child health 1970 Community-based surveillance Population-wide 1998 Integrated disease surveillance and response Child health 2000 PMTCT/VCT Maternal and neonatal health Nutrition and micronutrient deficiency Child health 2007 Integrated maternal and child health campaign Maternal and child health 2004 Child health promotion week Child health

Source: Field data, 2007

50

A148 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

6.0 Conclusion Ghana has been implementing almost the full range of cost-effective evidence-based maternal, neonatal and child health interventions and this combined with an increasing expenditure on health should have resulted in improved maternal and child health indicators for the country. If there is evidence to demonstrate that the child health indicators for Upper East region are better than the rest of the country and that the difference is significantly attributable to the ACSD intervention then the health systems for delivery of the interventions evidenced to reduce the morbidity and mortality burden in the rest of Ghana should be revised to include systems that would increase coverage to 90-99%. It would prudent to increase coverage so that those who need the interventions most (the poor and vulnerable) and who ultimately determine the rate of improvement in maternal and child health indicators are adequately covered so that Ghana can achieve MDGs 4 and 5.

51

IIP-JHU | Retrospective evaluation of ACSD in Ghana A149 Health and development programs and policy mapping 1996­2007

Appendix 1: SELECTED REVIEWED DOCUMENTS

1. DANIDA (2006) Health Sector Programme Support Annual Report 2005 2. DANIDA (2007) Health Sector Programme Support Annual Report 2006 3. Ghana Health Service (2006) Expanded Programme on Immunization Annual Report 2005. Accra 4. Ghana Health Service (2007) Expanded Programme on Immunization Annual Report 2006. Accra. 5. Ghana Red Cross Society-Upper East Region (2001) Annual Report 2000 6. Ghana Red Cross Society-Upper East Region (2002) Report on Child Survival Project 2000-2001 7. Ghana Red Cross Society-Upper East Region (2003) Annual Report 2002 8. Ghana Red Cross Society-Upper East Region (2004) Annual Report 2003 9. Ghana Red Cross Society-Upper East Region (2006) Annual Report 2005 10. Ghana Red Cross Society-Upper East Region (2007) Annual Report 2006 11. GHS (2002) Annual Review 2001 12. GHS (2003) Programme of Work 2004 13. GHS (2003) Reproductive and Child Heal Unit Annual Re port 2002 14. GHS (2003) Review of Health Sector Programme of Work 2002 15. GHS (2004) Main Sector Review 2003 16. GHS (2004) Policies and Priorities for 2005 17. GHS (2004) Review of Health Sector Programme of Work 2003 18. GHS (2005) Community-based Surveillance in Ghana 19. GHS (2006) Annual Report 2005 20. GHS (2006) Expanded Programme on Immunization Annual Progress Report 21. GHS (2006) Facts and Figures 2005 22. GHS (2006) Programme of Work 2007 23. GHS (2007) Review of Health Sector Programme of Work 2006 24. GHS (2007) Upper East Region Annual Report 2006 25. GHS 5-year Programmes of Work (I II and III) 26. GOG Ghana Poverty Reduction Strategy I and II 27. GSS/MOH/ORC Macro (2003) Ghana Service Provision Assessment Survey 2002 28. GSS/NMIMR/ORC Macro (2004) Ghana Demographic and Health Survey 2003. Calverton, Maryland 29. JICA (2007) Health Interventions in Ghana 1999-2006 52

A150 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

30. MOH (1995) Medium Term Health Strategy 31. MOH (1999) Policy and Strategies for Improving the health of Children under-five in Ghana 32. MOH (2000) IMCI Strategic Plan for Ghana 2002-2006 33. MOH (2000) Roll Back Malaria Strategic Plan for Ghana 34. MOH (2006) C-IMCI Annual Report 2005 35. MOH (2006) Directory of Policies, Standards and Guidelines 36. Navrongo-Bolgatanga Diocese (2001) Annual Report 2000 37. Navrongo-Bolgatanga Diocese (2002) Annual Report 2001 38. Navrongo-Bolgatanga Diocese (2003) Annual Report 2002 39. Navrongo-Bolgatanga Diocese (2004) Annual Report 2003 40. Navrongo-Bolgatanga Diocese (2005) Annual Report 2004 41. Navrongo-Bolgatanga Diocese (2006) Annual Report 2005 42. Navrongo-Bolgatanga Diocese (2007) Annual Report 2006 43. The Abuja Declaration (extract from the Africa Summit on RBM Abuja 2000 44. UNDP (2005) MDG Report Ghana 2004 45. UNICEF/USAID/MICS/GSS (2007) Multiple Indicator Cluster Survey Preliminary Report 46. WHO (2002) First Two Years of IMCI Implementation in Ghana 47. WHO (2004) IMCI documentation; progress, experiences and lessons learnt 48. WHO/UNICEF (2006) Review of National Immunization Coverage Ghana 1980-2005 49. World Bank (2003) Ghana Health Sector Programme Support Project II 50. World Vision International (1998) Bongo ADP Annual Report 1997 51. World Vision International (2000) Bongo ADP Annual Report 1999 52. World Vision International (2001) Bongo ADP Annual Report 2000 53. World Vision International (2001) Bongo ADP Mid-Term Evaluation Report 54. World Vision International (2002 Bongo ADP Annual Report 2001 55. World Vision International (2003) Bongo ADP Annual Report 2002 56. World Vision International (2004) Bongo ADP Annual Report 2003 57. World Vision International (2005) Bongo ADP Annual Report 2004 58. World Vision International (2006) Bongo ADP Annual Report 2005 59. World Vision International (2007) Bongo ADP Annual Report 2006 60. World Vision International (2007) Bongo ADP Profile 2006

53

IIP-JHU | Retrospective evaluation of ACSD in Ghana A151 Health and development programs and policy mapping 1996­2007

Appendix 2: Summary results from Focus Group Discussions of mothers in five communities in Bongo district

1. What proportion of children under-five do you think receive the following interventions or experience the following? a. Vitamin A capsules (80%) b. Exclusive breast feeding (30%) c. Iodized salt (2%) d. Pneumonia treatment with antibiotics (60-70%) e. Measles vaccination (80%) f. Oral rehydration Salt for diarrhoea (60-70%) g. Penta 3 vaccination (60%) h. Skilled attendance at birth (70%) i. Die before the first month of life (20%) j. Die before the first year of life (30%) k. Sleep under ITN (70%) l. Are low birth weight babies (40%) m. Have access to clean drinking water (20%)

2. What proportion of mothers do you think receive or know the following?

a. At least 2 TT injections before delivery (60-70%) b. Receive IPT (70%) c. At least two danger signs of pregnancy (20%) d. At least two danger signs of newborns (90%) e. At least four danger signs for children under-five (90%) f. Proportion of women accompanied by their husbands to ANC (20%) g. Information on birth preparedness (100%) h. Practice family planning (60-70) i. Sleep under ITN (70%) j. Deliver through caesarean section (50%) k. Bathe their newborns within 24 hours of delivery (80%)

54

A152 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Appendix 3: INTERVIEW GUIDES

HEALTH PROGRAMMES AND POLICY MAPPING EXERCISE IN UPPER EAST REGION AND THE REST OF GHANA

SEMI-STRUCTURED INTERVIEW GUIDE

RESPONDENT: NOGOs/CSOs/FBOs IN MATERNAL AND CHILD HEALTH IN UPPER EAST REGION

1. What type of health related interventions are you into? 2. What specific types of your interventions relate to maternal and child health?

3. Please give a chronological account of these interventions since 1996? 4. In which districts have you been focusing?

5. What are your estimates of coverage for your interventions? 6. What is your relationship with the Ghana Health Service?

7. Who are your major donors? 8. Who are your main collaborators in the field?

9. What are the funding forecasts? 10. Have you in way been involved with the UNICEF funded ACSD in the region?

RESPONDENT: GHS PROGRAMME MANAGERS (EPI, Child Health, RBM)

1. What are GHS’ strategies and priorities in child health? 2. What is the place of ACSD in GHS’ child health policies?

3. What is the role of your programme in child health? 4. What coverage has your programme experienced?

5. What are the achievements and constraints of your programme? 6. Could you share with us copies of reports or documents covering your programme activities since 1997?

55

IIP-JHU | Retrospective evaluation of ACSD in Ghana A153 Health and development programs and policy mapping 1996­2007

Appendix 4: FOCUS GROUP DISCUSSION GUIDE FOR COMMUNITIES IN BONGO DISTRICT, UER

1. What proportion of children under-five do you think receive the following interventions or experience the following;

a. Vitamin A capsules

b. Exclusive breast feeding

c. Iodized salt

d. Pneumonia treatment with antibiotics

e. Measles vaccination

f. Oral rehydration salt for diarrhoea

g. Penta 3 vaccination

h. Skilled attendance at birth

i. Die before the first month of life

j. Die before the first year of life

k. Sleep under ITN

l. Are low birth weight babies

m. Have access to clean drinking water

2. What proportion of mothers receive or know the following;

a. at least 2 tetanus injections before delivery

b. Receive IPT

c. Know at least two danger signs of pregnancy (mention two signs)

d. Know at least two danger signs of newborns (mention two signs)

e. Know four danger signs for children under-five (mention four)

f. Proportion of mothers are accompanied by their husbands to ANC

g. Information on birth preparedness

56

A154 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

h. Practice family planning

i. Sleep under ITN

j. Deliver through Caesarean section

k. Bathe their newborns within 24 hours of delivery

57

IIP-JHU | Retrospective evaluation of ACSD in Ghana A155 Health and development programs and policy mapping 1996­2007

Appendix 5: Dummy table to record health interventions and coverage at national level Year Intervention Activity Indicator Region District

58

A156 IIP-JHU | Retrospective evaluation of ACSD in Ghana Health and development programs and policy mapping 1996­2007

Appendix 6: Dummy table to record interventions and coverage in Upper East Region Year Intervention Activity Indicator District Community

59

IIP-JHU | Retrospective evaluation of ACSD in Ghana A157