Infant Male Circumcision Morbidities and Care Seeking
Infant male circumcision morbidities and care seeking
behaviour in sub-Saharan Africa: a population-based
cross sectional study in rural Ghana
Thomas Kwasi Gyan BA MPH
This thesis is presented for the degree of Doctor of Philosophy of The
University of Western Australia
Division of Paediatrics
Faculty of Health and Medical Sciences
University of Western Australia
2018 Thesis declaration
I, Thomas Kwasi Gyan, certify that this thesis has been substantially accomplished during enrolment in the degree.
This thesis does not contain material which has been accepted for the award of any other degree or diploma in my name, in any university or other tertiary institution. No part of this work will, in the future, be used in a submission in my name, for any other degree or diploma in any university or other tertiary institution without the prior approval of The University of Western Australia and where applicable, any partner institution responsible for the joint-award of this degree.
This thesis does not contain any material previously published or written by another person, except where due reference has been made in the text.
The work(s) are not in any way a violation or infringement of any copyright, trademark, patent, or other rights whatsoever of any person.
The research involving human data reported in this thesis was assessed and approved by The University of Western Australia Human Research Ethics
Committee. Approval number: RA/4/1/7911. Ethical approvals were also obtained from the Ghana Health Service Ethical Review Committee. Approval numbers: GHS-ERC 05/07/11 & 06/07/11 and GHS-ERC: 3, the Institutional
Ethics Committee of Kintampo Health Research Centre. Approval numbers:
2011-03/Extension 1 and 2012-01, and the Research Ethics Committee of
London School of Hygiene and Tropical Medicine. Approval number: A318. ii
Written patient consent has been received and archived for the research involving patient data reported in this thesis.
The work described in this thesis was funded by an Australian Government
Research Training Program (RTP) Scholarship, Kintampo Health Research
Centre, Ghana Health Service and the London School of Hygiene and Tropical
Medicine through the Neovita trial.
This thesis contains published work and a manuscript submitted for publication, all have been co-authored.
Signature:
Date: 28th February 2018
iii
Abstract
Aims: The overall aim of this thesis was to identify the burden and key determinants of morbidity associated with infant male circumcision, and care seeking in a community based setting in sub-Saharan Africa.
Background: Male circumcision services have expanded throughout Africa as part of a long-term HIV prevention strategy. Yet, there have been no studies that have examined the burden and determinants of morbidity associated with infant male circumcision in community-based settings. There is also no evidence about the factors that influence choice of circumcision providers. Data about care seeking patterns once circumcision morbidities are observed are also limited.
Information about the effect of health service provider training in infant male circumcision on circumcision related morbidity also appears largely anecdotal.
To address these gaps, this thesis sought to identify the key determinants of morbidity associated with infant male circumcision and care seeking behaviour in sub-Saharan Africa through a population based cohort study of 3141 male infants in central rural Ghana. Also, in order to inform recommendations to reduce morbidity and improve care seeking, a systematic review of the published literature investigating the effectiveness of health service provider circumcision training was conducted.
Methods: This thesis analysed data from 3141 male infants born in the Brong
Ahafo region of central rural Ghana from 21st May 2012 to 31st December 2012.
Multivariable logistic regression models were used to assess: (i) the burden and iv key determinants of morbidity associated with infant male circumcision; (ii) the effect of key determinants on choice of circumcision provider; and (iii) the health care seeking behaviour of families with infants who had circumcision-related morbidity. In the third study, similar methods were used to analyse data from a larger study of all infants born from 16th August 2010 to 7th November 2011, who experienced acute illness in the same geographic area. In addition, a systematic review was conducted to examine the effect of education and training in infant male circumcision on morbidity and mortality outcomes. Studies were published in standard databases and clinical trial registries in all languages from January
1985 to June 2016.
Results:
(i) Burden and determinants of infant male circumcision-related morbidity: The
results indicated a double burden of high prevalence of infant male
circumcision (93% by 20 days of age) and high risk of morbidity (8.1%)
including two deaths after circumcision. Hygiene-related healthcare
practices such as not washing hands (44.5%), the use of uncleaned
instruments (56.7%) and not cleaning the penile area (66.9%) were very
poor amongst circumcision providers and clearly linked to the
morbidity risk. Not cleaning hands was associated with a twofold
increased risk of morbidity. Socio demographic risk factors such as age
at circumcision and maternal education were clearly linked.
(ii) Choice of circumcision provider: Results showed an independent influence of both socio-economic status and geographic access to health facilities on the
v choice of circumcision provider. The risk of receiving a circumcision from an informal provider increased with each level of deprivation (aOR 1.34, 95% CI
1.25-1.43 p=<0.001) and with the distance that families lived from health facilities (aOR 1.25, 95 CI 1.30–1.38 P = <0.001). The research also found that families with the lowest household income were the most likely to pay for their circumcision (aOR 0.35, 95% CI 0.23-0.53 p=<0.001). Poor families were also most likely to pay additional in-kind contributions (aOR 0.41, 95% CI 0.25–0.67 p=<0.001).
(iii) Health care seeking practices: Only 35 families sought care for circumcision related morbidity. A much lower proportion (15%) of families with infants experiencing circumcision-related problems sought healthcare for the morbidities compared to families of infants with non-circumcision related acute illnesses (88.1%). However, numbers were small and statistical tests could not be performed. In the circumcision study, more families appeared to seek care from a formal (68.6%) than informal (31.4%) health care provider. There were no clear determinants of healthcare seeking for acute illnesses or circumcision related morbidities.
(iv) Systematic review: A total of 1,229 studies were identified. There were no randomised controlled trials (RCT). Only three observational studies satisfied the inclusion criteria; all of which examined the effect of training on the skills and knowledge of medical doctors. No study involved informal circumcision providers. Information on the duration of training required for both medical
vi doctors and informal providers to become competent in infant male circumcision was not available. All included studies were low quality.
Conclusions: This thesis presents the first community level population based study of infant male circumcision in rural Ghana. Given the high morbidity risk and poor care seeking, interventions to reduce morbidity and improve care seeking are urgently needed. The development and rigorous testing of simple training packages to improve education and training of circumcision providers, especially informal non-medical providers in low income countries, are required to improve hygiene practices and reduce circumcision related morbidities.
vii
Table of contents
Thesis declaration...... ii
Abstract ...... iv
Table of contents ...... viii
Acknowledgements ...... xii
Authorship declaration: co-authored publications ...... xiv
Presentations arising from this research ...... xx
List of figures ...... xxi
List of tables ...... xxii
List of abbreviations ...... xxiv
Chapter 1: Introduction ...... 1
1.2 Aims of this research ...... 7
Chapter 2: Literature Review ...... 9
2.1 History and key determinants of male circumcision ...... 9
2.2 Prevalence of male circumcision ...... 10
2.3 Types of circumcision provider ...... 11
2.4 Circumcision methods ...... 13
2.5 Importance and burden of morbidity associated with infant male circumcision ...... 14
2.6 Determinants of morbidity associated with infant male circumcision ...... 16
2.7 Factors that influence choice of health service provider for infant male circumcision ...... 20
2.8 Determinants of care seeking for morbidity ...... 22
viii
2.9 Effect of health service provider education and/or training in infant male circumcision ...... 24
2.10 Policies and guidelines on infant male circumcision ...... 26
Chapter 3: Methods ...... 28
3.1 Study setting ...... 28
3.2 Conceptual framework ...... 37
3.3 Thesis structure ...... 40
Chapter 4. The burden and risk of morbidity associated with infant male circumcision in a community-based setting ...... 43
4.1 Abstract ...... 43
4.2 Introduction ...... 44
4.3 Methods ...... 46
4.4 Results ...... 50
4.5 Discussion...... 62
Chapter 5. Factors which influence choice of infant male circumcision provider in rural Ghana ...... 67
5.1 Abstract ...... 67
5.2 Background ...... 68
5.3 Methods ...... 70
5.4 Results ...... 74
5.5 Discussion...... 86
5.6 Conclusions ...... 90
Chapter 6. Health care seeking behaviour of families ...... 91
ix
6.1 Abstract ...... 91
6.2 Introduction ...... 92
6.3 Methods ...... 94
6.4 Results ...... 99
6.5 Discussion ...... 109
6.6 Conclusions ...... 113
Chapter 7. Effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes: Protocol for systematic review...... 115
7.1 Abstract ...... 115
7.2 Background ...... 117
7.3 Methods ...... 120
7.4 Discussion ...... 129
Chapter 8. Health service provider education and/or training in infant male circumcision to improve short and long term morbidity outcomes: A systematic review ...... 131
8.1 Abstract ...... 131
8.2 Background ...... 132
8.3 Methods ...... 136
8.4 Results ...... 145
8.5 Discussion ...... 156
8.6 Conclusions ...... 161
Chapter 9. Discussion, conclusions and recommendations ...... 163
x
9.1 Main findings in comparison with other epidemiological studies ...... 163
9.2 Strengths and limitations of this research ...... 178
9.3 Future research ...... 181
9.4 Recommendations and implications for policy and practice ...... 183
9.5 Conclusions ...... 185
References ...... 186
Appendix A. Circumcision data collection form ...... 205
Appendix B. PRISMA-P (Preferred Reporting Items for Systematic review and
Meta-Analysis Protocols). Protocol checklist 2015...... 205
Appendix C. MEDLINE search strategy ...... 205
Appendix D. PRISMA-P (Preferred Reporting Items for Systematic review and
Meta-Analysis). Final publication checklist 2017...... 205
Appendix E. MEDLINE search strategy used for the systematic review ...... 205
Appendix F. PhD publications ...... 205
xi
Acknowledgements
This research was supported by an Australian Government Research Training
Program (RTP) Scholarship, Kintampo Health Research Centre, Ghana Health
Service and the London School of Hygiene and Tropical Medicine. This thesis would not have been completed if it was not for the incredible and generous support from my supervisors, Professor Karen Edmond and Dr Kimberley
McAuley. I sincerely thank them for all their time, expertise, friendship and the financial support to undertake this study. I am indebted, in particular to
Professor Karen Edmond for her continued support and dedication for the PhD right from 2010, when I worked as the head of fieldwork of the Neovita trial, a study in which Karen was the lead investigator. I have been fortunate to benefit from Karen’s passion for teaching and training of young researchers and I am extremely grateful for the sound epidemiological, public health and scientific advice she has provided. I also thank Dr Kimberley McAuley for her ongoing support, encouragement, statistical expertise and academic advice throughout the PhD. I also acknowledge and thank Dr. Natalie A. Strobel for her continued support of this study. If it wasn’t for her enthusiasm and support in the systematic review aspect of this research and beyond, it may never have happened.
I thank Associate professor Deborah Lehmann and Dr Anita Van Den Biggelaar of Telethon Kids Institute for the expertise and advice during the initial stages of the PhD.
xii
I would also like to express my gratitude to Dr Sam Newton of the School of
Public Health, Kwame Nkrumah University of Science and Technology, Kumasi
Ghana. I first met Dr Sam Newton when I joined Kintampo Health Research
Centre in 2003, as a Research Assistant on the World Health Organization funded
Expanded Program on Immunization Double Plus project, a study that Sam was the Principal Investigator. This was my first exposure to research. Sam is a mentor and we have work together over many years. His research and academic experience, support and dedication has been an invaluable guide. My gratitude also goes to Dr Maureen O’ Leary and Dr Caitlin Shannon for their support and encouragement.
I am very grateful for the support and encouragement of my Family; Ante Ama,
Evelyn, Zena, Nana Gyan, Millicent and Vida over the years. Special thanks also goes to all Neovita and circumcision staff; data managers, site leaders, site coordinators, field supervisors, fieldworkers, support staff, drivers and all the study participants. I also acknowledge the support of the management and staff of Kintampo Health Research Centre and Ghana Health Service during this period. The staff of Division of Paediatrics, University of Western Australia, provided expert assistance and advice at all times. Special thanks to Associate
Professor Sunalene Devadason and Estelle Dawes.
xiii
Authorship declaration: co-authored publications
This thesis contains works that have been published or submitted for publication in peer reviewed Journals. The authors acknowledge the contribution of each published paper for this thesis and have given approval for this research to be submitted for the degree of Doctor of Philosophy.
Prof Karen M Edmond
Dr Kimberley McAuley
Dr Natalie A. Strobel
Dr Sam Newton
Mr Seeba Amenga-Etego
Prof David Forbes
Prof Seth Owusu-Agyei
Dr Maureen O'Leary
Dr Caitlin Shannon
xiv
Prof Betty Kirkwood
Ms Charlotte Tawiah-Agyemang
1. Gyan Thomas, McAuley Kimberley, Strobel A. Natalie, Shannon Caitlin,
Newton Sam, Tawiah‐Agyemang Charlotte, Amenga‐Etego Seeba, Owusu‐
Agyei Seth, Kirkwood Betty, Edmond M. Karen. Determinants of morbidity
associated with infant male circumcision: community level population-based
study in rural Ghana. Trop Med Int Health 2017; 22(3): 312-22.
Location in thesis: chapter 4 (Appendix F)
Student contribution to work: 80%.
Co-author signatures and dates:
12/02/2018 12/02/2018 12/02/2018 12/02/2018 12/02/2018
Gyan Thomas McAuley Strobel A. Newton Sam Edmond M.
Kimberley Natalie Karen
15/02/2018 16/02/2018 18/02/2018 19/02/2018 14/07/2018
Amenga-Etego Owusu-Agyei Caitlin Kirkwood Tawiah-
Seeba Seth Shannon Betty Agyemang
Charlotte
xv
2. Gyan Thomas, McAuley Kimberley, Strobel A. Natalie, Newton Sam,
Owusu-Agyei Seth, Edmond M. Karen. The influence of socioeconomic factors on choice of infant male circumcision provider in rural Ghana; a community level population based study. BMC Pediatr 2017; 17(1): 185.
Location in thesis: chapter 5 (Appendix F )
Student contribution to work: 80%.
Co-author signatures and dates:
12/02/2018 12/02/2018 12/02/2018 12/02/2018 12/02/2018
Gyan Thomas McAuley Strobel A. Newton Sam Edmond M.
Kimberley Natalie Karen
Owusu-Agyei
Seth
16/02/18
3. Gyan Thomas, McAuley Kimberley, O'Leary Maureen, Strobel A. Natalie,
Edmond M. Karen. Healthcare seeking patterns of families of infants with circumcision-related morbidities from two population-based cohort studies in
Ghana. BMJ Open 2017; 7(8): e018185.
Location in thesis: chapter 6 (Appendix F)
Student contribution to work: 80%.
Co-author signatures and dates:
xvi
12/02/2018 12/02/2018 12/02/2018 12/02/2018 17/02/2018
Gyan Thomas McAuley Strobel A. Edmond M. O'Leary
Kimberley Natalie Karen Maureen
4. Gyan Thomas, Strobel A. Natalie, McAuley Kimberley, Shannon Caitlin,
Newton Sam, Tawiah-Agyemang Charlotte, Amenga-Etego Seeba, Owusu-
Agyei, Forbes David, Edmond M. Karen. Health service provider education and/or training in infant male circumcision to improve short-and long-term morbidity outcomes: protocol for systematic review. Syst Rev 2016; 5(1): 41.
Location in thesis: chapter 7 (Appendix F)
Student contribution to work: 80%.
Co-author signatures and dates:
12/02/2018 12/02/2018 12/02/2018 12/02/2018 12/02/2018
Gyan Thomas McAuley Strobel A. Newton Edmond M.
Kimberley Natalie Sam Karen
15/02/2018 15/02/208 16/02/108 18/02/2018 14/7/2018
Amenga- Forbes Owusu- Shannon Tawiah-
Etego David Agyei Caitlin Agyemang
Seeba Seth Charlotte
xvii
5. Gyan Thomas, Strobel A. Natalie, McAuley Kimberley, Shannon Caitlin,
Newton Sam, Tawiah-Agyemang Charlotte, Amenga-Etego Seeba, Owusu-
Agyei Seth, Forbes David, Edmond M. Karen. Health service provider education and/or training in infant male circumcision to improve short and long term morbidity outcomes: A systematic review. Submitted to PLoS One,
27th September 2017.
Location in thesis: chapter 8.
Student contribution to work: 80%.
Co-author signatures and dates:
12/02/2018 12/02/2018 12/02/2018 12/02/2018 12/02/2018
Gyan Thomas McAuley Strobel A. Newton Edmond M.
Kimberley Natalie Sam Karen
15/02/2018 15/02/2018 16/02/2018 18/02/2018 14/107/2018
Amenga-Etego Forbes David Owusu- Shannon Tawiah-
Seeba Agyei Caitlin Agyemang
Seth Charlotte
Student signature:
Date: 28th February 2018
xviii
I, Dr Kimberley McAuley certify that the student statements regarding their contribution to each of the works listed above are correct
Coordinating supervisor signature:
Date: 28th February 2018
xix
Presentations arising from this research
2015
Infant male circumcision morbidities and care seeking behaviour in sub-
Saharan Africa: a population-based cross sectional study in rural Ghana
(Presenter).
2015 Child & Adolescent Health Research Symposium Perth, Australia.
Princess Margaret Hospital Foundation, Princess Margaret Hospital,
Telethon Kids Institute and the University of Western Australia.
2017 Determinants of morbidity associated with infant male circumcision;
community level population-based study in rural Ghana (Presenter).
15th World Congress on Public Health Melbourne, Australia.
Healthcare seeking patterns of families of infants with circumcision
related morbidities from two population-based cohort studies in Ghana
(Presenter).
11th Annual Telethon Kids Institute Student Circle Symposium and
Development Program Rottnest Island, Western Australia.
Healthcare seeking patterns of families of infants with circumcision
related morbidities from two population-based cohort studies in Ghana
(Presenter).
10th European Congress on Tropical Medicine and International Health
Antwerp, Belgium.
xx
List of figures
Figure 3.1 Map of Ghana and the study area ...... 28
Figure 3.2a Macro level conceptual framework of determinants of circumcision related morbidity and care seeking…………………………………………………39
Figure 3.2b Micro level conceptual framework of determinants of circumcision morbidity……………………………………………………………………………...40
Figure 4.1 Flow diagram for the circumcision study……………………………...52
Figure 5.1 Flow diagram for the study: factors which influence choice of circumcision provider………………………………………………………………..75
Figure 6.1 Flow diagram for live births from August 2010 to November 2011 included in the Neovita study……………………………………………………..100
Figure 6.2 Flow diagram for live births from May 2012 to December 2012 included in the circumcision study study………………………………………...101
Figure 8.1 Study flow diagram, showing the results of the searches for the systematic review…………………………………………………………………...146
xxi
List of tables
Table 4.1 Definitions of circumcision morbidity ...... 49
Table 4.2 Infant and maternal characteristics in the study ...... 53
Table 4.3 Circumcision related morbidities and health care practices by provider type in the study population ...... 57
Table 4.4 Health service determinants of circumcision associated morbidity in the study population ...... 60
Table 4.5 Socio demographic determinants of circumcision associated morbidity in the study population ...... 61
Table 5.1 Infant and maternal characteristics in the study population ...... 77
Table 5.2 Determinants of choice of informal provider for infant male circumcision ...... 79
Table 5.3 Details of cash payments and in-kind payments by provider type ..... 82
Table 5.4 Cost of circumcision by household income status ...... 83
Table 5.5 Distance to health facility by household income status ...... 85
Table 6.1 Characteristics of families and infants who sought care for acute illness in the Neovita trial ...... 102
Table 6.2 Characteristics of families and infants who sought care for morbidity in the circumcision study ...... 104
Table 6.3 Predictors of care seeking behaviour in the Neovita trial population
...... 105
Table 6.4 Type of health care provider for circumcision-related morbidity ..... 108
xxii
Table 8.1 Description of included studies ...... 148
Table 8.2 ROBINS-I risk of bias assessment ...... 151
Table 8.3 GRADE findings ...... 155
xxiii
List of abbreviations
WHO World Health Organization
UNAIDs Joint United Nations Programme on HIV/AIDS
USAID United States Agency for International Development
HIV Human Immunodeficiency Virus
GHS Ghana Health Service
KHRC Kintampo Health Research Centre
UWA University of Western Australia
OR Odds ratios
CI Confidence intervals
AOR Adjusted odds ratio
PCA Principal Component Analysis
GIS Geographic Information System
KM Kilometre
Ghs Ghana Cedis
AIDS Acquired immunodeficiency syndrome
MD Mean difference
RCT Randomised controlled trial
PRIMSA-P Preferred Reporting Items for Systematic Reviews and Meta-
Analyses Protocols
GRADE Grades of Recommendation Assessment Development and
Evaluation
xxiv
ROBINS-I Risk Of Bias In Non-Randomized studies-of Interventions
RR Risk ratio
SMD Standardised mean difference
SE Standard error
CI Confidence interval
UTI Urinary tract infection
ICC Intracluster correlation coefficient
xxv
Chapter 1: Introduction
Male circumcision involves the complete or partial removal of the foreskin of the penis1. The procedure has historically been practiced for cultural, religious, social and medical reasons2,3 and is often performed during infancy or in adolescence as a rite of passage4. An estimated 30% of males were reported to be circumcised worldwide including; 665 million infants, children and adults in 20102. A total of
90.7% of infant males were circumcised in Ghana in 20125. In comparison, the prevalence of infant male circumcision in the United States was 58%; 25% in the
Philippines, 20% in Australia and 6% in the United Kingdom between 2006 and
20122,6,7. Randomised controlled trials conducted in sub-Saharan Africa have shown that medically performed circumcision is safe and can reduce men’s risk of acquiring HIV infection from heterosexual exposure by about 60%8-10. A systematic review of the evidence to data also indicated that female partners of circumcised men were at decreased risk of HIV and five other diseases including cervical cancer, cervical dysplasia, hypes simplex virus type 2, chlamydia and syphilis11.
Other benefits of male circumcision include penile hygiene, protective effects for cancer of the penis, reduced urinary tract infections and protective effects for sexually transmitted infections12. Although the procedure is safe, there are risks associated with male circumcision at all ages such as pain, bleeding, infection, and inadequate skin removal2,13-17.
1
There has been an expansion of adult and infant male circumcision services throughout Africa as part of a long-term Human Immunodeficiency Virus (HIV) prevention strategy18-20. However, little is known about the burden and determinants of morbidity associated with infant male circumcision in low and middle income countries including West Africa. Acute morbidities associated with infant male circumcision include pain, bleeding, swelling, infection, tetanus, skin bridges, meatitis and inadequate skin removal2,14-17,21. A systematic review published in 2010 reported a morbidity rate of 1.5% after infant male circumcision14. Nine of the studies in this review were from Africa but all were based in urban hospitals and none included follow-up after discharge14,22-29. It was also unclear whether the data on morbidities were obtained by parental report or if the infants were examined by trained fieldworkers.
There is also no reliable evidence about the factors that influence choice of circumcision providers in poor rural areas in Africa. In sub Saharan Africa, infant male circumcisions are performed by formal health service providers including nurses, doctors and medical assistants in health facilities or informal providers including: pharmacists, domestic staff, family members, and village based traditional (traditional and informal are used interchangeably) service providers in the home21,30. Initiatives to improve the health care practices of both informal and formal circumcision providers are underway in Africa21,30. These include training on
2 infection control, instruments to perform circumcision and hygiene. However, other strategies to influence family’s care seeking patterns, improve use of health facilities, and improve use of trained circumcision providers are also needed. This requires an understanding of the key factors which influence a family’s choice of circumcision provider. Another systematic review published in 2010 reported that socio- economic factors such as income, location (rural and urban), and cost of the circumcision were key determinants of choice of health service provider for infant male circumcision2. Socio-economic status, cost, and geographical access are also key determinants of care seeking for antenatal and birthing care in sub-Saharan African populations31-36. However, to my knowledge, there have been no studies from poor rural Africa that have investigated the effect of these factors on choice of infant male circumcision provider.
In addition, studies conducted in rural Ghana have reported that poorer families believed infant illnesses related to traditional practices are not meant to be managed in the hospital or clinic37-39. Although, no other data on care seeking for infant male circumcision related morbidity in rural African areas are available, other studies have examined the determinants of care seeking for acute infectious diseases and newborn complications in rural Africa40,41. A recent Ghanaian study identified poor health care seeking behaviour in families of infants with acute illness during the postpartum period. In this study 29% and 54% of sick infants
3 received health care within two and fourteen days, respectively42. Another
Ghanaian study identified poverty and personal preferences as key determinants of poor care seeking behaviour by families of infants with illness43. A study from
Burkina Faso also linked traditional concepts of illness in Africa to delays in seeking professional medical care by families of infants with malaria44. Distance to health facility, health provider attitude and inadequate supplies have also been reported as key determinants of delayed care seeking for malaria cases in infants in sub-Saharan African populations45. However, it appears that there have been no studies which have compared health seeking behaviour between families of infants with circumcision related morbidities and families of infants with acute illnesses that are not circumcision related.
Furthermore, information about the effect of health service provider training in infant male circumcision on circumcision related morbidity appears largely anecdotal in community based settings. As a surgical procedure, it is expected that male circumcision is performed at the required standard by all circumcision service providers who have completed requisite training. For example, health service providers need to be trained on: how to perform circumcision with aseptic techniques; anaesthesia; use of equipment and tools; measures to reduce complications; follow-up care; informed consent and education for families. These
4 skills appear to be simple and easy to implement by a range of health care professionals. Infant male circumcision is commonly practiced in many parts of the world by formal health service providers such as doctors/surgeons, medical assistants, midwives/nurses. Yet, published studies have indicated that formal health service providers often do not receive the required training and education to perform these procedures to optimal standards resulting in complications ranging from 1.5 to 20.2% after circumcision14,27. Other studies have also revealed that even doctors providing infant male circumcision receive minimal training which is generally informal and unstructured2,46-48.
The World Health Organization (WHO) has developed a simple training program and manual to improve the skills of health service providers in performing infant male circumcision in a variety of settings1. However, this training package does not appear to have been formally evaluated. What is also not yet clear is the effectiveness of training to health professionals who have had no prior training in surgery, including informal providers such as traditional circumcision providers, traditional birth attendants, religious leaders, other health facility staff and the duration and intensity of the training required. It is also unclear whether the provision of education and/or training using simple packages such as the WHO manual would improve non-surgical health service provider skills in infant male circumcision and would lead to reduced rates of short and long-term morbidity
5 outcomes. Currently no systematic review has been published on this topic. A group of authors systematically assessed RCTs of interventions to improve the safety and efficacy of nontherapeutic male circumcision in persons of any age and eight RCTs were located in 201049. Only two of these studies were conducted among infants, and involved an assessment of circumcision on pain response to vaccination and the relationship between circumcision and pain distress; and mother and infant bonding. None of the studies included training of service providers49. Another group of authors searched for RCTs of routine neonatal circumcision for the prevention of urinary tract infections in infancy and found no studies in 201250.
This thesis therefore aimed to provide an in depth assessment of the burden and determinants of morbidity associated with infant male circumcision and care seeking in community based settings, the factors that influence choice of circumcision providers, and training requirements. This research will be used to inform the development of policies and facilitate the training and monitoring of male circumcision services in countries where circumcision is widely practiced. The study will also be beneficial to the WHO/UNAIDS circumcision expansion program in parts of Africa where the prevalence of both adult and infant male circumcision is low. This will be achieved through the provision of population based data on the burden and determinants of morbidity associated with infant male circumcision,
6 care seeking practices of families and training requirements in low and middle income countries.
1.2 Aims of this research
The overall aim of this doctoral study was to identify the burden and key determinants of morbidity associated with infant male circumcision. The contribution of care seeking by families for infant circumcision and circumcision related morbidities in a community based setting in sub-Saharan Africa was also explored. There were five specific objectives:
1. To assess the burden and key determinants of morbidity associated with infant
male circumcision in a community based setting in rural Ghana. The research
conducted to address this objective is described in thesis chapter four,
publication one.
2. To determine the key factors which influence the choice of infant male
circumcision provider in rural Ghana. The research conducted to address this
objective is described in thesis chapter five, publication two.
3. To investigate and compare the health care seeking behaviour of families with
infants who had circumcision-related morbidity and families with infants who
experienced acute illness in the study area. The research conducted to address
this objective is described in thesis chapter six, publication three.
7
4. To investigate and identify gaps in the literature about the effectiveness of health
service provider education and/or training in infant male circumcision on
morbidity or mortality outcomes in any setting. The systematic review
conducted to address this objective is described in thesis chapter seven,
publication four and chapter eight, submitted for publication thesis manuscript.
5. To make recommendations and describe the implications for further research,
policy and practice to reduce morbidity associated with infant male circumcision
and improve care seeking in sub Saharan Africa. These have been discussed in
chapter nine as part of the thesis discussion and conclusions from the research
findings.
8
Chapter 2: Literature Review
2.1 History and key determinants of male circumcision
Circumcision comes from the latin word “circumcidere” which means to cut around51. There are varied reports about the origin of circumcision, although it is well known that the practice has been ongoing for several centuries52. Some reports indicate that circumcision originated from Egypt about fifteen thousand years ago before the practice spread globally53. Egyptian pharaonic paintings depicting circumcision have provided evidence that the practice has been performed for at least six thousand years52. Several studies have also described that male circumcision has been performed in North Africa for many centuries54.
Male circumcision is practiced in many countries across the world including Africa,
Middle East, Asia, Australia and United States of America (USA) as a cultural practice or for health benefits2,21,52,55. In addition, religions such as Islam and Jewish faiths have traced the practice of male circumcision to Ibrahim in the Quran and
Abraham in the Bible, respectively52. Practicing Muslims are required by the teachings of the Quran to be circumcised preferably on the seventh day after birth.
Muslims are reported to be the largest religious group to perform circumcision2,14.
Circumcision is also required among practicing Jews56. Other religions including
Christianity, Hinduism and Buddhism are described as not having any religious requirements related to circumcision of males2,52. Other reasons for families initiating
9 infant male circumcision have been hygiene, family tradition, medical benefits, improved sexual performance and perceived better looks of the penis after circumcision57. Today, male circumcision is commonly performed for cultural, religious, social and medical reasons2.
The practice of male circumcision in Ghana started from the 20th century21. Some authors have described male circumcision in Ghana to have been introduced by
Hausa Muslims who arrived from the northern neighbouring countries of Ghana21.
Despite Islam not being the dominant religion in Ghana the practice of circumcision has evolved and many different reasons including religious, socio-cultural and medical have been reported to account for the practice of male circumcision in
Ghana2,58.
2.2 Prevalence of male circumcision
An estimated 665 million infants, adolescents and adults were circumcised worldwide in 20102. Male circumcision is widespread in West Africa, North Africa, some Central and Eastern African countries, and less common in Southern
Africa2,30,52,58,59. Sixty eight percent of Muslim men in North Africa, Middle East and parts of Asia were circumcised in 20102 . In addition, an estimated 80% of men in the
United States were also circumcised in 20126. Circumcision is performed in adolescence as a rite of passage in countries such as South Africa, Kenya and
10
Tanzania, whereas in West Africa the procedure is conducted at a much younger age with the majority done in infancy2,30,60. Approximately, 90% of infant males were circumcised in West Africa in 20102 and over 90% in Ghana in 20125. In 2010, it was reported that infant male circumcision was rare in most parts of Asia (5-50%), South
America (19%), Central America (20%), and most of Europe (19%)2 and has low prevalence rates in Australia (10-20%)6,52,61,62. In Australia, Great Britain and the USA, infant male circumcision prevalence has fallen in recent years due to concerns about pain, complications and reduction of insurance rebates61-64.
A systematic review and one other study in Ghana have reported that male infants are circumcised most commonly on the eighth day after birth2,21. In Nigeria, infant males are circumcised during a ceremony on the seventh day after birth2. The
American Academic of Pediatrics issued a technical report on infant male circumcision stating that “Evaluation of current evidence indicates that the health benefits of newborn male circumcision outweigh the risks; furthermore, the benefits of newborn male circumcision justify access to this procedure for families who choose it”6.
2.3 Types of circumcision provider
Infant male circumcision is a simple procedure and is currently performed by both formal and informal circumcision providers, the latter particularly in low and
11 middle income countries2,16. Circumcision in Ghana and other parts of West Africa is performed by informal providers (including village based traditional providers) and formal health care providers (including doctors, nurses and medical assistants)2,5,16,21,65. A recent study from Ghana reported that 58% of infant male circumcisions were performed by informal providers and 42% by formal providers5.
In contrast, a higher proportion (67.1%) of formal healthcare providers performed infant male circumcision in Nigeria65. A systematic review involving studies from all regions including Africa and Asia also reported a higher proportion of infant male circumcisions being performed by formal healthcare providers (91%) and a lower proportion (9%) by informal providers2. However, all the African studies in the review were based in urban hospitals and cannot be representative of the whole of Africa. Religion, cost, perception of quality, household income and location (rural and urban) were noted as the main factors that influenced choice of circumcision provider in the review2.
Informal providers conduct infant male circumcision particularly in rural areas in many Muslim countries in the Middle East, Egypt, parts of Asia including Pakistan and in Ghana2,21,66. For example in Ghana they are called Wanzams21, in Israel they are called Mohel (religious circumcision provider)2 and in Pakistan they are called circumcision barbers or drummers67. In Ghana, Pakistan and other low and middle income countries, informal providers are reported to have no formal training but
12 learn the trade through apprenticeship from another experienced informal provider2,21. However in Israel, the Mohel, who performs infant male circumcision, is reported to receive training and supervision from the Israeli Health Ministry2.
2.4 Circumcision methods
The Plastibell device, Mogen clamp and the Gomco clamps are devices widely used for infant male circumcision in hospital and clinic settings under local anaesthesia1,2,6. The clamp methods are described by some authors to cause less pain, prevent bleeding and protect the glans of the penis1,6,68. The Plastibell device is commonly used in many settings. Complications associated with the improper use of the Plastibell device have been reported including infection and penile necrosis due to retained Plastibell device2,69. The Mogen clamp is largely used in developed countries including USA and Canada2,6. The use of the Mogen clamp is reported to result in less complications compared to other methods2,69-71. The Mogen clamp has also been reported to be reusable, faster and reduces pain compared to the Gomco clamp1,2,6,68,70,71. Overall, there is little data on the effectiveness, cost and use of these devices in low and middle income countries.
Freehand instruments such as scalpels, scissors, knives and blades are commonly used for circumcision by informal providers in sub-Saharan Africa, North Africa, parts of Middle East and Asia2,5,21. The use of locally made circumcision instruments
13 under unhygienic conditions have also been reported in Pakistan and other Middle
Eastern countries2. Infant male circumcision performed by informal providers is usually done without the use of local or general anaesthesia2,21. After circumcision, wound care and activities to prevent bleeding including the application of local medicinal treatments and dressings have been reported in some low and middle income countries2,21. For example, in Ghana and other West African countries, traditional circumcision providers apply herbal preparations to the circumcision wound to prevent bleeding21. In Israel, the Mohel has been described to wrap a bandage on the circumcision wound to prevent bleeding2. However, there are no data from poor rural areas in Africa.
2.5 Importance and burden of morbidity associated with infant male circumcision
The scale-up of voluntary medical male circumcision as a method of combating
HIV/AIDS was recommended by the global community following three randomised controlled trials that showed a 60% protective effect of circumcision during adulthood on HIV acquisition in sub Saharan Africa8-10. Recommendations from the research were that, male circumcision was effective in reducing heterosexual transmission of HIV by 60% and circumcision services should be integrated with other HIV/AIDS prevention strategies8-10. The WHO, USAID and UNAIDS are now expanding adult and infant circumcision services in regions of high HIV prevalence
14 as a longer-term HIV prevention strategy2,20. Other benefits of male circumcision include penile hygiene, protective effects for cancer of the penis, reduced urinary tract infections, protective effects for sexually transmitted infections and reduction of genital herpes12.
There are a number of acute morbidities associated with infant male circumcision2,13-
17. However, few studies use standardised case definitions. A systematic review involving quantitative methods conducted in 2010 in males aged under twelve years reported a median morbidity risk of 1.5% (range 0-16%) after circumcision14. The large studies in the review came from the USA, UK and Israel where the morbidity risk were reported to be around 0.2% in the USA, 0.9% in UK and 0.1% in Israel14.
Nine studies were from Africa, with the number of participants within each study ranging from 95 to 1,56314,22-29,72. The average prevalence of complications ranged from 0% to 13% in these African studies14. Serious adverse events such as penile amputation, urethral laceration (0-2%) were also recorded. However, none of the
African studies included infants who were circumcised in the community-based rural settings. All studies were conducted in urban settings and no studies included follow-up after hospital discharge14,65. It was also unclear whether the data on morbidities were obtained by parental report or if the boys were examined by trained field workers. Parental report has been shown to over- and under-estimate complication rates73.
15
Another systematic review assessed eight randomised controlled trials (RCTs) of interventions aiming to improve the safety and efficacy of nontherapeutic male circumcision49. However, only two of the studies were conducted among infants, and these were hospital based trials of 87 infants in Canada and 60 in USA. The review reported that infants have similar responses to pain to adults and that circumcision performed without anaesthesia is likely to affect infant behaviour, mother-infant bonding and increase response to pain49.
Limited age specific prevalence data are available for morbidities associated with infant male circumcision. A systematic review reported lower frequencies of morbidity among infants compared to older children14, and a study from the US reported lower incidence of morbidities in newborns than in boys aged 3-8.5 months74. Overall, there is a paucity of high quality data from community-based studies examining morbidities associated with infant male circumcision in sub-
Saharan Africa. Data collected by trained field workers using standardised definitions for adverse events are especially lacking.
2.6 Determinants of morbidity associated with infant male circumcision
Site of circumcision (hospital, clinic, home) and access to equipment’s have been reported to be important determinants of morbidity associated with circumcision2,14,48. This was highlighted in a Kenyan study among boys in which
16
34.3% of morbidities (including excessive incisions, swelling and inadequate skin removal) were attributed to circumcision performed at home compared to 11.1% at hospitals and clinics48. In addition, two systematic reviews2,14 and the Kenyan study48 described increased risk of morbidity after circumcision by the use of inadequate and inappropriate equipment’s. This was clearly indicated by the Kenyan study in which incorrect length of suture material made stitches rupture with blunt scissors causing rough wound edges48. There have been no other community-based studies examining the determinants of morbidity associated with infant male circumcision in low income countries.
Type of service provider, informal versus formal, have also been shown to be a determinant for infant male circumcision related morbidities. A mixed methods systematic review of neonatal and child male circumcision reported a higher frequency of morbidities following circumcision by traditional providers (29%) compared to formal health service providers (6%)2. Another study from Kenya also reported a post circumcision complication of 34.3% by traditional informal providers compared to formal health service providers 11.1%48. The systematic review also included a study from Turkey involving 407 boys circumcised by informal providers during a traditional ceremony. The frequency of morbidity within this population was 73%2. Similar high morbidities (60% overall) were also seen in another study from the Philippines involving 114 boys. In the Pilipino study,
17
68% of the circumcisions were performed by informal providers2. But there are no data from poor rural areas in Africa.
Data from several observational studies have indicated that training and skills of the health service providers were key determinants of morbidity associated with infant male circumcision2,14,25,27,48,72. This was clearly reported by a systematic review14 conducted in 2010 and indicated an average of 1.6% post circumcision related morbidity among doctors in a teaching hospital compared to 20% at private clinics where the level of training was described to be low14. Other studies have also described that health care professionals who perform infant male circumcision often do not receive enough training to perform the procedure to required standards resulting in complications14,27,48. Morbidity risk has been shown to range from 1.5 to 20.2% after circumcision because of inadequate training of circumcision providers14,27. Authors have also described how some doctors who provide infant male circumcision receive minimal training which is generally informal and unstructured2,46-48.
Systematic reviews14,75 and other hospital based studies from Kenya71, Pakistan76 and the USA74,77 have indicated that circumcision performed in infants less than three weeks after birth was associated lower risk of morbidity compared to circumcision in older infants. A systematic review in 2010 found 1.5% (range 0-16%)
18 morbidity risk following circumcision in infants aged under 11 months compared to 14% in older boys aged 1 to 12 years14. The lower frequency of morbidity reported in younger babies was attributed to the simpler nature of the procedure during early infancy and faster healing.
Hygiene related care practices have been described as being an important determinant for morbidities in infant male circumcision. For example, in a 2010 systematic review it was reported that there was a 80% risk of morbidity after circumcision by traditional informal providers in unhygienic conditions in low income countries14. Although there is limited empirical evidence, studies have highlighted the ongoing important hygiene issues such as lack of access to clean water, no hand washing, uncleaned instruments and general unsterile conditions14,21,48,78.
It is important to note that the data quality of assessing determinants of morbidity varies with assessments made by trained formal health care providers14,73, trained fieldworkers73 and parents14,73. Parental report of circumcision related morbidities have been described to under estimate the true prevalence of morbidities after circumcision14. For example, a Kenyan study conducted among boys aged 5 to 21 years indicated a higher morbidity risk (88%) among circumcisions that were observed by physician and a research assistant (12 circumcisions by formal health
19 care providers and 12 by informal providers) compared to self-reported morbidities (25%) in 1007 boys after 62 days post circumcision73.
2.7 Factors that influence choice of health service provider for infant male circumcision
A mixed methods systematic review assessed 293 studies for the key determinants for choice of circumcision provider2. This review included 40 studies from Africa however only two were from rural settings2. The review indicated that families in northern Nigeria used traditional providers for infant male circumcision because they are often highly valued by families for cultural and religious reasons2,79. A study conducted in Northern, Eastern and Central regions of Tanzania also reported that
71% of participants preferred traditional provider for circumcision services because the participants believed circumcision was a “ritual” practice80. The systematic review also reported an increase in the use of traditional providers by families with lower level of education and poorer households in Western Iran2. In the review, richer families were reported to be four fold more likely to have their infant male circumcisions performed in hospital settings compared to poorer families. The review also indicated that 72% of families with low education will choose traditional providers compared to 18% of families with higher level of education2. The review further reported a threefold reduction in the use of traditional providers for circumcision services among families with higher education compared to families
20 with lower educational level in cities in Turkey2. In addition to education, increased distance from a medical facility has been shown to contribute to the general use of traditional providers for infant male circumcision, especially within Africa and other low and middle income countries, and among poorer families2,79. However, there remains a lack of data specifically focusing on the determinants of health service provider selection for infant male circumcision in rural areas in low and middle income countries, especially in Africa.
Initiatives to improve the health care practices of informal circumcision providers
(Wanzams) and other circumcision providers are underway in Ghana21,30. These include training on infection control, instruments to perform circumcision and hygiene. However, other strategies to influence family’s care seeking patterns, improve use of health facilities, and improve use of trained circumcision providers are also needed. This requires an understanding of the key factors which influence a family’s choice of circumcision provider. A systematic review reported that socio- economic factors such as income, location (rural and urban), and cost of the circumcision were key determinants of choice of health service provider for infant male circumcision2. Socio-economic status, cost, and geographical access are also key determinants of care seeking for antenatal and birthing care in sub-Saharan African populations31-36. However, to my knowledge, there have been no studies from rural
21
Africa that have investigated the effect of these factors on choice of infant male circumcision provider.
2.8 Determinants of care seeking for morbidity
Each year, a concerning high risk of circumcision related morbidities are reported in low and middle income countries including Africa after infant male circumcision. A systematic review of quantitative methods reported, on average, a
1.5% (range 0-16%) frequency of morbidity after infant male circumcision in 201014.
Another study from Ghana in 2017 described an 8.1% risk of morbidity after infant male circumcision5. In the systematic review, frequency of morbidities after circumcision by an informal provider was 29% compared to 6% by a formal health care provider14. In the Ghanaian study, 53% of morbidities were due to circumcision by an informal and 47% by formal health care provider5.
Appropriate health care seeking for infants with circumcision-related morbidities could minimise the occurrence of severe short and long-term morbidities. Studies conducted in rural Ghana have reported that poor women believe infant illnesses related to traditional practices are not meant to be managed in the hospital or clinic37,38. However, no other additional data on care seeking for circumcision-related morbidities in rural Africa are available. There are more studies that have examined the determinants of care seeking for acute infectious diseases and newborn
22 complications in rural Africa40,41. A recent study involving mothers and infants in
Ghana identified poor health care seeking behaviour in families of infants with acute illness during the postpartum period. The percentage of sick infants receiving appropriate care within 2 and 14 days postpartum was reported to be 29% and 54% respectively42. Another study investigating infants aged 1 to 28 days after birth with serious illness in Ghana revealed poor care seeking behaviour among families with infants experiencing acute illness. Only 39% of sick infants received care in hospital in this study43.
A Ghanaian study also identified poverty and personal preferences as key determinants of poor care seeking behaviour by families of infants with illness43. A study from Burkina Faso also linked traditional concepts of illness in Africa to delays in seeking professional medical care by families of infants with malaria44. Distance to health facility, health provider attitude and inadequate supplies have also been reported as key determinants of delayed care seeking for malaria cases in infants in sub-Saharan African populations45. However, it appears that there have been no studies which have compared health seeking behaviour between families of infants with circumcision related morbidities and families of infants with acute illnesses that are not circumcision related in poor rural areas in Africa.
23
2.9 Effect of health service provider education and/or training in infant male circumcision
Education and training to improve patient health outcomes has been shown to be effective for surgeons and doctors1,81-84. Yet, there is limited information about the effectiveness of this intervention on infant male circumcision outcomes. As a surgical procedure, it is expected that male circumcision is performed at the required standard by all circumcision service providers and that they have completed the requisite training. For example, they would need to be able to complete circumcision with aseptic techniques; anaesthesia; use of equipment and tools; processes to reduce complications; follow-up care; informed consent and education for families. Although these skills appear to be simple and easy to implement, published studies have shown that health care providers are not properly trained to perform circumcision to acceptable standards14,46-48. As a result, it has been reported that complications range from 1.5% to 20.2% after circumcision14,27.
The WHO has developed a simple training program and manual to improve the knowledge and competency of health care providers in performing infant male circumcision in a variety of settings1. However, this training package does not appear to have been formally evaluated. What is also not yet clear is the effectiveness of training to health professionals who have had no prior training in surgery,
24 including informal providers and the duration and intensity of the training required.
It is also unclear whether the provision of education and/or training using simple packages such as the WHO manual would improve non-surgical health service provider skills in infant male circumcision, leading to reduced rates of both short- and long-term morbidity outcomes.
To date, there has been no systematic review investigating the effectiveness of health service provider training in infant male circumcision to improve morbidity outcomes. Only two other systematic reviews on infant male circumcision have examined any element of training on morbidity outcomes49,50. The first review conducted in 2010 found eight RCTs of interventions aimed at improving the safety and efficacy of nontherapeutic male circumcision at any age. Only two of the studies were conducted among infants and assessed the relationship between circumcision and pain distress, and mother and infant bonding49. The second review conducted in 2012 found no studies available to determine the effect of infant male circumcision in preventing urinary tract infections50 . None of the studies involved training of circumcision service providers.
The WHO office of Reproductive Health and Research sponsored stakeholder meetings in Ghana and Nigeria in 2008 to discuss infant male circumcision in sub
Saharan Africa for HIV prevention program and policy2,30. A recommendation from
25 these meetings included additional research in infant male circumcision in Africa to guide the scale up of circumcision services. This study will therefore explore the gaps in the literature on the effectiveness of health service provider education and training in infant male circumcision on morbidity and mortality outcomes.
2.10 Policies and guidelines on infant male circumcision
There is no regulation in Ghana to guide the practice of infant male circumcision2,21,30.
The lack of regulation for infant male circumcision practice also exists in other parts of sub Saharan Africa and several low and middle income countries2,80. To address this gap, the WHO developed a training manual to guide the practice of infant male circumcision1. In the Middle East, there are regulations that guide the practice of male circumcision at all ages2. For example in Israel, it has been reported that the performance of male circumcision up to six months of age is regarded as a religious practice and could be performed by both medical and registered religious providers.
However, male circumcision after six months was regarded as a surgical procedure and required a qualified doctor to perform the procedure2. A committee made up of both medical professionals and religious leaders also exists to supervise the provision of circumcision services in Israel2. In Saudi Arabia, it has been reported that untrained and unlicensed health professionals are not permitted to perform circumcision2.
26
In South Africa, there are existing laws that requires traditional circumcision providers to be registered with the country’s Health Department60. Both the South
African constitution and the children’s rights policy have made circumcision of male children under 16 years of age illegal except for religious and medical reasons60. The
WHO consultative workshop organised in Ghana in 2008 and other studies have recommended the enactment of legislation and regulations to guide the practice of circumcision providers in Ghana and the rest of sub Saharan Africa2,21,30,80.
27
Chapter 3: Methods
3.1 Study setting
The study was conducted at the Kintampo Health Research Centre (KHRC) of the
Ghana Health Service (GHS). The KHRC study area (Figure 3.1) covers nine
Figure 3. 1 Map of Ghana
28 contiguous districts in the Brong Ahafo Region of central Ghana: Kintampo North,
Kintampo South, Wenchi, Tain, Banda Ahenkro, Techiman North, Techiman South,
Nkoranza North, and Nkoranza South covering 12,000 square kilometres85. These districts also formed the study area for the “Neovita trial”. The Neovita trial was an individually randomised controlled trial (RCT) of vitamin A administered to infants within two days of birth. In the RCT, a total of 22,955 infants were followed up monthly for twelve months post birth where data on infant’s vital status and health were recorded86. The Neovita study recruited infants from 16th August 2010 to 6th
November 2011 and the last 12 month follow up visit occurred on 31st December
2012. In addition the funders (the World Health Organization and the Bill and
Melinda Gates Foundation) requested that recruitment of infants at birth be continued for an additional 12 months to increase the sample size of the mortality and morbidity data set. In this thesis all infants recruited from 16th August 2010 to
31st December 2013 are referred to as the Neovita trial.
The study area lies within the forest-savannah transitional ecological zone, and has two distinct rainy seasons from April to July and from September to October. The area is densely populated (175 people/square mile)87 with a total population of approximately 750,000 persons88. The annual population growth rate is currently
3.1%. The prevalence of HIV in Ghana was low (2.1% in 2012)89. The area also
29
recorded a low HIV prevalence in 2012 (2.0% in women attending antenatal care)89.
An estimated 95% of men in the study area were circumcised in 200890.
Over the period from 2010 to 2013, 20% of the population in the study area lived in
the urban district administrative centres of Kintampo, Wenchi, Techiman and
Nkoranza40. The remaining 80% of the population lived outside these four towns in
rural settlements. The rural population lived in compounds, containing houses with
mud walls, and thatch or aluminium roofs, in dispersed villages surrounded by
farming land. Subsistence farming was the main occupation of the population and
the main crops were yam, maize and millet. The population was multi-ethnic and
education levels were generally low85,86.
Ghana is predominantly Christian (77.6%)90. Muslims also account for 15.0% of the
population90. Other religious affiliations included Traditional African religion (4.2%)
and 3.2% of the population had no religion90. The distribution of religious affiliation
in the study area was not different from the national breakdown. Majority of the
population in the study area were Christians (70%)86,91. Muslims accounted for 24%,
Other religions including traditional African and none constituted (6%)86,91.
There were four district hospitals and a dozen small health facilities that provided clinical (outpatient and inpatient) and maternity services and acted as the first referral
30 point for sub-districts and community based health care facilities. There were 30 doctors and 44 other formal health service providers (medical assistants and nurses) providing curative services in the four district hospitals at the time of the circumcision study. There were also several informal care providers such as drug sellers, traditional healers and religious leaders in the study area. There were approximately
60 village based traditional informal circumcision providers (called ‘Wanzams’ in
Ghana) and 100 formal (doctors, nurses and medical assistants) circumcision providers.
3.1.1 Data collection
All births in the Neovita study area were reported to the trial team via a network of fieldworkers and key informants. Fieldworkers visited families at home between two hours and two days after birth and collected data about the birth, socio- demographic characteristics and socioeconomic information (using an asset index).
All families of enrolled babies were visited every month by study fieldworkers until the infant reached 12 months of age to ascertain the vital status of the baby. Any infant death was followed-up by fieldworkers who conducted verbal autopsy interviews to ascertain the circumstances leading to the infant’s death.
All male liveborn Neovita infants were eligible to be included in the circumcision study. During a home visit when the infant was 8-11 weeks of age, the families were
31 asked by trained senior fieldworkers for detailed information on: age at circumcision; place of circumcision; type of circumcision provider; methods and instruments used; hygiene practices before, during and after the circumcision; amount of foreskin removed; medicines prescribed. The fieldworkers asked about complications or morbidities and whether these problems had been diagnosed or confirmed by a trained health professional. Standardised paper based data collection tools (including a standardised list of closed ended morbidity questions) were used for all interviews. A copy of the circumcision data collection form is provided in Appendix A. Visual aids (photographs and drawings) were used to assist with understanding all complications and morbidities. The fieldworkers also examined the infant’s penile area and recorded their findings. Families were asked if they sought care for circumcision related morbidities and if they had been given any advice about potential problems or care seeking, site of health care sought, type of health care provider, admissions and surgery. Consent to access the baby’s Neovita data including socioeconomic and demographic data was obtained. My contribution to data collections were the design of the data collection form and the development of visual aids.
3.1.2 Supervision and training
All the fieldworkers were trained for two weeks in all study procedures prior to the commencement of the study including recording of data on circumcision related
32 morbidities and care seeking appropriately. Interrater reliability was checked between all fieldworkers. During the study, field workers had scheduled and unscheduled supervisory visits from the study coordinator (TG) to assess data quality and consistency. In addition, TG was trained by medically skilled staff to identify morbidities that required immediate referral. TG was also accompanied by a medically trained member of staff for the first five physical examinations to give guidance and advice. Charts were provided which contained text descriptions and illustrations/photos of potential morbidities. If a rare or unusual morbidity
(morbidities other than bleeding or infection) was observed the medical staff were phoned to discuss the diagnosis and agree on the appropriate course of action.
My contribution to supervision and training included; i) training of all the fieldworkers and, ii) supervised each fieldworker during two visits in a month as part of the study scheduled and unscheduled supervisory visits. During these visits,
I observed the fieldworker interviewing mothers, and examined infants and recorded findings independently. Recorded data were compared between the fieldworker and TG/field supervisor and also used to provide feedback after the home visits and at weekly fieldworker meetings.
33
3.1.3 Data management.
Every Friday, field staff came to four central offices and attended weekly meetings.
At this time all data collection forms were reviewed by field supervisors and the forms were then transferred to the data centre at KHRC for double data entry. Data were entered by dedicated data entry clerks. A database separate to the main Neovita database was used to enter data. Standard checks were used at the data entry and verification stages. Central cleaning of the data, including additional running of range and consistency checks was also performed. My contributions to data management included; i) design of the data table, ii) design of range and consistency checks, iii) reviewed all the generated queries from the range and consistency checks and either resolved them directly or send the queries back to the field to be resolved by field supervisors, and iv) supervised the changes in the database and recorded the changes.
3.1.4 Analytical program of work for the thesis
There were two distinct parts to this thesis: a population based cohort study in
rural Ghana and a systematic review of the published literature of the effectiveness
of health service provider education and training in infant male circumcision on
morbidity and mortality outcomes. The first part of this thesis involved three
separate studies which analysed data from 3141 male infants born in the Brong
Ahafo region of central rural Ghana from 21st May 2012 to 31st December 2012. For
all analyses, I used multivariable logistic regression models to assess: (i) the
34 burden and key determinants of morbidity associated with infant male circumcision; (ii) the effect of key determinants on choice of circumcision provider; and (iii) the health care seeking behaviour of families with infants who had circumcision-related morbidity. In the third study similar methods were used to analyse data from a larger study of all infants born from 16th August 2010 to 7th
November 2011, who experienced acute illness in the same geographic area.
The explanatory variables chosen a priori were based on findings from systematic reviews of complications of infant male circumcision and care seeking behaviour.
Furthermore, the explanatory models constructed were not to identify which confounders to include for a particular risk factor nor to identify combinations of exposure variables that work. The model was to understand the causes of circumcision related morbidity (outcome) and to examine the exposures that influence this outcome. The model assigned meaning to the explanatory variables chosen apriori for inclusion. Formal stepwise methods are not recommended because of the problems with stepwise variable selection including i) the derived model given an over-optimistic impression, ii) regression coefficients are too large
(too far away from their null values), iii) changes in the data might lead to important changes in the variables selected for the final model and iv) stepwise procedures are not to be used as a substitute for thinking about the problem92.
35
During the initial development of the models collinearity was found with some variables in these models and in consultation with my supervisory panel I removed some of the exposure variables from the final models that were constructed. These included religion, maternal occupation, place of birth, birthing assistant, having a living child and having a child died.
The second part of the thesis involved a systematic review of studies published in standard databases and clinical trial registries in all languages from January 1985 to June 2016. Inclusion criteria were limited to any primary study that aimed to improve health service provider skills in performing infant male circumcision. I have described in detail the analytical approach for each study in the published papers (Appendix F), chapter four to chapter seven and chapter eight, submitted for publication thesis manuscript.
3.1.5 Ethics
The research conducted for this thesis obtained ethical approvals from Ghana
Health Service (GHS) Ethical Review Committee, the Institutional Ethics
Committee of Kintampo Health Research Centre (KHRC), the Research Ethics
Committee of London School of Hygiene and Tropical Medicine, and the Human
Research Ethics Committee of University of Western Australia (UWA).
36
Written informed consent was obtained from the legal guardian of male infants in the Neovita study to ask questions about their infant’s circumcision and associated morbidities. They were assured of their right to refuse consent without affecting their participation in the Neovita study and any other health services.
3.1.6 Funding
The circumcision study was nested into the Neovita trial, the parent study. The
Neovita trial was funded by the World Health Organisation (WHO) and Bill and
Melinda Gates foundation. The Neovita study funded all the necessary fieldwork and data management structures implemented for the field study. The doctoral component of the study was funded by an Australian Government Research
Training Program (RTP) Scholarship, Kintampo Health Research Centre and the
Ghana Health Service.
3.2 Conceptual framework
The macro level conceptual framework for this thesis is depicted in Figure 3.2a.
The macro level framework shows how the research findings related to each other; determinants of circumcision related morbidity, determinants of choice of circumcision provider, the effect of health service provider training and education and care seeking behaviour of families of infants who experienced circumcision related morbidity. The micro level conceptual framework for the research is
37 depicted in Figure 3.2b. This includes the determinants of morbidity associated with infant male circumcision; the determinants of type of circumcision provider and the determinants of care seeking by families of infants who have morbidity after circumcision.
38
Socioeconomic status Distance to health facility Maternal education Religion
Choice of circumcision Health care related care Training and/or provider practices education
Circumcision related morbidity
Care seeking from health care provider
Legend Did not assess Severe morbidities Insufficient evidence including death No effect Likely effect
Figure 3.2a Macro level conceptual framework of determinants of circumcision related morbidity and care seeking
39
Figure 3.2b Micro level conceptual framework of determinants of circumcision related morbidity 40
3.3 Thesis structure
This thesis is organised as a thesis by publication. Each of the papers is presented as a chapter on its own. I drafted all the papers with inputs from my supervisors and co-authors. Four papers have been published by peer reviewed journals and the final paper has been submitted to a peer reviewed journal.
As presented above, chapter one outlines the introduction to the research and aims and objectives. Chapter two presents the review of literature on male circumcision, the burden and key determinants of morbidity associated with infant male circumcision and care seeking. The effect of health service provider education and training on infant male circumcision, male circumcision policies and guidelines are also presented in chapter two. Chapter three provided details of the study methods and the conceptual framework showing how the research findings related to each other.
Following this is the results section which comprises five chapters. Chapter four to eight described the four papers which have been published and the final manuscript that has been submitted to a peer reviewed journal. The following results are presented; the burden and determinants of circumcision-related morbidity, factors that influence choice of circumcision provider and factors that influence families care seeking behaviour. The results of a systematic review of published literature that investigated the effectiveness of health service provider
41 education and training in infant male circumcision to improve short and long- term morbidity is also presented. The final chapter (chapter nine) presents the overall conclusions from each of the chapters, and the research limitations, strengths, recommendations, and implications for further research, policy and practice.
42
Chapter 4. The burden and risk of morbidity associated with infant male circumcision in a community-based setting
This chapter presents the first of five PhD papers that have been published or submitted for publication in peer reviewed journals. The paper presents findings of the research conducted to address the first objective of this thesis; to assess the burden and key determinants of morbidity associated with infant male circumcision in a community based setting in rural Ghana. This paper was published in the Journal of Tropical Medicine & International Health in 20175
(Appendix F).
4.1 Abstract
Background: Male circumcision services have expanded throughout Africa as part of a long-term Human Immunodeficiency Virus (HIV) prevention strategy.
However, there have been no community based studies of morbidities associated with infant male circumcision in sub-Saharan Africa.
Methods: The study assessed the effect of type of service provider (formal and informal) and hygiene practices on circumcision related morbidities in rural
Ghana. This population based cross-sectional study was conducted between May to December 2012 and involved 2850 circumcised infant males aged under 12 weeks. Multivariable logistic regression models were adjusted for maternal age, maternal education, income, birth weight and site of circumcision.
Findings: A total of 2850 (90.7 %) infant males were circumcised. Overall, the risk of experiencing a morbidity (defined as complications occurring during or after
43 the circumcision procedure as reported by the primary caregiver) was 8.1% (230).
Risk was not significantly increased if the circumcision was performed by informal providers (121, 7.2%) compared to formal health service providers (109,
9.8%) (adjusted odds ratio [aOR] 1.11, 95% CI 0.80–1.47, p=0.456). Hygiene practices were associated with significantly increased risk of morbidity, including: no hand washing (148 [11.7%]) (aOR 1.78, 95% CI 1.27–2.52, p=0.001); not cleaning circumcision instruments (174 [10.6%]) (aOR 1.80, 95% CI 1.27– 2.54, p=0.001); and uncleaned penile area (190 [10.0%]) (aOR 1.84, 95% CI 1.25–2.70, p=0.002).
Conclusion: The risk of morbidity following infant male circumcision in rural
Ghana was high. Poor hygiene practices were associated with increased morbidity risk. Government and non-government organisations need to improve training of circumcision providers in hygiene practices in sub-Saharan Africa.
4.2 Introduction
Male circumcision is the surgical removal of all or part of the foreskin of the penis and can be performed for cultural, religious, social and medical reasons2,14. Acute morbidities associated with infant male circumcision include: pain, bleeding, swelling, infection, tetanus and inadequate skin removal13,14. In sub-Saharan
Africa infant circumcisions are performed by formal health service providers including nurses, doctors and medical assistants in health facilities. They are also
44 performed by informal providers including: pharmacists, domestic staff, family members, and village based traditional service providers in the home21,30.
Approximately 90% of infant males were circumcised in West Africa and 95% in
Ghana in 20102. There has been an expansion of adult and infant male circumcision services throughout Africa as part of a long-term Human
Immunodeficiency Virus (HIV) prevention strategy18-20. Little is known about the burden and determinants of morbidity associated with infant male circumcision in low and middle income countries including West Africa. A recent global systematic review of 52 studies described an overall morbidity rate of 1.5% after infant male circumcision14. Nine of the studies in this review were from Africa but all were based in urban hospitals and none included follow-up after discharge14,22-29. It was also unclear whether the data on morbidities were obtained by parental report or if the infants were examined by trained fieldworkers.
The overall aim of this study was to assess the burden and key determinants of morbidity associated with infant male circumcision in a community based setting in rural Ghana. The primary objective was to assess the effect of type of service provider (formal and informal) on circumcision related morbidities. The secondary objective was to describe the effect of specific hygiene practices on circumcision related morbidities in the study population.
45
4.3 Methods
4.3.1 Setting
This was a population based cross-sectional study, conducted over a six month period from 21st May to 31st December 2012. Data were collected during the
Neovita trial which was a large population based randomised controlled trial investigating the impact of vitamin A administration to infants within two days of birth on infant mortality. Details of the Neovita trial are published elsewhere86.
The study area comprised nine contiguous districts in the Brong Ahafo Region of
Ghana covering 12,000 square kilometres85,86. Only 20% of the study population lived in the urban district administrative centres40. Subsistence farming was the main occupation of the population which was multi-ethnic with low educational levels. Four major district hospitals and 80 small health facilities provided health care services to the population. There were approximately 60 village based traditional informal circumcision providers (called ‘Wanzams’ in Ghana) and 100 formal (doctors, nurses and medical assistants) circumcision providers.
4.3.2 Data collection
All births in the Neovita study area were reported to the trial team via a network of fieldworkers and key informants. Fieldworkers visited families at home between two hours and two days after birth and collected data about the birth, socio-demographic characteristics and socioeconomic information (using an asset index). All families of enrolled babies were visited every month by study 46 fieldworkers until the infant reached 12 months of age to ascertain the vital status of the baby. Any infant death was followed-up by fieldworkers who conducted verbal autopsy interviews to ascertain the circumstances leading to the infant’s death.
All male liveborn Neovita infants were eligible to be included in the circumcision study. During a home visit when the infant was 8-11 weeks of age, the families were asked by trained senior fieldworkers for detailed information on: age at circumcision; place of circumcision; type of circumcision provider; methods and instruments used; hygiene practices before, during and after the circumcision; amount of foreskin removed; medicines prescribed. The fieldworkers asked about complications or morbidities and whether these problems had been diagnosed or confirmed by a trained health professional. Standardised paper based data collection tools (including a standardised list of closed ended morbidity questions) were used for all interviews. Visual aids (photographs and drawings) were used to assist with understanding all complications and morbidities. The fieldworkers also examined the infant’s penile area and recorded their findings. Families were asked if they sought care for circumcision related morbidities and if they had been given any advice about potential problems or care seeking. Consent to access the baby’s Neovita data including socioeconomic and demographic data was obtained.
47
Field workers were trained for two weeks in all study procedures including recording of data on circumcision related morbidities. Interrater reliability was checked between field workers. During the study field workers had scheduled and unscheduled supervisory visits from the study coordinator to assess data quality and consistency.
4.3.3 Study definitions
Reported morbidities were classified using specific case definitions. Case definitions were based on World Health Organization (WHO) and Joint United
Nations Programme on HIV/AIDS (UNAIDS) and the World Health
Organization program and manual for early infant male circumcision definitions for infant male circumcision1,2. An infant male was considered circumcised if all or part of the foreskin of the penis was removed. A formal health service provider referred to professionally trained, licensed and regulated providers of medical services. This included: doctors, medical assistants, or nurses. An informal provider was an untrained, unlicensed, unregulated private provider of medical services including: pharmacists, domestic staff, family members and Wanzams (village based traditional circumcision providers).
Circumcision morbidities were defined as complications occurring during or after the circumcision procedure as reported by the primary caregiver. Confirmed circumcision morbidities were defined as morbidities confirmed either by a senior fieldworker or health professional. The morbidity definitions are presented in
Table 4.1.
48
Table 4.1 Definitions of circumcision morbidity1,2 Circumcision morbidity Definition Excess bleeding Bleeding that cannot be stopped with compression Excess skin removal or incision Too much skin removed during circumcision or the entire skin on the penis shaft removed Inadequate skin removal Some foreskin remaining on penis after procedure or re-circumcision needed to remove residual foreskin Glans amputation Part or all of the glans penis was incised Ulcer Sores developed on the penis following circumcision Foreskin adhesions Small pieces of skin formed between any residual foreskin and the glans Fistula Hole in the wrong place on the penis where the infant urinates Urinary tract infection infection in the urine Infection of the penis Redness, tenderness or pus discharge from the wound Abnormal stream of urine Unusual strength and direction of urine stream Ballooning of the foreskin Foreskin separated from the glans of the penis but the foreskin opening remains narrowed
4.3.4 Statistical analysis
Circumcised infants who had died before the field worker visits were not excluded; the cause of morbidities and deaths were available through the WHO standard verbal autopsy form. The primary outcome measure was the prevalence of confirmed circumcision related morbidities among male infants aged under 12 weeks. This was compared between infants circumcised by formal health service providers and infants circumcised by informal providers.
Crude logistic regression models were first used to examine key determinants of morbidity associated with infant male circumcision. Odds ratios (ORs) and
95% confidence intervals (95% CI) were calculated. Multivariable logistic regression models were constructed apriori to adjust for the effect of important explanatory variables that were known to be associated with choice of health care provider and morbidity: maternal age, maternal educational level, income status, birth weight and site of circumcision. All analyses were conducted in
STATA version 13.
49
The research calculated that the almost 3000 infants included in this study would provide 80% power to detect a 33% relative effect on morbidity associated with infant circumcision when performed by formal compared to informal providers.
The study assumed a 5% significance level, a baseline circumcision morbidity risk of 10% and that the ratio between formal and informal providers would be
1:1.
4.3.5 Ethics
Ethical approvals were obtained from Ghana Health Service (GHS) Ethical
Review Committee, the Institutional Ethics Committee of Kintampo Health
Research Centre (KHRC), the Research Ethics Committee of London School of
Hygiene and Tropical Medicine and the Human Research Ethics Committee of
University of Western Australia (UWA).
Written informed consent was obtained from the legal guardian of male infants in the Neovita study to ask questions about their infant’s circumcision and associated morbidities. They were assured of their right to refuse consent without affecting their participation in the Neovita study and any other health services.
4.4 Results
From 21st May to 31st December 2012, there were 9100 live births in the
Neovita trial study area (Figure 4.1). A total of 8,110 (89%) infants were
50 recruited into the Neovita study. Of these, 4005 (49%) were male infants and
3142 (79%) were aged under 12 weeks. The families of 3141 (79%) male infants consented for their baby to be included in the circumcision study. Infant and family characteristics are presented in Table 4.2.
Overall, 2850 (90.7%) infants were circumcised and of those circumcised 93% were circumcised by 20 days of age. Three babies (0.1%) had no socio economic or demographic data collected and were excluded. Of the remaining 2847 infants, 230 (8.1%) had confirmed morbidities related to circumcision.
Age at circumcision ranged from 1 hour after birth to four weeks of age. One hundred and seventeen (4.0%) of the infants were circumcised between 0-6 days, 2558 (90.0%) between 7-20 days, and 172 (6.0%) 20 days and later. The mean age of circumcision was 11.3 days, standard deviation 5.96.
There were ten deaths in the 2847 infants who were circumcised. However, the fieldworkers were still able to collect information on circumcision practices and morbidities from the primary caregiver and were able to administer the verbal autopsy questionnaire. Two infants had excessive incisions and haemorrhagic complications from their circumcision which were considered to be a direct cause of their death. It was considered that haemophilia was not the cause of the excessive bleeding.
51
9100 live births in the study population
8110 (89.1%) live births enrolled in Neovita study
4105 (51%) female
4005 (49%) male live births
863 (21.5%) not seen within 12 weeks post birth 1 (0.1%) moved out of the study area
3141 (80%) males aged under 12 weeks enrolled in the circumcision study 291 (9.3%) not circumcised. 54 (18.6%) died, 153 (52.6%) decided to perform circumcision at a later date, 84 (28.9%) decided not to circumcise
2850 (90.7%) circumcised by age 12 weeks
3 infants without socio economic or morbidity data due to field worker error
2847 (99.9%) included in analysis
230 (8.1%) with confirmed circumcision associated morbidities including 2 deaths
Figure 4.1 Flow diagram for the circumcision study
52
Table 4.2 Infant and maternal characteristics in the study Characteristics Circumcision Uncircumcised Circumcised Missing data* Formal health service Informal provider study population infants infants provider n=3141 n=291 (9.3%) n=2850 (90.7%) n=3 (0.1%) n=1177 (41.3%) n=1670 (58.6%) Birth weight Less than 2.5kg 255 (8.1%) 41 (16.1%) 214 (83.9%) 0 (0.0%) 78 (36.5%) 136 (63.5%) 2.5kg or greater 2886 (91.9%) 250 (9.3%) 2636 (91.3%) 3 (0.1%) 1099 (41.7%) 1534 (58.3%) Maternal age (years) Less than 20 360 (11.5% 41 (11.4%) 319 (88.6%) 1 (0.3%) 126 (39.6%) 192 (60.4%) 20-29 1603 (51.0%) 145 (9.0%) 1458 (91.0%) 0 (0.0%) 604 (41.4%) 854 (58.6%) 30 or more 1170 (37.2%) 104 (8.9%) 1066 (91.1%) 2 (0.2%) 442 (41.5%0 622 (58.5%) Missing data 8 (0.3%) 1 (12.5%) 7 (87.5%) 0 (0.0%) 5 (71.4%) 2 (28.6%) Mother’s highest educational level None 601 (19.1%) 89 (14.8%) 512 (85.2%) 0 (0.0%) 153 (29.9%) 359 (70.1%) Primary 1619 (51.5%) 138 (8.5%) 1481 (91.5%) 1 (0.1%) 598 (40.4%) 882 (59.6%) Secondary 913 (29.1%) 63 (6.9%) 850 (93.1%) 2 (0.2%) 421 (49.7%) 427 (50.3%) Missing data 8 (0.3%) 1 (12.5%) 7 (87.5%) 0 (0.0%) 5 (71.4%) 2 (28.6%) Wealth quintile of household 1 (Lowest) 503 (16.0%) 115 (22.9%) 388 (77.1%) 1 (0.3%) 62 (16.0%) 325 (84.0%) 2 608 (19.4%) 75 (12.3%) 533 (87.7%) 1 (0.2%) 183 (34.4%) 349 (65.6%) 3 676 (21.5%) 47 (7.0%) 629 (93.1%) 1 (0.2%) 273 (43.5%) 355 (56.5%) 4 725 (23.1%) 38 (5.2%) 687 (94.8%) 0 (0.0%) 306 (44.5%) 381 (55.5%) 5 (Highest) 629 (20.0%) 16 (2.5%) 613 (97.5%) 0 (0.0%) 353 (57.6%) 260 (42.4%) Missing data 0 (0.0%) - - - - - Site of delivery Health facility 2479 (78.9%) 187 (7.5%) 2292 (92.5%) 1 (0.0%) 1050 (45.8%) 1241 (54.2%) Home 650 (20.7%) 101 (15.5%) 549 (84.5%) 2 (0.4%) 122 (22.3%) 425 (77.3%) Missing data 12 (0.4%) 3 (25.0%) 9 (75.0%) 0 (0.0%) 5 (55.6%) 4 (44.4%) Religion Christian 2234 (71.1%) 186 (8.3%) 2048 (91.7%) 3 (0.1%) 964 (47.1%) 1081 (52.9%) Muslim 727 (23.1%) 61 (8.4%) 666 (91.6%) 0 (0.0%) 170 (25.5%) 496 (74.5%) Traditional 171 (5.4%) 42 (24.6%) 129 (75.4%) 0 (0.0%) 38 (29.5%) 91 (70.5%) African/none Missing data 9 (0.3%) 2 22.2%) 7 (77.8%) 0 (0.0%) 5 (71.4%) 2 (28.6%) * Three circumcised infants had no socioeconomic and demographic data due to field worker error
53
1670 (58%) circumcisions were performed by informal providers (Wanzams 979
[34%]; domestic staff 454 [16%]; pharmacists 175 [6%]; 62 [2%] other informal providers) and 1177 (42%) were performed by formal health service providers
(nurses 958 [34%]; doctors 136 [5%]; and medical assistants 83 [3%]). A total of
1369 (48%) circumcisions were performed in a health facility and 1478 (52%) at home. Of those circumcised in a health facility, 516 (18%) were circumcised by informal providers [domestic staff 454 (16%), and 62 (2%) by other informal providers]. Sutures were used in 81 (3%) of the circumcisions. Two hundred and thirty (8.1%) of the circumcised infants had confirmed circumcision related morbidities (Table 4.3).
The most common morbidities were excess skin removal or incision (97, 42.2%), excess bleeding (92, 40%), inadequate skin removal (33, 14.3%), infection (23,
10%) and abnormal stream of urine (23, 10%) (Table 4.3). Haematomas were not reported as a complication by any mother in this study. One hundred and thirty three (57.8%) morbidities were reported by the primary caregiver to have occurred less than 24 hours after circumcision.
The study obtained complete data for most health care practices (handwashing, use of gloves, use of cleaned instruments, cleaning of the penile area and use of clean bandages). However, 1279 (44.9 %) respondents could not recall whether the circumcision providers used a circumcision device and 1256 (44.1 %) could not recall whether the circumcision providers used a free hand instrument
54
(n=2847). Overall, 1579 (55.5%) of the providers were reported to have washed their hands before the circumcision. Nine hundred and thirty nine (33%) cleaned the penile area before the circumcision and 1415 (88.9%) used a freehand instrument (i.e. used a scalpel without any supportive device) (Table
4.3).
Formal health care providers were significantly less likely to wash their hands
(adjusted odds ratio [aOR] 0.46, 95% CI 0.40–0.55, p < 0.001), clean circumcision instruments (aOR 0.71, 95% CI 0.61–0.83, p < 0.001), and clean the penile area
(aOR 0.59, 95% CI 0.50–0.70, p < 0.001) compared to informal providers (Table
4.3). Informal providers had significantly worse health care practices for use of gloves (52.2%), use of circumcision devices (17.0%) and application of clean bandages (61.4%). The use of free hand instruments appeared similar between the formal and informal providers. (Table 4.3).
One hundred and nine (47.4%) morbidities followed circumcision by formal health service providers (nurses [84, 77.1%], doctors [20, 18.3%] and medical assistants [5, 4.6%]. One hundred and twenty one (52.6%) morbidities followed circumcision by informal providers including Wanzams (54, 44.6%), domestic staff (50, 41.3%), pharmacists (13, 10.7%) and other informal providers (4, 3.3%).
Overall, the risk of experiencing a morbidity following circumcision was not significantly increased if the circumcision was performed by an informal
55 provider (109 [9.3%]) compared to a formal health service provider (121 [7.2%])
(aOR 1.11, 95% CI 0.80–1.47, p=0.456) (Table 4.4).
56
Table 4.3 Circumcision related morbidities and health care practices by provider type in the study population Number of Number of Formal health informal Unadjusted odds p-value Adjusted odds p-value infants with morbidities / service provider provider ratio (95% CI) ratio (95% CI) data* care practices n = 1177 (41.3%) n = 1670 (58.7%) Morbidities ** Total morbidities 2847 230 (8.1%) 109 (9.3%) 121 (7.2%) 1.31 (1.00-1.71 ) 0.05 0.89 (0.66-1.21) 0.47 Excess skin removal or incision 2847 97 (3.4%) 54 (4.6%) 43 (2.6%) 1.45 (1.10-1.92) 0.01 1.02 (0.75-1.39) 0.91 Excess bleeding 2847 92 (3.2%) 46 (3.9%) 46 (2.8%) 1.44 (0.95-2.18) 0.09 1.27 (0.78-2.05) 0.33 Inadequate skin removal 2847 33 (1.8%) 11 (0.9%) 22 (1.3%) 0.71 (0.43-1.46) 0.35 0.43 (0.19-0.97) 0.04 Abnormal stream of urine 2847 23 (0.8%) 11 (0.9%) 12 (0.7%) 1.17 (0.58-2.36) 0.65 1.20 (0.54-2.70) 0.65 Infection 2847 23 (0.8%) 11 (0.9%) 12 (0.7%) 1.00 (0.76-1.32) 0.99 0.96 (0.68-1.34) 0.81 Glans amputation 2847 21 (0.7%) 13 (1.3%) 8 (0.4%) 2.32 (0.96-5.62) 0.06 1.17 (0.43-3.17) 0.76 Ballooning of the foreskin 2847 21 (0.7%) 8 (0.7%) 13 (0.8%) 0.87 (0.36-2.11) 0.06 0.50 (0.19-1.33) 0.17 Ulcer 2847 18 (0.6%) 10 (0.9%) 8 (0.5%) 1.78 (0.70-4.52) 0.23 0.92 (0.33-2.54) 0.87 Other urination problem 2847 14 (0.5%) 6 (0.5%) 8 (0.5%) 4.99 (1.03-24.06) 0.05 4.08 (0.62-26.94) 0.14 Other problem related to 2847 9 (0.3%) 7 (0.6%) 2 (0.1%) 1.06 (0.37-3.11) 0.91 0.43 (1.14-1.25) 0.12 circumcision Health care practices ***
Handwashing 2847 1579 (55.5%) 508 (43.2%) 1071 (64.1%) 0.42 (0.36-0.49) <0.01 0.47 (0.40-0.55) <0.01
Use of gloves 2847 1772 (62.2%) 900 (76.5%) 872 (52.2%) 2.97 (2.25-3.50) <0.01 2.69 (2.27-3.18) <0.01
Cleaning of instruments 2847 1208 (42.4%) 427 (36.3%) 781 (46.8%) 0.65 (0.56-0.76) <0.01 0.71 (0.61-0.83) <0.01
Cleaning of penile area 2845 939 (33.0%) 292 (24.9%) 647 (38.8%) 0.52 (0.44-0.62) <0.01 0.59 (0.50-0.70) <0.01
Use of circumcision device 1568 377 (24.0%) 196 (39.1%) 181 (17.0%) 3.15 (2.47-4.00) <0.01 3.01 (2.33-3.89) <0.01
Use of free hand instruments 1591 1415 (88.9%) 460 (91.3%) 955 (87.9%) 1.45 (1.00-2.07) 0.04 1.65 (1.13-2.40) 0.01 without support (eg scalpel) Application of bandage to penis 2847 1995 (70.1%) 969 (82.3%) 1026 (61.4%) 2.92 (2.44-3.50) <0.01 2.54 (2.11-3.05) <0.01 after circumcision * Infants were excluded if the carer stated that they could not recall / did not know the answer to the question **Adjusted for maternal education, maternal age, income status, birth weight, site of circumcision ***Adjusted for income status
57
There was a 1.5-2.0 fold increased risk of morbidity associated with key health care practices: the provider not having washed their hands 1268 (44.5%) (aOR
1.78, 95% CI 1.27 – 2.52, p < 0.001); not cleaning instruments 1639 (57.6%) (aOR
1.80, 95% CI 1.27–2.54, p = 0.001); not cleaning the penile area 1906 (66.9%) (aOR
1.84, 95% CI 1.25–2.70, p = 0.002); not using freehand instruments 176 (11.1%)
(aOR 1.89, 95% 1.09-3.26, p=0.002) (Table 4.4).
Site of circumcision was significantly associated with circumcision related morbidity. Babies who were circumcised in a health facility (1369, 48.1%) had a
2.4 fold greater risk of circumcision related morbidity compared to babies who were circumcised at home (1478, 51.9%) (aOR 2.42, 95% CI 1.80-3.27 p < 0.001)
(Table 4.4). A likely reason for the increased frequency of morbidity in health facilities and non-health facilities in the study area is inadequate training in surgery and/ or poor supervision.
Babies circumcised at 20 days and later (172, 6.0%) had an increased risk of morbidity compared with infants who were circumcised earlier (2558, 90.0%)
(aOR 1.89, 95% CI 1.18–3.03, p = 0.008) (Table 4.4). Excess skin removal or incision, excess bleeding, abnormal stream of urine, other urination problems and infection were reported more often in the older age group (>20 days) compared to younger infants (0-6 days).
58
Babies born to mothers with no education (512, 18.0%) had a 2.4 fold increased risk of circumcision related morbidity compared to babies born to mothers with secondary level education (850, 29.9%) (aOR 2.02, 95% CI 1.31 – 3.11 p = 0.001), and a 1.5 fold increased risk compared to mothers who had received primary level education (1481, 52.1%) (aOR 1.48, 95% CI 1.05 – 2.07 p = 0.024) (Table 4.5).
59
Table 4.4 Health service determinants of circumcision associated morbidity in the study population Determinants Number of Number of % Unadjusted Odds Ratio p-value Adjusted Odds Ratio* p-value infants morbidities (95% CI) (95% CI) n = 2847 n = 230 Type of health service provider Formal 1177 109 9.3% 1.00 0.05 1.00 0.60 Informal 1670 121 7.2% 0.76 (0.58 – 1.00) 1.08 (0.80 – 1.47) Site of circumcision Health facility 1369 155 11.3% 2.39 (1.79 – 3.17) 2.47 (1.84 – 3.31) Home 1478 75 5.1% 1.00 <0.01 1.00 <0.01 Handwashing Yes 1579 82 5.2% 1.00 <0.01 1.00 <0.01 No 1268 148 11.7% 2.41 (1.82 – 3.19) 1.79 (1.27 – 2.53) Use gloves Yes 1772 143 8.0% 1.00 0.98 1.00 0.33 No 1075 87 8.1% 1.00 (0.75 – 1.32) 1.15 (0.86 – 1.54) Instruments cleaned before circumcision Yes 1208 56 4.6% 1.00 <0.01 1.00 <0.01 No 1639 174 10.6% 2.44 (1.79 – 3.33) 1.80 (1.27 – 2.55) Penile area cleaned before circumcision Yes 939 40 4.3% 1.00 <0.01 1.00 <0.01 No 1906 190 10.0% 2.48 (1.75 – 3.53) 1.84 (1.25 – 2.71) Carer did not know 2 0 0.0% - Use of circumcision device Yes 377 24 6.4% 1.06 (0.66 – 1.70) 1.01 (0.62 – 1.63) No 1191 72 6.1% 1.00 <0.01 1.00 0.69 Carer did not know 1279 134 10.7% Use of free hand instrument Yes 1415 78 5.5% 1.00 <0.01 1.00 0.05 No 176 18 10.2% 1.95 (1.14 – 3.35) 1.87 (1.09 – 3.22) Carer did not know 1256 134 10.7% Application of bandage to penis after circumcision Yes 1995 174 8.7% 1.00 0.05 1.00 0.76 No 852 56 6.6% 0.73 (0.53-1.00) 0.95 (0.68-1.32) Age at circumcision 0-6 days 117 13 11.1% 1.53 (0.84 – 2.77) 1.75 (0.96 – 3.21) 7-20 days 2558 193 7.5% 1.00 0.01 1.00 0.01 >20 days 172 24 14.0% 1.99 (1.26 – 3.13) 1.89 (1.18 – 3.04) *Adjusted for maternal education, maternal age, income status, birth weight, site of circumcision 60
Table 4.5 Socio demographic determinants of circumcision associated morbidity in the study population Determinants Number of Number of % Unadjusted Odds Ratio p-value Adjusted Odds Ratio p-value infants morbidities (OR) (95% CI) (95% CI) n = 2847 n = 230 Birth weight Less than 2.5kg 214 21 9.8% 1.24 (0.78 – 1.99) 1.10 (0.68 – 1.79) 2.5kg or greater 2633 209 7.9% 1.00 0.38 1.00 0.62 Maternal age (years) Less than 20 318 26 8.2% 1.20 (0.76 – 1.90) 1.35 (0.83 – 2.17) 20-29 1458 130 8.9% 1.28 (0.95 – 1.71) 1.36 (1.00 – 1.83) 30 or more 1064 74 7.0% 1.00 0.26 1.00 0.11 Missing data 7 0 0.0% - - - Mother’s highest educational level None 512 50 9.8% 1.56 (1.05 – 2.33) 2.01 (1.31 – 3.10) Primary 1481 127 8.6% 1.37 (0.99 – 1.90) 1.48 (1.06 – 2.08) Secondary 850 53 6.2% 1.00 0.06 1.00 0.01 Missing data 4 0 0.0% - - - - Wealth quintile of household 1 (Lowest) 387 28 7.2% 0.87 (0.54 – 1.40) 0.97 (0.58 – 1.62) 2 532 41 7.7% 0.92 (0.60 – 1.41) 0.89 (0.57 – 1.39) 3 628 47 7.5% 0.89 (0.59 – 1.34) 0.80 (0.53 – 1.22) 4 687 62 9.0% 1.07 (0.73 – 1.57) 1.02 (0.69 – 1.51) 5 (Highest) 613 52 8.5% 1.00 0.86 1.00 0.77 Missing data 0 0 0.0% - - - - Site of delivery Facility delivery 2292 196 8.6% 1.00 0.11 1.00 0.44 Home delivery 549 34 6.2% 0.75 (0.52 – 1.08) 0.87 (0.58 – 1.31) Missing data 6 0 0.0% - - - Religion Christian 2045 172 8.4% 1.00 0.54 1.00 0.86 Muslim 666 49 7.4% 0.85 (0.61 – 1.18) 0.93 (0.65 – 1.31) Traditional African/None 129 9 7.0% 0.80 (0.40 – 1.61) 0.91 (0.44 – 1.85) Missing data 7 0 0.0% - - - - *Adjusted for maternal education, maternal age, income status, birth weight, site of circumcision
61
4.5 Discussion
In this population based study in rural Ghana infant male circumcision was almost universal. A total of 93% of infant males in the study area were circumcised by 20 days of age. The study reported a high risk of morbidity
(8.1%) and two deaths following circumcision. Important determinants included poor hygiene practices such as not washing hands, the use of uncleaned instruments and not cleaning the penile area. Circumcision performed by formal providers was associated with similar risks to those performed by informal providers. However, hygiene practices appeared poorer in formal providers. Babies born to women with low education levels were particularly at risk.
It was not surprising that the morbidities and mortality observed in this study was higher than hospital-based studies from high-income settings including
USA, UK and Israel14,77,93. However, the morbidity risk was lower than the three published African studies which were conducted in large urban hospitals in
Nigeria and Botswana. These studies reported an 18% (range 10-24%) risk of morbidity 6-16 weeks post infant male circumcision27,65,69. The case definitions used in those studies were similar to this study. However, infants with high complication risks were admitted to hospital and cannot be considered representative of the whole population. This study appears to be the first to analyse the population level burden associated with circumcision in the African community setting. 62
Overall, the prevalence of circumcision was similar to published rates from other parts of West and North Africa2,27,65,94. The most common morbidities in this study were: excess skin removal, excess bleeding, inadequate skin removal, infection and abnormal stream of urine. These findings were similar to morbidities reported by other similar studies27,65,69,93,95. The reasons for this are likely to be poor training and quality of care. Infection and hygiene practices do play a part in causing poor wound healing and excess bleeding.
The study reported two deaths due to excessive incisions and haemorrhage but no deaths due to infection or tetanus.
Almost 60% of infant male circumcisions were performed by informal providers. Other studies report that Wanzams were the most common providers of circumcision in rural West Africa2,21. In this study, the risk of experiencing a morbidity was not significantly increased if circumcision was performed by informal providers compared to formal health service providers.
Other studies have reported both higher27,65,95 and lower2,14,17,52 risks of morbidity after circumcision performed by informal providers. However, these findings were consistent with a Lancet review published in 2007 which reported that circumcision related morbidities were common in both formal and informal providers96.
Overall, hygiene practices were inadequate in the study area. These findings were consistent with recent studies which reported unclean and unsterile 63 circumcision practices amongst urban and rural African service providers2,21,73.
Concerningly, the research found that hygiene practices such as handwashing and cleaning of the penile area were poorer in formal health care providers compared to informal providers. Reasons for this were unclear and will be investigated in a future study. The higher morbidity amongst the informal providers is likely to be due to a different reason. It was likely to be due to the unclean environment in which the informal circumcisions were being conducted and the unclean home environments of the infants.
There was a twofold increased risk of morbidity from circumcisions conducted after 20 days of life. Similar findings were reported in a recent systematic review and studies from Kenya, Pakistan and the USA14,71,74-76. Older infants are at increased risk of complications because the procedure is more difficult in older infants. That is, it is harder for primary care providers to sedate or anaesthetise older infants and older infants struggle more. Furthermore, older infants are more mobile and the wounds are more likely to get dirty and infected. The reasons for the lower risk of morbidity in younger infants was considered to be due to the simpler nature of the procedure during this period and faster healing14,29,75. Infants born to mothers with no formal education had a greater risk of circumcision related morbidity. The reasons were likely to be due to the fact that women with no formal education were more likely to be from
64 poor families who have (i) less knowledge about hygiene practices and (ii) who are less able to create a clean uncontaminated living environment.
The study had some limitations. The research was not able to conduct in depth qualitative interviews to explore the experiences of families and circumcision providers. The study was unable to ascertain other important outcomes such as pain experienced by the infants. This study was observational and cross- sectional and does not provide proof of causation. Cross-sectional studies are also prone to differential misclassification and may report greater associations between potentially concerning health care practices (e.g. lack of handwashing) and maternal reports of morbidity. However, the research did not rely on maternal recall alone, visual aids were used to clarify responses, and all infants were examined by trained fieldworkers. Families were blinded to the study hypotheses and questions about risk factors and determinants were all asked before the questions about morbidity. The association between poor health care practices and morbidity risk has strong biological plausibility e.g. the risk of infection associated with poor handwashing and lack of cleaning of the penile area and instruments. Forty four percent of respondents could not recall the type of instruments used for the circumcision but data on all other variables were almost complete. Other strengths of this study included the standardised questionnaires, and well-trained, experienced fieldworkers who have been involved in many large trials in the study area. The fieldworkers were blinded to the study hypothesis therefore reporting bias was low. This study was also 65 population based and provided data on infants circumcised in both the formal and informal healthcare systems in rural Africa.
The study has implications for policy and programme development. Infant male circumcision is almost universal in many parts of sub Saharan Africa including our study area in rural Ghana. The study reported two deaths and a high risk of morbidity following circumcision in male infants. Hygiene related health care practices were poor amongst circumcision providers and clearly linked to the morbidity risk. These findings indicated that the Ghanaian
Ministry of Health and other non-government organisations need to improve training of formal and informal care providers in hygiene practices including hand washing and cleaning of circumcision instruments and the penile area.
Investment in training will improve service provider skills, reduce circumcision risk and improve safety during circumcision procedures for young infants in rural African settings.
66
Chapter 5. Factors which influence choice of infant male circumcision provider in rural Ghana
This chapter presents the second PhD paper which was the research conducted to address the objective two of this thesis; to determine the key factors which influence the choice of infant male circumcision provider in rural Ghana. This paper was published in BMC Pediatrics in 201797 (Appendix F).
5.1 Abstract
Background: The influence of socio-economic determinants on choice of infant male circumcision provider is not known in areas with high population coverage such as rural Africa. The overall aim of this study was to determine the key socio-economic factors which influence the choice of infant male circumcision provider in rural Ghana.
Methods: The study investigated the effect of family income, distance to health facility, and cost of the circumcision on choice of infant male circumcision provider in rural Ghana. Data from 2847 circumcised infant males aged under 12 weeks and their families were analysed in a population based cross-sectional study conducted from May to December 2012 in rural Ghana. Multivariable logistic regression models were adjusted for income status, distance to health facility, cost of circumcision, religion, maternal education, and maternal age.
Results: Infants from the lowest income households (325, 84.0%) were more likely to receive circumcision from an informal provider compared to infants from the highest income households (260, 42.4%) even after adjusting for religious affiliation (adjusted odds ratio [aOR] 4.42, 95% CI 3.12-6.27 p=<0.001). 67
There appeared to be a dose response with increasing risk of receiving a circumcision from an informal provider as distance to a health facility increased
(aOR 1.25, 95 CI 1.30-1.38 P=<0.001). Only 9.0% (34) of families in the lowest socio-economic quintile received free circumcision services compared to 27.9%
(171) of the highest income families.
Conclusions: The Government of Ghana and Non-Government Organisations should consider additional support to poor families so they can access high quality free infant male circumcision in rural Ghana.
5.2 Background
Globally, male infants are circumcised mostly for medical and religious reasons2,55. Male circumcision has been reported in a number of high quality trials to reduce HIV infection in adult males who live in communities with high HIV prevalence such as South and East Africa3,9. Other benefits of male circumcision include penile hygiene, protective effects for cancer of the penis, reduced urinary tract infections, protective effects for sexually transmitted infections and reduction of genital herpes12. Approximately 91% of infant (age 0-11 months) males are circumcised in Ghana5 and other West African countries2. My previous study reported high risks of concerning health care practices and morbidities following infant male circumcision in our community based study in rural
Ghana5. Fifty eight percent of circumcisions were performed by informal providers; including Wanzams (village based traditional circumcision providers), family members, and drug sellers. 68
Initiatives to improve the health care practices of Wanzams and other circumcision providers are underway21,30. These include training on infection control, instruments to perform circumcision and hygiene. However, other strategies to influence family’s care seeking patterns, improve use of health facilities, and improve use of trained circumcision providers are also needed.
This requires an understanding of the key factors which influence a family’s choice of circumcision provider. A recent systematic review reported that socio- economic factors such as income, location (rural and urban), and cost of the circumcision were key determinants of choice of health service provider for infant male circumcision2. Socio-economic status, cost, and geographical access are also key determinants of care seeking for antenatal and birthing care in sub-
Saharan African populations31-36. However, to my knowledge, there have been no studies from rural Africa that have investigated the effect of these factors on choice of infant male circumcision provider.
Thus, the overall aim of this study was to determine the key socio-economic factors which influence the choice of infant male circumcision provider in rural
Ghana. The primary objective was to determine if socio-economic status was an important determinant of choice of circumcision provider. The secondary objectives were to assess the associations between distance to health facilities and cost of circumcision on choice of circumcision provider.
69
5.3 Methods
5.3.1 Study design and setting
This was a community level population based cross-sectional study conducted in the Brong Ahafo Region of central Ghana from 21st May 2012 to 31st
December 2012. Data were collected during a large neonatal vitamin A supplementation trial (Neovita) and full details are published elsewhere86. At the time of the circumcision study, 80% of the study population lived in rural settlements and almost 20% of mothers did not have primary school education.
Four major district hospitals and 80 small health facilities provided health care services to the population. There were approximately 60 Wanzams and 100 formal circumcision providers (doctors, nurses, and medical assistants) at the time of the study.
5.3.2 Data collection
All births in the Neovita study area were reported to the trial team via a network of fieldworkers and key informants. Fieldworkers visited all families at home between two hours and two days after birth and interviewed the mother of the infant, or the primary care giver. Fieldworkers weighed the baby and asked the mother or the primary care giver about: date of birth, site of birth, current address, distance to health facilities, socio-demographic characteristics, and socio-economic information (using an asset index). The fieldworkers also collected data on the vital status of the baby (including if the baby was alive, dead, or hospitalised).
70
Only male liveborn Neovita infants who were aged under 12 weeks were included to ensure the most accurate recall of circumcision related events.
Infants were included in the Neovita trial if they were aged under three days, able to feed, were staying in the study area for at least six months after enrolment and their mother provided written informed consent. Follow-up visits were scheduled between eight to eleven weeks post birth and trained senior fieldworkers asked for consent to collect additional detailed data on: age at circumcision, site of circumcision, and type of circumcision provider. Infant male circumcision was supposed to be covered under the Ghana Health
Insurance Scheme but it was well known that fees for circumcisions were charged by some formal and informal providers. So we also asked families if they had to pay any fees or “in-kind” contributions for the circumcision.
Families were also asked if the study team could have access to the baby’s
Neovita data including socio-economic, and socio-demographic data.
Fieldworkers were trained for two weeks in all study procedures prior to the commencement of the study. Interrater reliability was checked between all fieldworkers. During the study fieldworkers received scheduled and unscheduled supervisory visits from the study coordinator to assess data quality and consistency. The fieldworkers used standardised paper based data collection tools (including a standardised list of closed ended questions) for all interviews.
71
5.3.3 Study definitions and categories
In this study a ’formal circumcision provider’ was defined as a professionally trained, licensed, and regulated provider of circumcision services. This included: doctors, medical assistants, or nurses2. An ‘informal circumcision provider’ was an untrained, unlicensed, unregulated private provider of circumcision services including: Wanzams (village based traditional circumcision providers), drug sellers, and family members2,4,21. To assess ‘income status’ an asset index was constructed based on data collected on household assets (ownership of animals, television, motorcycle, etc) and housing material
(walls, floor, windows, and roof). The index was analysed using principal component analysis (PCA) in Stata version 13 and categorised into five income quintiles98. ‘Distance to a health facility’ was measured in kilometres using
Geographic Information System (GIS) software and the most commonly used roads from each village to the nearest health facility. It was categorised into four levels (<1km, 1-4.9km, 5-9.9km, 10 km or more). Many of the families in our study had limited recall about the exact cash amounts they paid for their circumcision but could categorise their responses. Thus information on the exact cash amounts for ‘cost of the circumcision’ was not collected and data were collected in the following categories: free, not free but less than 10 Ghana Cedis
(Ghs), between 10 and 20 Ghs, 20 Ghs or more (at the time of conducting the study 1 Ghs = 0.6 $US)99. ‘In kind contributions’ were defined as any non-cash payment to the formal or informal provider for the circumcision (e.g. bars of soap, chickens, kola nuts, and corn). 72
5.3.4 Statistical analysis
Crude logistic regression models were used to examine the effect of income status on type of circumcision provider (informal vs formal). Odds ratios (ORs) and 95% confidence intervals (95% CI) were calculated. Multivariable logistic regression models were constructed apriori to adjust for the effect of important explanatory variables (income status, cost of circumcision, religion, maternal education, maternal age and distance to health facility). Model one assessed each of the infant and maternal characteristics as determinants of choice of informal provider, adjusting for income status, cost of circumcision, religion, maternal education and maternal age. Model two is the same as model one with an additional adjustment for distance to health facility. All analyses were conducted using STATA version 13.
I calculated that the 2800 infants included in this study would provide 80% power to detect at least a 20% effect due to income status on choice of circumcision provider. I assumed a 5% significance level and a baseline 60% risk of receiving circumcision from an informal circumcision provider5.
5.3.5 Ethical issues
Ethical approvals were obtained from Ghana Health Service Ethical Review
Committee, the Institutional Ethics Committee of Kintampo Health Research
Centre, the Research Ethics Committee of London School of Hygiene and
Tropical Medicine, and the Human Research Ethics Committee of the 73
University of Western Australia. Written informed consent was obtained from all the families of the circumcised male infants.
5.3.6 Role of funding source
The funders had no role in data gathering, data analysis, or writing of the report. The corresponding author had full access to all the data in the study, and for the decision to submit for publication.
5.4 Results
There were 9100 live births in the Neovita trial study area from 21st May to 31st
December 2012 (Figure 5.1). A total of 8,110 (89%) liveborn infants were recruited into the Neovita study. Forty nine percent (4005) were male infants and 78% (3141) were aged under 12 weeks. Of the 3141 eligible male infants,
2850 (90.7%) were circumcised. Two hundred and ninety one (9.3%) infants were not circumcised within 12 weeks after birth. Of these, 153 (52.6%) were circumcised at a later date, 84 (28.9%) were never circumcised and 54 (18.6%) died. Three circumcised babies (0.1%) had no socio-economic or demographic data collected and were excluded in the statistical analysis of associations between socio-economic or demographic factors and choice of circumcision provider.
74
9100 live births in the study population
8110 (89.1%) live births enrolled in Neovita study
4105 (51%) female
4005 (49%) male live births
863 (21.5%) not seen within 12 weeks post birth. 1 (0.0%) moved out of the study area
3141 (78.4%) males aged under 12 weeks enrolled in the circumcision study
291 (9.3%) not circumcised. 54 (18.6%) died, 153 (52.6%) decided to perform circumcision at a later date, 84 (28.9%) decided not to circumcise 2850 (90.7%) circumcised by age 12 weeks
3 infants without socio economic, demographic or circumcision provider data due to field worker error
2847 (99.9%) included in the analysis of determinants of circumcision provider*
*Includes the 54 (18.6%) who died. These families were still interviewed and provided full information about circumcision thus their data were included
Figure 5.1 Flow diagram for the study: factors which influence choice of circumcision provider
Of the remaining 2847 circumcised infants, 1670 (59%) were circumcised by informal providers and 1177 (41%) by formal health service providers (Table
5.1). Three hundred and eighty seven (13.6%) were in the lowest socio-economic quintile, 186 (6.7%) lived 10km or more from a health facility, and 512 (18.0%)
75 mothers of circumcised infants had no primary school education (Table 5.1). A total of 666 (23.4%) mothers of circumcised infants were Muslim, and 549
(19.3%) delivered at home (Table 5.1). Five hundred and thirty nine (18.9%) infants received their circumcision free of charge (Table 5.2). A total of 2229
(78.3%) families paid some form of cash currency (between 1-100 Ghana Cedis
[approximately 0.60 to 55 $US]) for their infant’s circumcision and 87 (3.1%) families paid in-kind contributions in the form of bars of soap, chickens, kola nuts, and corn (Table 5.3).
Infants from the lowest income households (quintile 1) (325, 84.0%) were four times more likely to receive a circumcision from an informal provider compared to infants from the highest income households (260, 42.4%) (adjusted odds ratio [aOR] 4.42, 95% CI 3.12-6.27 p=<0.001) (Table 5.2). There also appeared to be a ‘dose response’ with increasing risk of receiving a circumcision from an informal provider as income status decreased (Table 5.2)
(aOR 1.34, 95% CI 1.25-1.43 p=<0.001). A total of 2229 (78.3%) families paid to receive circumcision services from both formal and informal circumcision providers (Tables 5.2 and 5.3). Five hundred and thirty nine (18.9%) infants received their circumcision free of charge (50.1% formal and 49.9% informal)
(Tables 5.2 and 5.3). Only 6.9% (68) of Wanzams provided free circumcisions.
76
Table 5.1 Infant and maternal characteristics in the study population Characteristics Uncircumcised Circumcised infants Total circumcised infants n=291 included in the infants analysis n=2847 n=2850* (99.9%) Income status of household (quintile) 1 (Lowest) 115 (39.5%) 387 (13.6%) 387 (13.6%) 2 75 (25.8%) 532 (18.7%) 532 (18.7%) 3 47 (16.2%) 628 (22.1%) 628 (22.1%) 4 38 (13.0%) 687 (21.1%) 687 (24.1%) 5 (Highest) 16 (5.5%) 613 (21.5%) 613 (21.5%) Missing data - - 3 (0.1%) Distance to health facility <1 km 101 (34.7%) 1444 (50.7%) 1444 (50.7%) 1 – 4.9 km 60 (20.6%) 741 (26.0%) 741 (26.0%) 5 – 10 km 84 (2.9%) 400 (14.0%) 400 (14.0%) 10 km or more 45 (15.5%) 186 (6.5%) 186 (6.5%) Missing data 1 (0.3%) 76 (2.7%) 79 (2.8%) Cost of circumcision** Free - 539 (18.9%) 539 (18.9%) Less than 10 Ghs - 145 (5.1%) 145 (5.1%) Between 10-20 Ghs - 1530 (53.7%) 1530 (53.7%) 20 Ghs or more - 554 (19.5) 554 (19.4) Missing data - 79 (2.8%) 82 (2.9%) Maternal occupation Gov’t/private 4 (1.4%) 105 (3.7%) 105 (3.7%) employed Self-employed 53 (18.2%) 927 (32.6%) 927 (32.5%) Farming 101 (34.7%) 589 (20.7%) 589 (20.7%) Not working 49 (16.8%) 484 (17.0%) 484 (17.0%) Missing data 84 (28.9%) 742 (26.0%) 745 (26.4%) Maternal highest educational level None 89 (30.6%) 512 (18.0%) 512 (18.0%) Primary 138 (47.2%) 1481 (52.0%) 1482 (52.0%) Secondary 63 (21.6%) 850 (29.9%) 852 (29.9%) Missing data 1 (0.3%) 4 (0.1%) 4 (0.1%) Religion Christian 186 (8.3%) 2048 (71.9%) 2048 (71.9%) Muslim 61 (8.4%) 666 (23.4%) 666 (23.4%) Traditional 42 (24.6%) 129 (4.5%) 129 (4.5%) African/none Missing data 2 (0.7%) 4 (0.1%) 7 (0.2%) Maternal age (years) Less than 20 41 (14.1%) 319 (11.2%) 319 (11.2% 20-29 145 (49.8%) 1458 (51.2%) 1458 (51.2%) 30 or more 104 (35.7%) 1066 (37.4%) 1066 (37.4%) Missing data 1 (0.3%) 4 (0.1%) 7 (0.2%) Site of delivery Health facility 187 (64.3%) 2292 (80.5%) 2292 (80.4%) Home 101 (34.7%) 549 (19.3%) 549 (19.3%) Missing data 3 (1.0%) 6 (0.2%) 9 (0.3%) Birth weight Less than 2.5kg 41 (14.1%) 214 (7.5%) 214 (7.5%) 2.5kg or greater 250 (85.9%) 2633 (92.5%) 2636 (92.5%) Missing data - - - Type of circumcision provider - Formal provider - 1177 (41.3%) 1177 (41.3%) Informal provider - 1670 (58.7%) 1670 (58.6%) Missing data - 0 (0.0%) 3 (0.1%) 77
Age at circumcision 0-6 days - 117 (4.1%) 117 (4.1%) 7-20 days - 2556 (89.8%) 2556 (89.7%) > 20 days - 172 (6.0%) 172 (6.0%) Missing data - 2 (0.1%) 5 (0.2%) *Three circumcised infants had no socioeconomic and demographic data due to field worker error
** 1 Ghs = 0.6 $US (2012)
78
Table 5.2 Determinants of choice of informal provider for infant male circumcision Determinants Total number Number (%) of Unadjusted odds Adjusted odds ratio Adjusted odds ratio of infants infants who received ratio (95% CI) (95% CI) model 1* (95% CI) model 2** circumcision from an informal provider n = 2847 n= 1670 (58.7%) Income status of household (quintile) 1 (Lowest) 387 325 (84.0%) 7.12 (5.19 – 9.76) 5.77 (4.15 – 8.02) 4.42 (3.12 – 6.27) 2 532 349 (65.6%) 2.59 (2.04 – 3.29) 2.26 (1.76 – 2.91) 1.89 (1.45 – 2.47) 3 628 355 (56.5%) 1.77 (1.41 – 2.21) 1.67 (1.32 – 2.11) 1.49 (1.17 – 1.90) 4 687 381 (55.5%) 1.69 (1.36 – 2.11) 1.58 (1.26 – 1.98) 1.48 (1.17 – 1.87) 5 (Highest) 613 260 (42.4%) 1.00 1.00 1.00 Missing data 0 - - - - Distance to health facility <1 km 1444 776 (53.8%) 1.00 1.00 1.00 1 – 4.9 km 741 433 (58.4%) 1.21 (1.01 – 1.45) 1.20 (1.00 – 1.45) 1.20 (1.00 – 1.45) 5 – 9.9 km 400 286 (71.5%) 2.16 (1.70 – 2.74) 1.28 (1.00 – 1.67) 1.28 (1.00 – 1.67) 10 km or more 186 154 (83.4%) 4.41 (2.94 – 6.61) 2.70 (1.76 – 4.12) 2.70 (1.76 – 4.12) Missing data 76 21 (26.6%) - Mean 3.2km sd 3.7 - - - Cost of circumcision*** Free 539 269 (49.9%) 1.00 1.00 1.00 Less than 10 Ghs 145 114 (78.6%) 3.69 (2.40 – 5.68) 2.43 (1.54 – 3.84) 2.41 (1.53 – 3.79) Between 10-20 Ghs 1530 888 (58.0%) 1.39 (1.14 – 1.69) 1.22 (0.91 – 1.38) 1.14 (0.92 – 1.41) 20 Ghs or more 554 337 (60.8%) 1.56 (1.23 – 1.98) 1.32 (1.03 – 1.69) 1.26 (0.97 – 1.63) Missing data 79 62 (78.5%) - - - Maternal occupation Gov’t/private 105 49 (46.7%) 0.63 (0.41 – 0.96) 1.19 (0.75 – 1.88) 1.41 (0.87 – 2.27) employed Self-employed 927 514 (55.4%) 0.89 (0.71 – 1.11) 1.19 (0.92 – 1.54) 1.21 (0.93 – 1.57) Farming 589 380 (64.6%) 1.31 (1.02 – 1.68) 0.91 (0.67 – 1.23) 0.87 (0.63 – 1.19) Not working 484 282 (58.3%) 1.00 1.00 1.00 Missing data 742 445 (59.9%) - Maternal educational level None 512 359 (70.1%) 2.31 (1.83 – 2.92) 1.36 (1.05 – 1.77) 1.30 (1.00 – 1.70) Primary 1481 882 (59.6%) 1.45 (1.23 – 1.72) 1.24 (1.04 – 1.48) 1.21 (1.01 – 1.45) Secondary 850 427 (50.4%) 1.00 1.00 1.00 Missing data 4 2 (28.6%) - - - Maternal religion 79
Christian 2048 1081 (52.9%) 1.00 1.00 1.00 Muslim 666 496 (74.5%) 2.60 (2.14 – 3.16) 2.26 (1.84 – 2.78) 2.40 (1.93 – 2.98) Traditional 129 91 (70.5%) 2.13 (1.45 - 3.15) 1.36 (0.90 – 2.04) 1.33 (0.88 – 1.99) African/None Missing data 4 2 (0.0%) - - - Maternal age (years) Less than 20 319 192 (60.4%) 1.00 1.00 1.00 20-29 1458 854 (58.6%) 0.93 (0.72 – 1.19) 1.10 (0.85 – 1.42) 1.05 (0.80 – 1.37) 30 or more 1066 622 (58.7%) 0.92 (0.72 – 1.19) 0.98 (0.74 – 1.28) 0.94 (0.71 – 1.23) Missing data 4 2 (28.6%) - - - Site of delivery Health facility 2291 1241 (54.2%) 1.00 1.00 1.00 Home 549 425 (77.7%) 2.95 (2.37 – 3.66) 2.04 (1.62 – 2.59) 1.89 (1.49 – 2.41) Missing data 6 4 (40.0%) - - - Birth weight Less than 2.5kg 214 136 (63.6%) 1.25 (0.94 – 1.67) 1.24 (0.91 – 1.69) 1.24 (0.91 – 1.70) 2.5kg or greater 2633 1534 (58.2%) 1.00 1.00 1.00 Missing data 0 0 (0.0%) - - - Age at circumcision 0-6 days 117 68 (58.1%) 1.00 1.00 1.00 7-20 days 2556 1483 (58.3%) 0.99 (0.68 – 1.45) 1.09 (0.74 – 1.62) 1.08 (0.71 – 1.64) > 20 days 172 119 (69.2%) 1.62 (0.99 – 2.64) 1.31 (0.78 – 2.21) 1.29 (0.75 – 2.22) Missing data 2 0 (0.0%) - - - *Model 1. Adjusted for income status, cost of circumcision, religion, maternal education and maternal age ** Model 2. Further adjusted for distance to health facility *** 1 Ghs = 0.6 $US (2012) Ref: Reference group, CI: Confidence interval; sd: Standard deviation
80
In contrast, 59.5% of circumcisions were provided free by doctors, 16.2% by nurses, 41.1% by medical assistants, 9.1% by drug sellers, and 40.3% by domestic helpers (Table 5.3). Families in the lowest income quintile also appeared to be the least likely to receive free circumcision services (Table 5.4).
Only 9.0% of families in the lowest quintile received free circumcision services compared to 27.9% in the highest quintile (aOR 0.40, 95% CI 0.28-0.58 p=<0.001).
There also appeared to be a ‘dose response’ where the likelihood of receiving a free circumcision decreased as income status decreased (aOR 0.35, 95% CI 0.23-
0.53 p=<0.001). 58.7% of families in the lowest quintile paid between 10-20
Ghana Cedis for their circumcision and 20.2% paid 20-100 Ghana Cedis.
Eighty seven (3.1%) families paid in-kind contributions in the form of bars of soap, chickens, kola nuts, and corn (Table 5.3). The payment of in-kind contributions was more common with Wanzams (7.4%) than doctors (0.7%), nurses (0.2%), medical assistants (0.0%), drug sellers (3.4%), and domestic helpers (1.5%) (Table 5.3). Families of low socio-economic status appeared to be more likely to pay additional in-kind contributions (31.0%) compared to highest income families (11.5 %) (aOR 0.41, 95% CI 0.25-0.67 p=<0.001).
Infants who lived 10 km or more from a health facility (154, 83.4%) were two times more likely to receive their circumcision from an informal provider compared to infants who lived less than 1 km from a health facility (776, 53.8%)
81
Table 5.3 Details of cash payments and in-kind payments by provider type
Provider Type
All providers Doctor Medical assistant Nurse Drug seller Domestic helper Wanzam Other**
Included Included Included Included Included Included Included Included Total in-kind Total in-kind Total in-kind Total in-kind Total in-kind Total in-kind Total in-kind Total in-kind Cost of payment payment payment payment payment payment payment payment circumcision* n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) 87 136 1 83 0 958 2 175 6 454 7 979 72 62 0 Total 2847 (3.1%) (4.8%) (0.7%) (2.9%) (0.0%) (33.6%) (0.2%) (6.1%) (3.4%) (15.9%) (1.5%) 34.4%) (7.4%) (2.2%) (0.0%) 539 16 81 1 34 0 155 1 16 2 183 3 68 10 2 0 Free (18.9%) (3.0%) (59.5%) (1.2%) (41.1%) (0.0%) (16.2%) (0.6%) (9.1%) (12.5) (40.3%) (1.6%) (6.9%) (14.0%) (3.2%) (0.0%) 1.00-9.9 Ghs 145 4 4 0 3 0 24 0 11 0 13 0 84 4 1 0 (5.1%) (2.9%) (2.9%) (0.0%) (3.6%) (0.0%) (2.5%) (0.0%) (6.2%) (0.0%) (2.9%) (0.0%) (8.6%) (4.6%) (1.6%) (0.0%) Between 10-20 1530 42 17 0 29 0 596 0 122 2 154 2 579 38 33 0 Ghs (53.7%) (2.8%) (12.5%) (0.0%) (34.9%) (0.0%) (62.2%) (0.0%) (69.0%) (1.6%) (33.9%) (1.3%) (59.0%) (6.6%) (53.2%) (0.0%) 20.00-100.00 554 23 31 0 16 0 170 1 23 2 100 2 188 18 26 0 Ghs (19.5%) (4.2%) (22.8%) (0.0%) (19.3%) (0.0%) (17.7%) (0.6%) (13.1%) (8.7%) (22.0%) (2.0%) (19.2%) (9.6%) (41.9%) (0.0%) 79 2 3 0 1 0 13 0 3 0 4 0 55 2 0 0 Missing data (2.8%) (2.5%) (2.2%) (0.0%) (1.2%) (0.0%) (1.4%) (0.0%) (1.7%) (0.0%) (0.9%) (0.0%) (5.6%) (3.6%) (0.0%) (0.0%)
*1 Ghs = 0.6 $US (2012) ** Other: These included family members and friends
82
Table 5.4 Cost of circumcision by household income status Cost of circumcision** Household Total Free Less than 10 Between 10-20 20.00 – 100 Missing income status* Ghs Ghs Ghs data
n=2847 n=539 n=145 n=1530 n=554 n=79 n (%) n (%) n (%) n (%) n (%) 1 (Lowest) 387 34 (9.0%) 41 (10.6%) 227 (58.7%) 78 (20.2%) 7 (1.8%) 2 532 55 (10.3%) 27 (5.0%) 319 (60.0%) 118 (22.2%) 13 (2.4%) 3 628 114 (18.2%) 25 (4.0%) 367 (58.4%) 104 (16.6%) 18 (2.9%) 4 687 165 (24.0%) 34 (4.9%) 336 (48.9%) 132 (19.2%) 20 (2.9%) 5 (Highest) 613 171 (27.9%) 18 (2.9%) 281 (45.8%) 122 (19.9) 21 (3.4%) * Wealth quintile calculated using principal components analysis ** 1 Ghs = 0.6 $US (2012)
(aOR 2.70, 95% CI 1.76-4.12 p=<0.001) (Table 5.2). There also appeared to be a dose response with increasing risk of receiving a circumcision from an informal provider as distance to a health facility increased (Table 5.2) (aOR 1.25, 95 CI
1.30-1.38 P=<0.001).
Household income status was closely associated with distance to a health facility (Table 5.5). Families in the lowest socio-economic quintile lived an average of 6.1 km (standard deviation [sd] 4.4km) from a health facility (median
6.9 km , interquartile range [IQR] 1–15.9 km) compared to an average of 1.1 km in families in the highest socio-economic quintile (sd 1.6km, median 0.6 km,
IQR 0–10.4 km) (Table 5.5). Families in the lowest socio-economic quintile (79,
42.5%) were 22 times more likely to live more than 10km from a health facility compared to families in the highest socio-economic quintile (5, 2.7%) (aOR 22.35
95% CI 8.84-56.54 p=<0.001) (Table 5.5). However, both socio-economic status
(aOR 1.32, 95 CI 1.23-1.41 P=<0.001) and distance to health facilities (aOR 1.28,
83
95 CI 1.14-1.43 P=<0.001) had independent effects on the choice of circumcision provider.
There was no statistical evidence of modification of the effect of distance from health facility on the choice of provider for circumcision by income status of the household (p-value for the interaction, 0.188). Infants were two times more likely to receive circumcision from an informal provider if the families were
Muslim (496, 74.5%) compared to Christian (1081, 52.9%) (aOR 2.40, 95% CI
1.93-2.98 p=<0.001) (Table 5.2). Mothers with no formal education (359, 70.1%) were 30% more likely to receive an informal circumcision provider compared to mothers with secondary level education (427, 50.4%) (aOR 1.30, 95% CI 1.00-
1.70 p=<0.049) even after adjusting for other variables. There were no obvious differences associated with other socio demographic characteristics (Table 5.2).
84
Table 5.5 Distance to health facility by household income status Household Total Mean distance Median distance Min. & max. < I Km 1 – 4.9 Km 5 – 9.9 Km 10 Km or more Not known/ income status (sd) (interquartile range) values missing n (%) km km km n (%) n (%) n (%) n (%) n (%)
Total 2847 3.1 (3.6) 1.1 (0 – 12.8) 0 & 18.9 1444 741 400 186 76
1 (Lowest) 387 (13.6%) 6.1 (4.4) 6.9 (1 – 15.9) 0 & 18.5 106 (7.3%) 49 (6.1%) 153 (38.3%) 79 (42.5%) 0 (0.0%)
2 532 (18.7%) 3.9 (3.8) 2.1 (0 – 12.3) 0 & 13.1 218 (15.1%) 117 (15.8%) 132 (33.0%) 55 (29.6%) 10 (13.2%)
3 628 (22.1%) 2.4 (3.0) 0.9 (0 – 11.3) 0 & 12.3 321 (22.2%) 203 (27.4%) 62 (15.5%) 35 (18.8%) 7 (9.2%)
4 687 (24.1%) 1.7 (2.2) 0.8 (0 – 10.9) 0 & 12.7 387 (26.8%) 228 (30.8%) 42 (10.5%) 12 (6.5%) 18 (23.7%)
5 (Highest) 613 (21.5%) 1.1 (1.6) 0.6 (0 – 10.4) 0 & 11.4 412 (28.5%) 144 (19.4%) 11 (2.8%) 5 (2.7%) 41 (53.9%)
85 5.5 Discussion
In this population based study in rural Ghana, infant male circumcision was almost universal (91%) and was performed by both formal (41%) and informal
(59%) circumcision providers. Both socio-economic status and geographic access to health facilities had important and independent effects on the choice of circumcision provider. The risk of receiving a circumcision from an informal provider increased with each level of deprivation and with the distance that families lived from health facilities. The study also found that families with the lowest household income were the most likely to pay for their circumcision.
Poor families were also most likely to pay additional in-kind contributions.
The relationship between socio-economic status2,52,73,100, geographic access2,48,101, and choice of informal provider for infant male circumcision has been reported in many studies in low and middle income countries. However, this study was the first to report data from a rural area in Africa with high population level coverage of infant male circumcision. This is also the first study to report the double burden that circumcision places on families of low socio-economic status. In this study poor families were more likely to receive a circumcision from an untrained informal provider and also more likely to incur a significant economic cost.
86 In 2008, the “Free Maternal Care Policy”102 was introduced into the Ghana
Health Insurance Scheme103. Under the policy, all pregnant women and their infants up to 90 days postpartum and all children aged 90 days to 18 years are meant to receive free care in accredited public and private healthcare facilities.
The services that are covered include antenatal care, delivery care, postnatal care, and infant male circumcision. Mothers and children just have to be registered and receive a registration card. The registration process is free and there are meant to be no out of pocket expenses. However, there have been difficulties in enrolling many families into the scheme. This has been attributed to difficulties in accessing many areas of Ghana, especially the poorest and most disadvantaged areas104,105. In 2011, close to the time of conducting this study, only 33% of Ghana’s population were registered with 4.2% coverage for the poorest105. The most recent data from 2013 indicate that the national coverage still remains limited with only 38% registered106. Inequity in health insurance coverage is likely to be an important driver of the costs of circumcision incurred by poor families that was reported. The study area was located in central rural
Ghana in the Brong Ahafo region and health insurance coverage in the Brong
Ahafo region was 45.9% in 2011105. However, there were no data on coverage of health insurance in the poorest families in the study area. Antenatal care and delivery services were also meant to be free under the Ghana health insurance scheme102,103 and similar inequities are also reported for these services. There were reports of poor women being charged unofficial and non-legitimate fees
87 for delivery and postnatal care services107,108. Reports of poor women and their babies being forcibly kept in birthing hospitals until their bills are settled have also been published107. Poor women have also been charged unofficial fees for antenatal109, delivery, and postnatal care services108 in accredited facilities because they were unable to confront authority figures110,111. Poor women were also less likely to be insured for delivery care compared to richer women in
Ghana112,113.
Additional economic costs of circumcision include the payment of ‘in-kind’ contributions. The payment of in-kind contributions was more common with
Wanzams (7%) than formal providers (3%) (doctors, nurses, and medical assistants) in this study. The poorest families also paid more in-kind contributions (31%) than the highest income families (12%). Two rural Kenyan studies have reported the payment of in-kind contributions (chickens, sheep, food and medical supplies) by families for circumcision48,73. In these studies medical practitioners (49%) and informal traditional providers (51%) received similar in-kind contributions. However, these studies did not provide any information on the in-kind contributions paid by poor and richer families within the same study area.
The study also reported that families of the Muslim religion were two-fold more likely to choose an informal provider than families with other religious
88 affiliations. The Muslim religion is a well-known determinant of use of informal providers for circumcision in urban and rural Africa2,52,114 and many Wanzams are Muslim themselves21. Approximately, 70% of Wanzams who performed circumcisions in our rural study area were Muslims. The study also reported that mothers with no formal education were more likely to choose an informal circumcision provider compared to mothers with secondary level education.
These data were also consistent with other African studies65. There were no obvious differences in choice of circumcision provider associated with other socio-demographic characteristics in this study.
The study had some limitations. Investigators from Egypt have reported a lack of confidence in the formal health care system as a reason for the use of informal circumcision providers who charge fees2,66. These studies also reported that traditional providers were perceived as more experienced and better in providing healthcare than formal health service providers66. However, this study was not able to conduct in depth qualitative interviews to explore perspectives and experiences of families and health service providers in our study.
The study was also unable to assess family’s perceptions of quality of care. The study was also unable to collect data on transport costs and other opportunity costs incurred by the families. This study was observational and cross-sectional
89 and does not provide proof of causation. However, the study controlled for a wide range of individual, household and community level confounders and strengths of this study included its large community and population based data collection system. In addition 22% of babies were not able to be visited within a
12 week period after birth. Anecdotal information from the study area indicated that these families needed to travel more for employment and they were of lower socio economic status and educational levels. The omission of these infants reduces the generalisability of this study a little but is unlikely to have introduced any systematic bias.
5.6 Conclusions
This study appears to be the first to analyse the “on the ground” “community level” influence of socioeconomic factors on choice of infant male circumcision provider in an area with almost total population coverage. It also appears to be the first study that has described the high and inequitable costs paid by the poorest families in rural Africa for infant male circumcision. The Government of Ghana and other Non-Government Organisations should provide additional support to poor families so they can access high quality free infant male circumcision in rural Ghana. This includes improved coverage of Ghana’s free maternal care policy and health insurance scheme for the poorest families.
90 Chapter 6. Health care seeking behaviour of families
The third PhD paper presented in this chapter is the findings of the research conducted to address the thesis objective three; to investigate and compare the health care seeking behaviour of families with infants who had circumcision- related morbidity and families with infants who experienced acute illness in the study area. This paper was published in BMJ Open in 2017 (Appendix F)115.
6.1 Abstract
Objective: This study assessed health care seeking patterns of families of infants with circumcision related morbidities and families of infants with acute illnesses in rural Ghana.
Design: Two population based cohort studies.
Setting: Brong Ahafo Region of central rural Ghana.
Participants: A total of 22,955 infants enrolled in a large population based trial
(Neovita trial) from 16th August 2010 to 7th November 2011 and 3141 infants in a circumcision study from 21st May 2012 to 31st December 2012.
Primary Outcome: Care seeking for circumcision related morbidities and acute illnesses unrelated to circumcision.
Results: Two hundred and thirty (8.1%) infants from the circumcision study had circumcision related morbidities and 6265 (27.3%) infants from the Neovita study had acute illnesses unrelated to circumcision. A much lower proportion
(35, 15.2%) of families of infants with circumcision related morbidities sought healthcare compared to families of infants with acute illnesses in the Neovita
91 study (5520, 88.1%). More families sought care from formal providers (24, 69%) compared to informal providers (11, 31%) for circumcision related morbidities.
There were no obvious determinants of care seeking for acute illnesses or circumcision related morbidities in the population.
Conclusions: Government and non-government organisations need to improve awareness about the complications and care seeking needed for circumcision related morbidities.
6.2 Introduction
Infant male circumcision is a common surgical procedure and is performed for cultural, religious and medical reasons. Approximately 90% of males aged under 12 weeks are circumcised in Ghana5 and much of West Africa2. A concerning high risk of circumcision related morbidity (8.1%) following infant male circumcision was reported in a recent community-based study in rural
Ghana5. Overall, 53% of the complications were observed following circumcision performed by an informal provider (including village-based traditional circumcision providers) and 47% after circumcision performed by formal healthcare providers including doctors, nurses, and medical assistants5.
Appropriate health care seeking for infants with circumcision related morbidities could minimise the occurrence of severe long and short-term morbidities. Studies conducted in rural Ghana have reported that poor women
92 believe infant illnesses related to traditional practices are not meant to be managed in the hospital or clinic37-39. However, no other data on care seeking for infant male circumcision related morbidity in rural African areas are available. There are more studies that have examined the determinants of care seeking for acute infectious diseases and newborn complications in rural
Africa40,41. A recent study conducted among mothers and infants in Ghana identified poor health care seeking behaviour in families of infants with acute illness during the postpartum period. In this study, 29% and 54% of sick infants received health care within two and fourteen days, respectively42. Another
Ghanaian study identified poverty and personal preferences as key determinants of poor care seeking behaviour by families of infants with illness43. A study from Burkina Faso also linked traditional concepts of illness in
Africa to delays in seeking professional medical care by families of infants with malaria44. Distance to health facility, health provider attitude and inadequate supplies have also been reported as key determinants of delayed care seeking for malaria cases in infants in sub-Saharan African populations45. However, it appears that there have been no studies which have compared health seeking behaviour between families of infants with circumcision related morbidities and families of infants with acute illnesses that are not circumcision related.
The overall aim of this study was to assess healthcare seeking patterns of families of infants with circumcision related morbidities and families of infants
93 with other acute illnesses in rural Ghana. The primary objective was to investigate whether there were differences in the healthcare seeking patterns of families of infants with circumcision related problems, including infants who died, and families of infants with acute illnesses. The secondary objective was to assess the factors that influence care seeking patterns of families in the study area.
6.3 Methods
6.3.1 Study design and setting
Data from two separate cohort studies in the same part of the Brong Ahafo
Region of central Ghana were analysed separately and compared descriptively.
The larger of the two cohorts includes data from a population based trial of newborn vitamin A supplementation (Neovita trial), conducted between 16th
August 2010 to 7th November 2011 and involving 22,955 mother infant pairs; full details have been published previously86. The smaller cohort was conducted between 21st May 2012 and December 31st 2012 and focused on circumcision.
This was implemented to obtain additional observational data on births, cause specific mortality and circumcision related morbidity in the study area. Over the period from 2010 to 2012, 80% of the study population lived in rural settlements and 20% of mothers had no primary school education. Four major district hospitals and over 80 small health facilities provided health care services to the population. There were 30 doctors and 44 other formal health
94 service providers (medical assistants and nurses) providing curative services in the four district hospitals at the time of the circumcision study. There were also approximately 120 informal care providers such as drug sellers, traditional healers and religious leaders in the study area.
6.3.2 Data collection
For the Neovita study, from August 2010 to November 2011, all births in the study area were reported to the trial team via a network of fieldworkers.
Fieldworkers visited all families at home between two hours and two days after birth and interviewed the mother of the infant or the primary care giver.
Fieldworkers weighed the baby and asked the mother or the primary care giver about: date of birth, site of birth, current address, distance to health facilities, socio-demographic characteristics and socio-economic information (using an asset index). At the monthly surveillance visits, the families were asked by fieldworkers for detailed information on; infant illness including start and end dates; and healthcare seeking during infant illness. Due to time constraints families were not asked exactly who they sought care from. The fieldworkers also asked the families who sought healthcare whether the infant was admitted to hospital. The fieldworkers next reviewed infant health records and checked for consistency with family’s report. The fieldworkers also collected data on the vital status of the baby (including if the baby was alive, dead or hospitalised).
95
The circumcision study included all male liveborn infants who were born in the study population from May to December 2012 and were aged under 12 weeks.
Follow-up visits were scheduled between eight to eleven weeks post birth and trained senior fieldworkers asked for consent to collect additional detailed data on: age at circumcision, site of circumcision and type of circumcision provider.
The fieldworkers asked about circumcision morbidities and whether these morbidities had been confirmed by a trained health professional. Families were also asked if they sought care for circumcision related morbidities and if they had been given advice about potential problems or care seeking. The fieldworkers also asked detailed information on the following; site of healthcare; type of healthcare provider; medicines prescribed; admissions; and surgery. Consent to access the baby’s Neovita data including socioeconomic, demographic and care seeking was obtained.
In both studies the fieldworkers were trained for two weeks in all study procedures prior to the commencement of the study. Interrater reliability was checked between all fieldworkers. We did not perform statistical evaluation of agreements between the fieldworkers and the study coordinator (TG). However, each fieldworker was supervised during 2 visits per month as part of the study scheduled and unscheduled supervisory visits. During these visits, TG and field supervisors observed the fieldworker interviewing mothers, and examined infants and recorded findings independently. Recorded data were compared
96 between the fieldworker and TG/field supervisor and used to provide feedback after the home visits and at weekly fieldworker meetings. The fieldworkers used standardised paper based data collection tools (including a standardised list of closed ended questions) for all interviews.
6.3.3 Study definitions
The study defined a formal health care service as medical care provided at hospitals, clinics or health posts by professionally trained, licensed and regulated providers of medical services including; doctors, medical assistants, nurses or pharmacists. An informal health care service was defined as care provided by untrained, unlicensed, unregulated private provider of care services including: drug sellers, traditional healers and religious leaders.
Circumcision related morbidities were defined as complications occurring during or after the circumcision procedure as reported by the primary caregiver including; excess skin removal or incision, excess bleeding, inadequate skin removal, infection, abnormal stream of urine, glans amputation, ulcer5. Acute illnesses were defined as illnesses or injuries other than circumcision related complications such as malaria, sepsis, acute respiratory tract infection or diarrhoea116,117. A formal circumcision provider referred to professionally trained, licensed and regulated providers of medical services. This included: doctors, medical assistants, or nurses. An informal circumcision provider was an untrained, unlicensed, unregulated private provider of medical services including: drug
97 sellers, domestic staff, family members and Wanzams (village based traditional circumcision providers).
6.3.4 Outcome
The primary outcome was care seeking. In the circumcision study, care seeking was defined as the percentage of families of infants who sought care from either a formal healthcare provider or an informal care provider among families who reported of an infant having a circumcision related morbidity after circumcision during the circumcision follow-up visits (occurred between eight and eleven weeks post birth). In the Neovita study, care seeking was defined as the percentage of families of infants with acute illnesses unrelated to circumcision who sought care outside the home of the infant as reported by care givers during the Neovita infant follow-up visits (from four weeks up to fifty two weeks of age).
6.3.5 Statistical analysis
Two hundred and thirty (8.1%) infants had circumcision related morbidities and of these, only 35 (15.2%) families sought care. Thus, the study reported care seeking patterns in the circumcised cohort only using simple proportions and descriptive analyses. Statistical analyses were used to assess care seeking patterns in the larger Neovita cohort of 22,955 infants. Multivariable logistic regression models were constructed apriori to assess the association between care seeking patterns and morbidity risk and adjusted for the effect of important explanatory variables including: religion, maternal education level, maternal
98 age, distance to health facility, site of delivery, sex, birth weight and income status. Adjusted odds ratios (ORs) and 95% confidence intervals (95% CI) were calculated. All analyses were conducted in STATA version 13.
6.3.6 Ethical issues
Ethical approvals were obtained from Ghana Health Service (GHS) Ethical
Review Committee, the Institutional Ethics Committee of Kintampo Health
Research Centre (KHRC), the Research Ethics Committee of London School of
Hygiene and Tropical Medicine, and the Human Research Ethics Committee of the University of Western Australia (UWA). Written informed consent was obtained from all the families of the circumcised male infants.
6.4 Results
From 16th August 2010 to 7th November 2011, 22,955 infants comprising 11,649
(50.8%) males and 11306 (49.3%) females were recruited into the Neovita trial
(Figure 6.1). A total of 4510 (19.7%) of the mothers were in the lowest socio- economic quintile, 17581 (76.6%) delivered in a health facility and 7127 (31.1%) had no primary education (Table 6.1). From 21st May to 31st December 2012,
3141 male infants were recruited in the circumcision study (Figure 6.2).
Overall, 2850 (90.7%) of these infants were circumcised. The study excluded three babies (0.1%) who had no socio-economic or demographic data. A total of
503 (16.0%) of mothers were in the lowest income quintile; 2479 (78.9%) delivered in a facility and 601 (19.1%) had no primary education (Table 6.2).
99
26414 live births in the study population from August 2010 to November 2011
3459 excluded (13.1%). 2500 would not stay for 6 months, 800 too old, 84 unable to feed, 75 other reason
22955 (86.9%) infants enrolled in the Neovita trial
11306 (49.3%) females 11649 (50.7%) males
2969 (26.3%) with acute 8337 (73.7%) with no 3296 (28.3%) with acute 8353 (71.7%) with no illness reported illness illness reported illness
2589 (87.2%) sought 380 (12.8%) did not seek 365 (11.1%) did not seek 2931(88.9%) sought care care care care
274 (10.6%) admitted 334 (10.1%) admitted
Figure 6.1 Flow diagram for live births from August 2010 to November 2011 included in the Neovita study
100
8110 live births in the study population from May to December 2012
4105 (51%) female
4005 (49%) male live births 863 (21.5%) not seen within 12 weeks post birth 1 (0.0%) moved out of the study area 3141 (78.4%) males aged under 12 weeks enrolled in the circumcision study
291 (9.3%) not circumcised
2850 (90.7%) circumcised by age 12 weeks
3 infants without socio economic or morbidity data due to field worker error
2847 (99.9%) included in analysis
230 (8.1%) with confirmed circumcision associated morbidities including 2 deaths
35 (15.2%) sought care for morbidity
3 (8.6%) admitted in health facility
Figure 6.2 Flow diagram for live births from May 2012 to December 2012 included in the circumcision study
6.4.1 Acute illness and care seeking
A total of 6265 (27.3%) Neovita trial infants had an acute illness during the first year of life. There was little difference in the prevalence of acute illness by sex in
101
the Neovita population; 3296 (28.3%) males and 2969 (26.3%) females (Figure
6.1 and Table 6.1).
Table 6.1 Characteristics of families and infants who sought care for acute illness in the Neovita trial Neovita study Neovita trial infants Neovita trial infants with population with acute illness acute illness whose families sought care
Characteristics n=22955 n=6265 (27.3%) n=5520 (88.1%) Religion Christian 15979 4314 (27.0%) 3820 (88.5%) Muslim 5511 1566 (28.4%) 1357 (86.7%) Traditional 385 (26.3%) 1465 343 (89.1%) Africa/none Missing data - - - Mother’s highest educational level None 7127 2009 (28.2%) 1752 (86.2%) Primary 4236 1241 (29.3%) 1088 (87.7%) Secondary 11578 3013 (26.0%) 2678 (88.9%) Carer didn’t know 14 2 (14.3%) 2 Income status of household 1 (Lowest) 4510 1316 (29.2%) 1147 (87.2%) 2 4550 1297 (28.5%) 1147 (88.4%) 3 4583 1274 (27.8%) 1130 (88.7%) 4 4644 1212 (26.1%) 1063 (87.7%) 5 (Highest) 4668 1166 (25.0%) 1033 (88.6%) Missing data - - - Maternal age (years) Less than 20 2646 760 (28.7%) 658 (86.6%) 20-29 12041 3269 (27.1%) 2900 (88.7%) 30 or more 8239 2228 (27.0%) 1954 (87.7%) Carer didn’t know 30 8 (26.7%) 8 Site of delivery Health facility 17 581 4775 (27.2%) 4213 (88.2%) Home 5374 1490 (27.7%) 1307 (87.7%) Missing data - - Distance to health facility <1 km 13880 3657 (26.3%) 3234 (88.4%) 1 – 4.9 km 5285 1497 (28.3%) 1324 (88.4%) 5 – 10 km 2950 752 (25.5%) 646 (85.9%) 10 km or more 1197 359 (30.0%) 316 (88.0%) Missing data - - - Birth weight Less than 2.5kg 3592 998 (27.8%) 867 (86.9%) 2.5kg or greater 19361 5267 (27.2%) 4653 (88.3%) Missing data - - - Sex Male 11649 3296 (28.3%) 2931 (88.9%) Female 11306 2969 (26.3%) 2589 (87.2%)
Healthcare was sought for 5520 (88.1%) Neovita trial infants with acute illness
and 608 (11.0%) were admitted to hospital (10.6% females and 10.1% males)
(Figure 6.1 and Table 6.1). A total of 1033 (88.6%) families who sought care for
acute illness were in the highest socio-economic quintile compared to 1147 102
(87.2%) in the lowest quintile, 4213 (88.2%) delivered in a health facility compared to 1307 (87.7%) who delivered at home and 2678 (88.9%) had more than primary school education compared to 1752 (86.2%) with no formal education (Table 6.1).
A slightly lower proportion of families of female infants with acute illness
(87.2%) sought care compared to families of male infants (88.9%) (aOR 0.85, 95%
CI 0.73, 0.99 p=0.039) (Table 6.3). Surprisingly, there were no obvious differences in the prevalence of health care seeking related to levels of education (aOR 0.89, 95% CI 0.72, 1.10 p=0.260), socio economic status (aOR
1.00, 95% CI 0.75, 1.33 p=0.848), maternal age (aOR 1.15, 95% CI 0.89, 1.48 p=0.131), site of delivery (aOR 1.01, 95% CI 0.83, 1.24 p=0.886), distance to health facility (aOR 0.96, 95% CI 0.68, 1.37 p=0.369) or birth weight (aOR 0.89, 95% CI
0.72, 1.09 p=0.260) in the Neovita trial population (Table 6.3).
6.4.2 Circumcision complications and care seeking
Of the 2847 circumcised male infants included in the circumcision analysis, 230
(8.1%) had confirmed morbidities related to circumcision and two died.
103
Table 6.2 Characteristics of families and infants who sought care for morbidity in the circumcision study Circumcision Circumcised Circumcised infants Circumcised infants study infants with morbidity with morbidity whose population families sought care Characteristics n=3141 n=2850* (90.7%) n=230 (8.1%) n=35 (15.2%) Religion Christian 2234 2048 (91.7%) 172 (8.4%) 24 (14.0%) Muslim 727 666 (91.6%) 49 (7.4%) 9 (18.4%) Traditional 171 129 (75.4%) 9 (7.0%) 2 (22.2%) Africa/none Missing data 9 7 (77.8%) - - Mother’s highest educational level None 601 512 (85.2%) 50 (9.8%) 6 (12.0%) Primary 1619 1481 (91.5%) 127 (8.6%) 21 (16.5%) Secondary 913 850 (93.1%) 53 (6.2%) 8 (15.1%) Carer didn’t know 8 7 (87.5%) - - Income status of household 1 (Lowest) 503 387 (76.9%) 28 (7.2%) 6 (21.4%) 2 608 532 (87.5%) 41 (7.7%) 4 (9.8%) 3 676 628 (92.9%) 47 (7.5%) 9 (19.1%) 4 725 687 (94.8%) 62 (9.0%) 8 (12.9%) 5 (Highest) 629 613 (97.5%) 52 (8.5%) 8 (15.4%) Missing data 0 3 (0.1%) - - Maternal age (years) Less than 20 360 319 (88.6%) 26 (8.2%) 3 (11.5%) 20-29 1603 1458 (91.0%) 130 (8.9%) 18 (13.8%) 30 or more 1170 1066 (91.1%) 74 (6.9%) 14 (18.9%) Carer didn’t know 8 7 (87.5%) - - Site of delivery Health facility 2479 2292 (92.5%) 196 (8.6%) 28 (14.3%) Home 650 549 (84.5%) 34 (6.2%) 7 (20.6%) Missing data 12 6 (50.9%) - - Distance to health facility <1 km 1545 1444 (93.5%) 110 (7.6%) 21 (19.1%) 1 – 4.9 km 801 741 (92.5%) 77 (10.4%) 7 (9.1%) 5 – 10 km 484 400 (82.6%) 27 (6.8%) 3 (11.1%) 10 km or more 231 186 (80.5%) 14 (7.5%) 3 (21.4%) Missing data 80 76 (95.0%) 2 (2.6%) 1 (50.1%) Birth weight - Less than 2.5kg 255 214 (83.9%) 21 (9.1%) 1 (4.8%) 2.5kg or greater 2886 2633 (91.2%) 209 (7.9%) 34 (16.3%) Missing data 0 - - - *Three circumcised infants had no socioeconomic and demographic data due to field worker error
104
Table 6.3 Predictors of care seeking behaviour in the Neovita trial population Neovita study Neovita trial Neovita trial infants Unadjusted p-value Adjusted Odds p-value population infants with acute whose families Odds ratio (OR) ratio (aOR)* illness sought care Characteristics n=22955 n=6265 (27.3%) n=5520 (88.1%) Religion Christian 15979 4314 (27.0%) 3820 (88.5%) 1.00 0.121 1.00 0.195 Muslim 5511 1566 (28.4%) 1357 (86.7%) 0.84 (0.70, 1.00) 0.86 (0.72, 1.03) Traditional 385 (26.3%) 1465 343 (89.1%) Africa/none 1.06 (0.76, 1.47) 1.10 (0.79, 1.55) Missing data - - Mother’s highest educational level None 7127 2009 (28.2%) 1752 (86.2%) 0.89 (0.73, 1.09) 0.89 (0.72, 1.10) Primary 4236 1241 (29.3%) 1088 (87.7%) 0.85 ( 0.72, 1.01) 0.84 (0.69, 1.03) Secondary 11578 3013 (26.0%) 2678 (88.9%) 1.00 0.174 1.00 0.260 Carer didn’t know 14 2 (14.3%) 2 - Income status of household 1 (Lowest) 4510 1316 (29.2%) 1147 (86.6%) 0.87 (0.69, 1.11) 1.00 (0.75, 1.33) 2 4550 1297 (28.5%) 1147 (86.4%) 0.98 (0.77, 1.26) 1.07 (0.82, 1.39) 3 4583 1274 (27.8%) 1130 (88.7%) 1.01 (0.79, 1.30) 1.08 (0.83, 1.39) 4 4644 1212 (26.1%) 1063 (87.7%) 0.92 (0.72, 1.18) 0.95 (0.74, 1.22) 5 (Highest) 4668 1166 (25.0%) 1033 (88.6%) 1.00 0.714 1.00 0.848 Missing data - - - Maternal age (years) Less than 20 2646 760 (28.7%) 658 (86.6%) 1.00 0.208 1.00 0.131 20-29 12041 3269 (27.1%) 2900 (88.7%) 1.22 (0.96, 1.54) 1.23 (0.97, 1.57) 30 or more 8239 2228 (27.0%) 1954 (87.7%) 1.11 (0.87, 1.41) 1.15 (0.89, 1.48) Carer didn’t know 30 8 (26.7%) 8 - Site of delivery Health facility 17 581 4775 (27.2%) 4213 (88.2%) 1.00 0.595 1.00 0.886 Home 5374 1490 (27.7%) 1307 (87.7%) 0.95 (0.80, 1.14) 1.01 (0.83, 1.24) Missing data - Distance to health facility <1 km 13880 3657 (26.3%) 3234 (88.4%) 1.00 0.278 1.00 0.369 1 – 4.9 km 5285 1497 (28.3%) 1324 (88.4%) 1.00 (0.83, 1.21) 0.99 (1.82, 1.20) 5 – 10 km 2950 752 (25.5%) 646 (85.9%) 0.80 (0.63, 1.00) 0.80 (0.62, 1.03) 10 km or more 1197 359 (30.0%) 316 (88.0%) 0.96 (0.69, 1.34) 0.96 (0.68, 1.37) Missing data - - - - - Birth weight Less than 2.5kg 3592 998 (27.8%) 867 (86.9%) 0.87 (0.71, 1.07) 0.89 (0.72, 1.09) 2.5kg or greater 19361 5267 (27.2%) 653 (88.3%) 1.00 0.194 1.00 0.260 Missing data - - - 105
Sex Male 11649 3296 (28.3%) 2931 (88.9%) 1.00 0.035 1.00 0.039 Female 11306 2969 (26.3%) 2589 (87.2%) 0.84 (0.73, 0.99) 0.85 (0.73, 0.99) *Adjusted for religion, maternal education, distance to health facility, site of delivery and sex.
106
Only 35 (15.2%) families of infants with circumcision related morbidities sought care and 3 (8.6%) were admitted to hospital (Figure 6.2 and Table 6.2).
The proportion of mothers aged less than 20 years who sought care for their infants with circumcision related morbidity (3, [11.5%]) was slightly lower than mothers aged 30 years or more (14, [18.9%]). Healthcare seeking of mothers of infants with low birth weight was lower (1, [4.8%]) than mothers of infants with normal birth weight though the total numbers were small (34, [16.3%]) (Table
6.2). There were no other obvious determinants of care seeking for circumcision related morbidity.
6.4.3 Type of provider
In the circumcision study, of the 35 families of infants who sought care for circumcision related morbidity, 20 (57.1%) infants were circumcised by an informal provider and 15 (42.9%) by a formal provider (Table 6.4). Most families of infants circumcised by a formal provider also sought care from a formal provider for complications (13, 87.7%). In contrast, only 45% (9) families of infants circumcised by an informal provider also sought healthcare for circumcision related morbidity from an informal healthcare provider (Table
6.4).
A similar proportion of mothers in the lowest income quintile (4, [66.7%]) sought healthcare for circumcision related morbidity from an informal care
107
Table 6.4 Type of health care provider for circumcision-related morbidity Circumcised infants Type of healthcare provider for circumcision morbidity whose families sought care Formal provider Informal provider* Characteristics n=35 24 (68.6%) 11 (31.4%) Religion Christian 24 17 (70.8%) 7 (29.2%) Muslim 9 6 (66.7%) 3 (33.3%) Traditional 1 (50.0%) 1 (50.0%) 2 Africa/none Missing data - - - Mother’s highest educational level None 6 2 (33.3%) 4 (66.7%) Primary 21 15 (71.4%) 6 (28.6%) Secondary 8 7 (87.5%) 1 (12.5%) Carer didn’t know - - - Income status of household 1 (Lowest) 6 2 (33.3%) 4 (66.7%) 2 4 2 (50.0%) 2 (50.0%) 3 9 7 (77.8%) 2 (22.2%) 4 8 7 (87.5%) 1 (12.5%) 5 (Highest) 8 6 (75.0%) 2 (25.0%) Missing data - - - Maternal age (years) Less than 20 3 3 - 20-29 18 11 (61.1%) 7 (38.9%) 30 or more 14 10 (71.4%) 4 (28.6%) Carer didn’t know - - - Site of delivery Health facility 28 22 (78.6%) 6 (21.4%) Home 7 2 (28.6%) 5 (71.4%) Missing data - - - Distance to health facility <1 km 21 18 (85.7%) 3 (14.3%) 1 – 4.9 km 7 3 (42.9%) 4 (57.1%) 5 – 10 km 3 1 (33.3%) 2 (66.7%) 10 km or more 3 1 (33.3%) 2 (66.7%) Missing data 1 1 - Type of circumcision provider Formal provider 15 13 (86.7%) 2 (13.3%) Informal provider 20 11 (55.0%) 9 (45.0%) Missing data 0 - - Birth weight Less than 2.5kg 1 - 1 (100.0%) 2.5kg or greater 34 24 (70.6%) 10 (29.4%) Missing data - - - * Informal care provider included; drug sellers, traditional healers and religious leaders
108 provider compared to mothers in the highest income quintile (2, [25.0%]), however numbers were low in this analysis (Table 6.4). A greater proportion of mothers who delivered their infants at home (5, [71.4%] sought healthcare for circumcision related morbidity from an informal care provider compared to mothers with health facility births (6, [21.4%]). The proportion of mothers who lived 10 km or more from a health facility (2, [66.7%]) and sought healthcare for circumcision related morbidity from an informal care provider was higher than those who lived within 1 km from a health facility (3, [14.3%]). There were no data on type of health care provider available from the Neovita trial population.
6.5 Discussion
There appeared to be substantial differences in healthcare seeking in families with infants experiencing circumcision related morbidities and those with other acute illnesses in the Brong Ahafo Region of central rural Ghana. A much lower proportion (15.2%) of families of infants with circumcision related morbidities sought healthcare compared to families who had an infant with an acute illness in the larger Neovita population (88.1%). However, the proportion of hospital admissions in infants of families who sought health care was almost the same
(11% in the Neovita study and 9% in the circumcision study). There was some indication that mothers of low birth weight infants had poorer care seeking for circumcision related morbidities than mothers of infants with normal birth
109 weight infants. There were no other clear determinants of healthcare seeking for acute illnesses or circumcision related morbidities in the study populations.
A study conducted in Nigeria also reported lower healthcare seeking by families of male infants with circumcision related morbidities compared to infants with other health problems46. Other studies have suggested that poor care seeking behaviour after circumcision can be due to family expectations that complications will heal without any intervention46,73, poor recognition of illness39,118,119, poor recognition of severity of illness39,120-122, and beliefs that some infant illnesses are not meant to be treated in hospitals37,39.
Interestingly, the proportion of hospital admissions post care seeking was almost the same in the Neovita and circumcision populations (11% of infants with acute morbidities whose families sought care were hospitalised and 9% of infants with circumcision related morbidities whose families sought health care were also hospitalised). This indicates that families may not be seeking care for many circumcised infants in the study area with morbidities that do require hospital admission and specialised hospital care. Overall, hospital admission rates were similar to a recently published study from Ghana that reported rates of 9 to 18%123.
The determinants of care seeking for acute illness have been reported in low and middle income countries39,124,125. However, few studies have reported
110 population based data on the care seeking patterns of families with infants who experience circumcision related morbidities and acute illness. There was some indication that mothers of low birth weight infants had poorer care seeking for circumcision related morbidities (5%) than mothers of infants with normal birth weight infants (16%) though numbers were small. These findings were similar to other studies that reported poor care seeking for acute illness91 and vaccination126 in families of low birth weight infants. The study found no evidence that other socioeconomic and demographic factors influenced families care seeking behaviour for circumcision related morbidity and acute illness in the population. These findings contrast with that of other earlier studies from
Ghana39,123,126,127 and Bangladesh128 that reported higher care seeking for infants with acute illness by richer families, and mothers with higher educational levels. This indicated that health care seeking behaviour for acute illness may be improving over time in rural Ghana.
Overall, the study found that more families sought care for a circumcision related morbidity from formal (68.6%) than informal (13.4%) healthcare providers. Unsurprisingly, most families of infants circumcised by a formal provider also sought care from a formal provider for complications. However, fewer families of infants circumcised by an informal provider sought care from an informal provider (45%) than a formal provider (55%). These findings differed from another published study from Kenya73 which reported families of
111 male infants who were circumcised by an informal provider were two fold less likely to seek care for a circumcision related morbidity from a formal healthcare provider73. The study was not able to assess the type of care provider for acute illness in the Neovita study as these data were not collected in the Neovita study.
Generally, healthcare seeking for acute illness in the Neovita study population was very high (88.1%). The proportion of families seeking health care in this study was higher than data reported by the Ghana Demographic and Health
Survey in 2008, close to the time of conducting the Neovita study (acute respiratory infection 51.0%, fever 51.0%, and diarrhoea 35.2%)90. Care seeking for acute illness in rural populations has also been reported to be around 61% in
Kenya122, 47% in Nigeria59 and 27% in Vietnam129. The high healthcare seeking patterns for acute illnesses seen in this study population might be due to recent communications and promotion of care seeking by the local government district health management teams. However, that these communications do not appear to have influenced care seeking for circumcision related morbidities to date.
The most important limitation to this study was the small sample size in the circumcision cohort due to the low level of health care seeking so the study was unable to statistically compare care seeking in the circumcision and total study cohort. However the study was able to assess the association between key socio
112 demographic factors and healthcare seeking patterns in the 22,955 infants in the community level population based Neovita cohort. Another limitation was the used of family level recall of health service utilisation which could have led to some episodes of care seeking been under reported in the circumcised cohort.
However, both families and data collectors were blinded to the study hypothesis and the study trained fieldworkers reviewed infant health records and checked for consistency with a family’s report. The Neovita study lacked data on the type of provider care was sought from. However, the population based nature of the data on whether care was sought for acute illnesses, serves as a proxy for care seeking patterns. Understanding the type of care provider sought for acute illnesses would be important to investigate in the future.
6.6 Conclusions
This study has implications for policy and program development. Infant male circumcision is commonly practiced in many parts of sub-Saharan Africa including this study area. The study reported substantial differences in healthcare seeking patterns in families in the two cohorts in the Brong Ahafo
Region of central rural Ghana. A much lower proportion of families of infants with circumcision related morbidity sought healthcare for acute morbidity compared to families of infants with acute illness. The findings from this study indicated that government and non-governmental organisations need to prioritise circumcision and improve awareness of this highly prevalent but
113 potentially dangerous practice. Circumcision carries a triple burden of high prevalence, high morbidity load and poor care seeking. Community awareness of the burden, consequences and solutions needs to increase. This requires the involvement of community leaders and other stakeholders and include behaviour change communication strategies.
114
Chapter 7. Effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes: Protocol for systematic review.
This chapter presents the fourth PhD paper, a protocol for systematic review published in the Journal of Systematic Reviews in 2016130 (Appendix F). This paper directly related to the thesis objective four; to investigate and identify gaps in the literature on the effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes in any setting.
7.1 Abstract
Background: There has been an expansion of circumcision services in Africa as part of a long-term HIV prevention strategy. However, the effect of infant male circumcision on morbidity and mortality still remains unclear. Acute morbidities associated with circumcision include pain, bleeding, swelling, infection, tetanus or inadequate skin removal. Scale-up of circumcision services could lead to a rise in these associated morbidities that could have significant impact on health service delivery and the safety of infants. Multidisciplinary training programmes have been developed to improve skills of health service providers but very little is known about the effectiveness of health service provider education and training in infant male circumcision on short and long term morbidity outcomes. This review aimed to evaluate the effectiveness of
115 health service provider education and/or training for infant male circumcision on short and long-term morbidity outcomes.
Methods: The review will include studies comparing health service providers who have received education and/or training to improve their skills for infant male circumcision with those who have not received education and/or training.
Randomised controlled trials (RCTs) and cluster RCTs will be included. The outcomes of interest are short-term morbidities of the male infant including pain, infection, tetanus, bleeding, excess skin removal, glans amputation and fistula. Long-term morbidities include: urinary tract infection (UTI), HIV infection, and abnormalities of urination. Databases such as MEDLINE (OVID),
PsycINFO (OVID), EMBASE(OVID) , CINAHL, Cochrane Library (including
CENTRAL and DARE), WHO databases and reference list of papers will be searched for relevant articles. Study selection, data extraction and synthesis and risk of bias assessment using the Cochrane risk of bias assessment tool will be conducted. We will calculate the pooled estimates of the difference in means and risk ratios using random-effect models. If insufficient data are available we will present results descriptively.
Discussion: This review appears to be the first to be conducted in this area. The findings will have important implications for infant male circumcision programmes and policy.
Systematic review registration: PROSPERO CRD42015029345
116
7.2 Background
There has been an expansion of circumcision services in areas of Africa as part of the long-term human immunodeficiency virus (HIV) prevention strategy.
This has been initiated from the results of three randomised controlled trials that showed a 60% protective effect of circumcision during adulthood on HIV acquisition8-10. Despite the outcomes of these previous randomized controlled trials (RCTs) the effect of infant male circumcision on HIV risk still remains unclear. Acute morbidities associated with circumcision also include pain, bleeding, swelling, infection, tetanus or inadequate skin removal. With increased circumcision rates there is potential for a rise in these associated morbidities that could have a substantial impact on health service delivery and patient safety.
Infant male circumcision is commonly practiced in many parts of the world, by both formal health service providers such as doctors/surgeons, medical assistants, midwives/nurses and informal health service providers including traditional circumcision providers, traditional birth attendants, religious leaders, traditional medicine men/women and other health staff. As a surgical procedure, it is expected that male circumcision be performed at the required standard by all circumcision service providers with the requisite training. For example, health service providers need to be trained on: how to perform circumcision in aseptic conditions; use of equipment and tools; measures to
117 reduce morbidity; follow-up care; and education to families. Yet, health service providers often do not receive the required training and education to perform these operations to an optimal standard resulting in complications ranging from
1.7% to 7.6%4 following infant male circumcision.
A recent prospective study of complications of neonatal circumcision in Nigeria found that complications due to male circumcision was 10.9% and as a result have recommended training of health service providers to reduce the rate of complications78. Another study conducted in the United States among doctors performing neonatal circumcision also indicated the lack of structured training on neonatal circumcision for doctors. The study attributed the high complication rate and poor paediatric urology results to uncoordinated and informal training schemes for doctors who performed neonatal circumcisions.
This study also recommended structured and formal training as strategy to reduce complications associated with infant male circumcisions47.
It is expected that morbidities associated with infant circumcision could be reduced if health service providers are given the necessary education and/or training to perform circumcision. Several training programmes have been developed to improve the skills of health service providers among diverse health interventions. What is not yet clear is whether the provision of education and/or training would improve health service provider skills in infant male
118 circumcision leading to reduce rates of short and long-term morbidity outcomes.
There has yet to be a systematic review on this topic. A recent Cochrane review in 2012 examined the effects of routine neonatal circumcision for the prevention of urinary tract infections in infancy. In this review, no RCT, cluster RCT or quasi-RCT were located50. Another systematic review assessed RCTs of interventions to improve the safety and efficacy of nontherapeutic male circumcision of any age was conducted in 2010 and 8 RCTs were located. Only two of the studies were conducted among infants. The review did not find any studies that reported on infant circumcision and sexually transmitted infection, penile cancer, urinary tract infections or HIV/AIDS49.
In 2008, the World Health Organisation (WHO) Department of Reproductive
Health and Research funded two workshops to document infant circumcision practices in Africa30. One recommendation from the workshops was that further study of current infant circumcision practices was needed to inform the development of circumcision policies. As infant male circumcision services are being promoted in areas of Africa, the review will help to provide an understanding of the gaps in the literature on the effect of health service provider education and/or training for infant male circumcision on short and long term morbidity outcomes
119
Objectives
The primary aim of the review is to assess the effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes in any setting.
Review question
What is the effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes in any setting?
7.3 Methods
7.3.1 Protocol development
We developed the systematic review protocol based on recommendations of the
Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols
(PRIMSA-P) statement131, and provided a populated checklist in Additional file
(Appendix B). The review protocol was revised by the authors and is registered on the PROSPERO database (CRD42015029345 which is available at http://www.crd.york.ac.uk/PROSPERO/register_new_review.asp?RecordID=29
345&UserID=14548)
7.3.2 Types of studies
Randomised controlled trials (RCTS) and cluster-randomised controlled trials will be included for the review. If possible, we will include published abstracts if there is sufficient information to allow us to assess study eligibility and risk of bias. If sufficient information is not available, the study would await assessment
120 pending the publication of the full trial report, or the provision of further information by trial authors.
7.3.3 Participants
Formal and informal health service providers will be included. Formal health service providers are professionally trained, licensed and regulated providers of medical services and includes; doctor, surgeon, medical assistant, midwife, nurse. Informal health service providers are untrained, unlicensed, unregulated or informal private providers of medical services and includes; traditional circumcision providers, traditional birth attendants, religious leaders, traditional medicine men/women and other health facility staff. Health service providers will be excluded if they are undergoing medical training.
7.3.4 Types of Interventions
Intervention
The intervention is additional education and/or training to improve health service provider skills in infant male circumcision. Training could be peer-to- peer, hands on experience, lectures, practical training, demonstrations, small- group discussions, structured training e.g. didactic learning, or simulation training. Training could occur in any environment e.g. hospital, clinic, family homes or provider’s homes. Education could include topics on how to perform circumcision, use of equipment, how to ensure sterility, morbidity control, communication, and follow-up care.
Control condition
121
There are two different types of control groups (i) Health service providers who have not received education and/or training; and (ii) Health service providers who have received standard education and/or training.
Studies of complex training interventions (in which training is combined with a larger health system intervention) will be excluded.
7.3.5 Types of outcome measures
The main (primary) outcomes are short- and long-term all-cause morbidity which may include pain, bleeding, excess skin removal, glans amputation, fistula, infection, tetanus, urinary tract infection (UTI), HIV infection, abnormalities of urination, other identified morbidities and all-cause mortality.
Short term morbidity is defined as a morbidity occurring within six months of the circumcision procedure. Long-term morbidity is defined as greater than six months from the time of the circumcision procedure
The secondary outcomes are defined as presentations to clinic or hospital, family satisfaction, health service provider satisfaction, skill improvement in health service provider, knowledge improvement in health service providers and training costs.
7.3.6 Search strategy
The databases to be used for searching the relevant trials include; Cochrane
Central Register of Controlled Trials (CENTRAL) (The Cochrane Library),
MEDLINE (Ovid), EMBASE (Ovid), and Cochrane Database of Systematic
122
Reviews, Database of Abstracts of Reviews of Effects (DARE). An example of the MEDLINE search strategy is in Additional file (Appendix C).
Clinical trial registries including: ClinicalTrials.gov (http://clinicaltrials.gov/),
Controlled Clinical Trials (http://www.controlled-trials.com),World Health
Organisation (WHO) International Clinical Trials Registry Platform
(who.int/ictrp/en/)132, and UK Clinical Research Network Study Portfolio
(public.ukcrn.org.uk/search/)133.
The search period will be from 1985 to 2015 in all languages. Translation assistance will be sought.
7.3.7 Searching other sources
We will hand search reference lists from relevant articles chosen for potential inclusion in this review to identify further relevant studies. We will contact authors of included studies to determine whether there are any additional studies published, ongoing or unpublished that may be relevant.
7.3.8 Study selection
All review titles and abstracts retrieved through the search strategy will be reviewed independently by two authors to identify studies that meet the inclusion criteria. Inclusion criteria at the title and abstract review level will be limited to any primary study (quantitative) reporting on intervention with focus on health service provider’s education and/ or training to improve provider skills on infant male circumcision. Exclusion criteria will include infant greater than 12 months, review articles, qualitative and opinion articles. Once articles have been
123 identified, full text articles will be found and independently assessed by two independent review authors. For any disagreement, a third reviewer will be asked to assess the trial to determine eligibility for inclusion in the review.
Authors will also be contacted for further clarification if necessary. We will document reasons for exclusion and reference manager Endnote X7 will be used during this selection process.
7.3.9 Data extraction and management
A data extraction form will be developed and pre-tested before it will be used to extract data from the eligible studies. Two review authors will independently extract data from the included studies. Any disagreement between the two authors will be discussed. Further disagreement will be resolved through discussion with a third reviewer. Retrieved information will be considered based on context, intervention and outcome and includes; methodology, type of randomisation, study setting, recruitment, loss to follow-up rates, study population, participant’s demographics, interventions; description of the intervention and control groups. The study outcomes; primary and secondary outcomes, time points of measurements, indicators of assessment of risk of bias.
Authors, study funding and any information by the authors that may be useful.
Non-English language articles will be translated. Authors will be contacted for additional information if required. All disclaimer statements will be assessed.
124
7.3.10 Risk of bias assessment
The Cochrane risk of bias assessment tool134 will be used to assess the risk of eligible studies by two review authors. In case of disagreement the third author will be involve in the assessment and resolution of the risk of bias assessments.
The assessment will be categorised as either low, high or unclear risk of bias with explanation for each domain. The risk of bias domains include: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (Detection bias), incomplete outcome data (Attrition bias), selective reporting (Reporting bias) and other bias.
Authors will be contacted to provide additional information if data for assessment of risk of bias from eligible studies is inadequate. These data will be included in the data extraction processes. We will not exclude studies based on risk of bias assessment. The information will be used in the analysis and report of the review findings.
7.3.11 Overall risk of bias
The risk of bias summary tool will be used to collate the risk of bias for all eligible studies. Studies with a high or unclear risk of bias in the domain for sequence generation, similarity of baseline outcome measurements, completeness of outcome data, and other risks of bias, will be considered high risk. Conclusions
125 will also take into account the impact of the bias, and whether it is likely to bias the findings of the study.
7.3.12 Measures of treatment effect
Data will be entered into RevMan 5.3 software. Risk ratios (RR) and 95% confidence intervals will be calculated for dichotomous data including the proportion of infants reported as having short or long-term morbidity outcome, infant mortality, presentations to clinic or hospital, skill improvement as a result of the intervention, adverse events and training costs.
We will report mean difference (MD) for continuous outcomes and standardised mean difference (SMD) if different scales are used (such as family satisfaction, health service provider satisfaction and increased knowledge) with
95% confidence intervals for outcomes in each of the studies. The level of significance will be provided in case the above is not available. Chi-square test will be used to assess heterogeneity.
7.3.13 Unit of analysis
It is possible that cluster RCTs may be included in this review. If they are to be included in the meta-analysis, we will first determine if the authors have appropriately controlled for effects of clustering in the study. If there is doubt the authors will be contacted for clarification. If the error has not been corrected and the data are available, we will derive an estimate of the intracluster correlation coefficient (ICC) from the study. If the data are not available, we will
126 determine the ICC using a similar trial or from a study with a similar population.
We will report whether an ICC has been used and conduct a sensitivity analysis to determine the effect of using an ICC. If the clustering has been accounted for we will determine whether this has been adequately completed. Where authors have controlled for clustering, there is little difference between the study designs, or where there is unlikely to be an interaction between the effect of the intervention and the choice of randomisation method, the data from cluster
RCTs may then be combined with data from individual RCTs. If we are unable to adjust for incorrect statistical methods used by the cluster trials, and cannot estimate the ICC with any a degree of confidence, we will exclude the trial134.
7.3.14 Dealing with missing data
We will attempt to contact authors by email for any missing data.
7.3.15 Assessment of heterogeneity
We will construct forest plots to examine heterogeneity between interventions and quantify the impact of the heterogeneity using the I² and Chi- square statistics. If we identify a substantial level of heterogeneity in trials (for example, the I² is more than 30% to 60%, the P value is less than 0.10 in the Chi- square test for heterogeneity, or there is a different magnitude and direction of effects), we will perform subgroup analyses to explore the possible causes of statistical heterogeneity135 .
127
7.3.16 Assessment of reporting biases
If we are able to pool more than 10 trials, we will create and examine a funnel plot to explore possible small study effects and publication biases.
7.3.17 Data synthesis
Where data is available, we will perform a single meta-analyses for each outcome using RevMan 5.3 software. For each outcome we will firstly group by the type of control and analyse each group separately. If there is no differences we will combine the two groups for an overall analysis. For quantitative data we will calculate the pooled estimates of the SMDs or MDs and 95% confidence intervals using the random-effect model. We will also calculate pooled RR estimates and 95% confidence intervals using the random-effect model. Where statistical pooling is not possible the findings will be analysed descriptively.
7.3.18 Assessing methodological quality
We will use the Grades of Recommendation, Assessment, Development, and
Evaluation (GRADE) approach to assess the quality of the outcomes reported in this review136.
7.3.19 Subgroup analysis and heterogeneity
We will undertake subgroup analysis for the primary outcome measures if significant heterogeneity is identified in the trials for the following categories; formal and informal health service providers and low, middle, and high income countries. PROGRESS-Plus: An assessment of participant’s socio-demographic characteristics and subgroup outcomes (including, socio-economic status,
128 provider education, occupation of provider, ethnicity, religion), settings of procedure, location of provider, risk groups of health system non-regulation or recognition, such as informal providers particularly traditional providers.
7.3.20 Sensitivity analysis
We planned to conduct sensitivity analysis to assess the impact of excluding studies with a high risk of bias assessment on the primary outcomes.
7.4 Discussion
The findings of our review will have significant implications for infant male circumcision programmes and policy in any setting. Both formal and informal health service providers will continue to play important roles in the provision of infant male circumcision services globally. The association between poverty, infant health status and access to appropriate health care are relevant for decisions regarding the education and training to improve health service provider skills on infant circumcision to reduce morbidity in any setting. In settings where there are few formal health service providers and/or poor access to health facilities and families continue to seek services from diverse health service providers, education and training may be an effective way to reduce both short and long-term morbidity outcomes. Similarly, in places where access and high quality health service system are available but families decide to rely on diverse health service providers for services, education and training, and
129 structured collaboration with the health system may be an effective way to reduce morbidity associated with infant male circumcision.
130
Chapter 8. Health service provider education and/or training in infant male circumcision to improve short and long term morbidity outcomes: A systematic review
This chapter presents the full report of the systematic review and was the research conducted to address thesis objective four; to investigate and identify gaps in the literature about the effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes in any setting. The systematic review report has been submitted to the peer reviewed journal PLoS One137.
8.1 Abstract
Background: Education and training to improve patient health outcomes has been shown to be effective for surgeons and doctors. Yet, there is limited information about the effectiveness of this intervention on infant male circumcision outcomes. This study was conducted to assess the effectiveness of education and/or training for informal and formal health service providers in infant male circumcision on morbidity or mortality outcomes through a systematic review of the literature.
Methods: We searched Cochrane Central Register of Controlled Trials
(CENTRAL), MEDLINE, EMBASE, CINAHL, Global Health, Cochrane
Database of Systematic Reviews and Database of Abstracts of Reviews of Effects and clinical trial registries in all languages from January 1985 to June 2016.
Inclusion criteria were limited to any primary study that aimed to improve
131 health service provider skills in performing infant male circumcision. The primary outcomes were all-cause morbidity and all-cause mortality.
Results: We identified 1229 studies. There were no randomised controlled trials
(RCT). We were able to include only three observational studies from the USA that satisfied our criteria; all of which examined the effect of training on the skills and knowledge of medical doctors. No study involved informal circumcision providers. Information on the duration of training required for both medical doctors and informal providers to become competent in infant male circumcision was not available. All included studies were low quality.
Conclusions: This review found low quality evidence of the effectiveness of health service provider education and training in infant male circumcision and no studies involving non-medical health care providers. High quality studies of simple training packages to improve education and training of circumcision providers, especially informal non-medical providers in low income countries are needed.
8.2 Background
Male circumcision is the complete or partial removal of the foreskin of the penis and is performed for cultural and medical reasons1,2,52. Within the last ten years, there has been additional focus on medical male circumcision as part of the
African human immunodeficiency virus (HIV) prevention programs, which is jointly rolled out by the World Health Organization (WHO) and the Joint
132
United Nations Programme on HIV/AIDS (UNAIDS)20,138. The scale-up of male circumcision services is the result of randomised controlled trials (RCTs) conducted in Africa showing that circumcision reduced the risk of acquiring
HIV infection in heterosexual men by 60%8-10,20. However, complications associated with circumcision are also experienced including excess bleeding and incisions, pain, infection, swelling, tetanus, or inadequate skin removal5,14.
As male circumcision services are expanded, there is a likelihood of complications resulting in poor outcomes for circumcised infants and men as well as an increased impact on health service delivery. This systematic review is the full report of our protocol published in 2016130.
Infant male circumcision is a surgical procedure performed by formal health service providers including medical assistants, doctors/surgeons, midwives/nurses, and informal health service providers such as religious leaders, traditional birth attendants, traditional circumcision providers, traditional medicine men/women and other health staff. As a result it would be expected that health care providers would have appropriate education and training needed to perform circumcision to optimal standards. For example, they would need to be able to complete circumcision with aseptic techniques; anaesthesia; use of equipment and tools; processes to reduce complications; follow-up care; informed consent and education for families. Although these skills appear to be simple and easy to implement, published studies have
133 shown that health care providers are not properly trained to perform circumcision to acceptable standards14,46-48. As a result, complications range from
1.5% to 20.2% after circumcision have been reported14,27.
The WHO has developed a simple training program and manual to improve the knowledge and competency of health care providers in performing infant male circumcision in a variety of settings1. However, this training package does not appear to have been formally evaluated. What is also not yet clear is the effectiveness of training to health professionals who have had no prior training in surgery, including informal providers and the duration and intensity of the training required. It is also unclear whether the provision of education and/or training using simple packages such as the WHO manual would improve non- surgical health service provider skills in infant male circumcision, leading to reduced rates of both short- and long-term morbidity outcomes.
To date, there has been no systematic review investigating the effectiveness of health service provider education and/or training in infant male circumcision to improve morbidity outcomes. We have identified only two other systematic reviews on infant male circumcision49,50. The most recent review conducted in
2012 found there were no studies available to determine the effect of infant male circumcision in preventing urinary tract infections50. The other review found eight RCTs of interventions aimed at improving the safety and efficacy of
134 nontherapeutic male circumcision at any age. Only two of the studies were conducted among infants and assessed the relationship between circumcision and pain distress, and mother and infant bonding49.
The WHO office of Reproductive Health and Research sponsored stakeholder meetings in Ghana and Nigeria in 2008 to discuss infant male circumcision in sub Saharan Africa for HIV prevention program and policy2,30. A recommendation from these meetings included additional research in infant male circumcision in Africa to guide the scale up of circumcision services.
Thus, this review will explore the gaps in the literature about the effectiveness of health service provider education and/or training in infant male circumcision on morbidity and mortality outcomes.
8.2.1 Objectives
The overall aim of this review was to assess the effectiveness of health service provider education and/or training in infant male circumcision on morbidity and mortality outcomes in any setting. Our secondary outcomes were: post- surgery presentations to clinic or hospital, knowledge improvement in health service providers, skill improvement in health service providers, health service provider satisfaction and family satisfaction, and training costs.
Pre-specified subgroups were: (i) health service providers with no prior training in surgery including traditional (informal) providers in low income countries
135 and (ii) simple training packages such as the WHO package. We also assessed the duration of the training provided for service providers.
8.2.2 What question does this review aim to answer?
What is the effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes in any setting?
8.3 Methods
8.3.1 Protocol
The review protocol was published in 2016 and was based on recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses
Protocols (PRIMSA-P) statement130,131 (Appendix B). The review protocol was also registered with PROSPERO (CRD42015029345).
8.3.2 Types of studies
We included all study designs: individual RCTs, cluster-RCTs, quasi-RCTs, non-RCTs, controlled before and after studies, prospective cohort studies, retrospective cohort studies, historical controlled trials, nested case-control studies, cross-sectional studies, and before and after comparisons. Qualitative studies, review and opinion articles without a quantitative component were specifically excluded.
8.3.3 Types of participants
We included studies of interventions that involved formal health service providers and informal providers. We defined formal health service providers
136 as professionally trained, licensed and regulated providers of medical services and included: doctors, surgeons, medical assistants, midwives and nurses.
Informal providers were defined as untrained, unlicensed, unregulated or informal private providers of medical services and included: traditional circumcision providers, traditional birth attendants, religious leaders, traditional providers and other health facility staff.
Male infants aged under twelve months were included as the circumcision population.
8.3.4 Types of interventions
Intervention
We included all interventions that involved education and/or training to improve health service provider skills in infant male circumcision. Training could be peer-to-peer, hands on experience, lectures, practical training, demonstrations, simulations, small-group discussions, structured training e.g. didactic learning. Training could occur in any environment e.g. hospital, clinic, family homes or provider’s homes. Education could include topics on how to perform circumcision, use of equipment, how to ensure sterility, morbidity control, communication, and follow-up care.
High duration training was defined as more than four hours of training, low duration training was defined as one hour or less of training. If the duration of training was unclear it was classified as ‘unknown’. Practical 'hands on training'
137 was defined as the provision of hands on demonstration to doctors and nurses to acquire surgical skills under supervision139-141. Simulation-based training was defined as means for doctors and nurses to participate in acquiring surgical skills without the use of live patients142,143. Video training was defined as the use of computer-based training model for learners to acquire surgical skills144.
Control condition
Our control groups were (i) Health service providers who had not received any education and/or training; and (ii) Health service providers who had received standard education and/or training.
We excluded studies in which training was combined with a multi-faceted health service delivery intervention.
8.3.5 Types of outcome measures
We included studies that reported data relating to: short- and long-term all- cause morbidity including pain, bleeding, excess skin removal, glans amputation, fistula, infection, tetanus, urinary tract infection (UTI), HIV infection, abnormalities of urination and other identified morbidities and all- cause mortality. Short term morbidity was defined as a morbidity occurring within six months of the circumcision procedure. Long-term morbidity was defined as greater than six months from the time of the circumcision procedure.
Our secondary outcomes were defined as: post-surgery presentations to clinic or hospital, knowledge improvement in health service providers, skill
138 improvement in health service providers, health service provider satisfaction, family satisfaction, and training costs.
8.3.6 Search strategy
We searched Cochrane Central Register of Controlled Trials (CENTRAL) (The
Cochrane Library), MEDLINE (Ovid), CINAHL (EBSCO), Global Health
(OVID), EMBASE (Ovid), Cochrane Database of Systematic Reviews and
Database of Abstracts of Reviews of Effects (DARE) and clinical trial registries in all languages from 1985 to June 2016.
Clinical trial registries searched were: ClinicalTrials.gov
(http://clinicaltrials.gov/), Controlled Clinical Trials (http://www.controlled- trials.com),World Health Organisation (WHO) International Clinical Trials
Registry Platform (who.int/ictrp/en/)132 and UK Clinical Research Network
Study Portfolio (public.ukcrn.org.uk/search/). We also searched article references and online resources.
8.3.7 Searching other sources
We hand searched reference lists from relevant articles to identify further relevant studies. We contacted authors of included studies to determine additional relevant studies.
8.3.8 Study selection, inclusion and exclusion criteria
All titles and abstracts were reviewed independently by two authors (TG, NS).
Inclusion criteria at the title and abstract review level were limited to any
139 primary study that aimed to improve provider skills in infant male circumcision. Exclusion criteria were: outcomes assessed only in persons greater than 12 months of age, review articles, qualitative and opinion articles.
Studies of interventions in which training was combined with a multi-faceted health service delivery intervention were also excluded. Full text articles were independently assessed by two independent review authors (TG, NS). If there were disagreements, a third reviewer (KE) was asked to assess the study to determine eligibility for inclusion in the review. We used reference manager
Endnote X7 during this selection process and documented reasons for exclusion.
8.3.9 Data extraction and management
A data extraction form was developed and pre-tested before it was used to extract data from the eligible studies. Two review authors (TG, NS) independently extracted data from the included studies. Any disagreement between the two authors were discussed. Further disagreement was resolved through discussion with a third reviewer (KE). Retrieved information was considered based on context, intervention and outcome. Methodology variables collected were type of randomisation, study setting and training could occur in any environment e.g. hospital, clinic, family homes or provider’s homes, recruitment and loss to follow-up rates. Study population variables collected were age, sex and year of training. In particular, we collected information on the type of health professional who received the training or education, the
140 timing of their training (e.g if it was pre-service [e.g during their university studies or qualification course] or in-service [while they were working in a defined position]); and if they had prior surgical training. We also classified service providers as formal health service providers or informal providers using the definitions described above.
8.3.10 Interventions including description of the intervention and control groups
We collected detailed information on the type of interventions and control group including who conducted the training/education (e.g peer-to-peer, trained educator); the mode of delivery (e.g practical ‘hands on’ training, simulation, video, demonstrations, small-group discussions, lectures); and the content (e.g how to perform circumcision, use of equipment, how to ensure sterility, morbidity control, communication, and follow-up care); the duration of the training/education and if there was follow up ‘on the job’ support, supervision or refresher training. We also assessed the duration of the training.
8.3.11 Outcomes including detailed primary and secondary outcomes
We also collected data on time points of measurements including age of the subject at assessment and study funding.
8.3.12 Risk of bias
Studies eligible for the review were assessed by two review authors (TG, NS) and if there were disagreements a third author (KE) assisted with resolution.
We used the ROBINS-I (Risk Of Bias In Non-Randomized studies-of
141
Interventions) assessment tool for studies that were included in this review145.
Seven domains were assessed including confounding, selection of participants into the study, deviations from the intended interventions, missing data, measurement of outcomes, and selection of the reported result. The categories for risk of bias judgements for the ROBINS-I are ‘No information’ ‘Low risk’,
‘Moderate risk’, ‘Serious risk’ and ‘Critical risk’ of bias'.
Authors were contacted to provide additional information and no studies were excluded based on risk of bias assessment. The information was used in the analysis and report of the review findings.
8.3.13 Measures of treatment effect
We provided the findings of studies in a narrative format as described by the study authors.
8.3.14 Dealing with missing data
We contacted authors of the included studies and requested for missing data.
All authors responded to our request by providing information on missing data by email.
8.3.15 Assessment of heterogeneity
We did not assess heterogeneity, as the three studies that were included either did not assess the same outcomes and those that did had substantial variation.
8.3.16 Assessment of reporting biases
We did not examine funnel plots to assess reporting bias for small study effects and publication bias as there were only three studies included in the review.
142
8.3.17 Data synthesis
We initially planned to perform a stratified meta-analysis, however the outcomes reported in the three studies included were very heterogeneous thus a narrative only synthesis was performed.
8.3.18 Subgroup and sensitivity analysis
Pre-specified subgroups were: (i) health service providers with no prior training in surgery including traditional (informal) providers in low income countries and (ii) simple training packages such as the WHO package. However, no studies were found.
8.3.19 Quality assessment
We used the Grades of Recommendation, Assessment, Development, and
Evaluation (GRADE) approach to assess quality of the body of evidence reported in this review136. Two review authors (TG, NS) independently assessed the certainty of the evidence (high, moderate, low, and very low) using the five GRADE considerations (risk of bias, consistency of effect, imprecision, indirectness, and publication bias)146.
We created a 'Summary of findings' table based on the primary outcomes and those outcomes that were reported in the included studies. We based our
Summary of findings tables on guidance provided by the Effective Practice and
Organisation of Care Cochrane group147.
143
8.3.20 Differences between protocol and review
The following discussed the differences between our original published protocol5 and this review.
Objectives
An additional objective was also considered to help assess the duration of the training required. We included this to enable clear recommendations to be formulated. We also added two specific subgroups: (i) health service providers with no prior training in surgery including traditional (informal) providers in low income countries and (ii) simple training packages such as the WHO package. We also assessed the duration of the training provided for service providers. We included these subgroups to ensure recommendations would be relevant to countries with the highest burden of complications and the highest number of informal providers.
Study design
Prior to constructing the final search strategy, we decided to include observational studies as well as RCTs. This was because an initial search revealed no RCTs and we wished to gather all information which were particularly relevant to our review. We have included a full list of included study designs under the 'Types of studies', section 8.3.2 under methods.
Population
144
We also decided to expand the population to include medical students as we realised that the number of studies were likely to be low and because this is an important population for training and quality assurance.
Risk of bias
As we included non-RCTs within our review, our risk of bias assessment was completed using the ROBINS-I tool145.
8.4 Results
We identified 1159 studies from the database searches and 70 through the clinical trials databases. After removal of duplicates we identified 821 studies for screening. We excluded 760 from the first screening and assessed 61 full text records. After reading full text records we excluded 58 records due to the following reasons; no intervention completed; no primary data collection
(editorial or review article); not education and/ or training intervention; education and/ or training model described but intervention not completed; unable to retrieve abstract, or conference abstract; post training results only and insufficient data for pre-post analysis (Figure 8.1). Three studies were found to be eligible and we performed full data extraction on these studies. The flow diagram is shown in Figure 8.1.
145
Records identified through database Additional records identified through searching other sources (n = 1159) (n = 70)
Records after duplicates removed (n = 821)
Records screened Records excluded (n = 821) (n = 760)
Full-text articles excluded, with reasons Full-text articles assessed for (n = 58) eligibility No intervention completed (n = 61) (n = 32) Review article, no primary data collection (n = 10) Not education and/ or training intervention (n = 6) Education and/ or training model described but intervention not completed Studies included in synthesis (n = 4) (papers; studies) Unable to locate full text (n (n = 3) = 1) Conference abstract only (n = 1) Post training results only (n=3) Insufficient data for pre-
Figure 8.1 Study flow diagram, showing the results of the searches for the systematic review
146
8.4.1 Included studies
Details of included studies can be found in Table 8.1. The three studies eligible for inclusion were all prospective cohort studies with historical controls and were published from 2002 to 2013148-150.
8.4.2 Location and participants
All three studies were conducted in the USA148-150 and took place within a hospital setting (Table 8.1). There were a total of 55 participants involved in all of the studies. The study participants were doctors only148-150. No study included informal health service providers or traditional providers. The recruitment methods of two of the studies were self-selection148 and convenience sampling149. The recruitment methods for the other study was not described150.
None of the studies reported on the criteria used to include or exclude participants. There were marked variations in age, gender, year of training and type of medical personnel (Table 8.1). Two studies149,150 did not report on the baseline characteristics of participants. One study reported on the duration and intensity of training needed148. No study reported on morbidity or mortality.
One study assessed hands on training combined with video training148, another study assessed simulation with video training150 and one other study assessed simulation training with educational materials149. None of the studies assessed other forms of training.
147
Table 8.1 Description of included studies Study Study design Participants Setting Type of Description of Duration of the Outcome measures Data Results intervention intervention intervention collection method Chandran 2002148 Prospective Paediatric residents Hospital Hands on - Self-directed learning on The duration 1. Knowledge Author 1. Improvement in (USA) cohort study (n=19): training circumcision, the included the improvement developed knowledge score, first with First year residents circumcision procedure following: regarding analgesia assessment year residents (from 5.0 historical (n=11); second year and anaesthetic techniques - Residents for circumcision, tools; non- to 8.8, p=0.003) and controls residents (n=2) using a 26-point observed a median care of the validated senior residents (from senior residents (n=6) procedural checklist of 5 (range 3-7) uncircumcised 5.0 to 7.6 p=0.010). Female: 80% (n=15) including steps for the circumcisions penis and 2. Mean test scores on Age <30 years: 58% correct performance of before training. complications of training items increased (n=11) circumcision. -Residents circumcision from 5 (standard - Viewing of videotape on performed a 2. Skill deviation [sd], 1.8) using Gomco clamp median of 6 (range improvement in before training to 8.3 (sd, techniques and dorsal 3-12) circumcisions performing 1.0) after training nerve block during training circumcision (p=<0.003 - Observation of an under supervision. 3. Health service 3. Improvement in instructor performing - Median time provider self-report proportion of residents three to five circumcisions. required to perform of competence felt competent to The instructor explained a circumcision, 30 perform circumcision the steps during the minutes (range 15- The evaluation of with minimal procedure using a 45). outcome measures supervision from 0 (0%) checklist. was performed to 17 (90%) (p= <0.001) - Three to five directly prior to circumcisions performed and one month under direct observation after the using the training checklist intervention. with feedback received.
148
Roca 2012149 Prospective Residents (n=29): Hospital Simulation - Reading materials Duration of the 1. Knowledge Author Marginal improvement (USA) cohort study Family medicine based - A brief lecture/discussion intervention: Not improvement developed in knowledge on with residents (n=23); training reviewing key facts about reported regarding basic and assessment circumcision (from historical obstetrics/ anatomy, informed relevant anatomy tool; non- mean pre-test score of controls gynaecology consent, existing evidence, and procedural validated 12.1 to mean post test residents (n=6) indications, concepts. score of 14.5) from a contraindications, total score of 15. techniques, complications, The outcome pain management, measure was documentation, and use of assessed after institutional order sets. reading materials - Demonstration of both were provided and Gomco and Mogen after techniques. demonstration. - Supervised practice sessions using life size manikins and penile models using procedural checklists with feedback provided. - Residents performed demonstration using preferred technique on the high-fidelity simulator and evaluated using checklist (outcomes assessed). Smith 2013150 Prospective Paediatric residents Hospital Simulation - A computerised Duration of the 1. Skill Author 1. Improvement in (USA) cohort study naïve to newborn based enhanced visual learning intervention: Not improvement of the developed technical competency with circumcision (n=7) training curriculum for the reported simulated assessment (from mean test score of historical simulation of newborn technique tools; non- 85.3 to 97.4) with controls clamp circumcisions. 2. Increased validated trainees achieving 96% - Residents trained satisfaction in competence by second independently and then performing simulation. performed two circumcision. 2. The authors also simulations with a mentor. reported an increased in - The initial simulation The outcome comfort and reduced was completed with measure was anxiety to perform scored feedback. evaluated prior to circumcision (No data - Residents later circumcision were provided). performed a second simulation and simulation. within 30 post simulation
149
8.4.3 Risk of bias assessment
We assessed the risk of bias using the seven domain criteria of the ROBINS-I tool. We rated the overall risk of bias for all three studies as “critical risk of bias” (Table 8.2).
8.4.4 Categorisation of the interventions
The training interventions were stratified into two categories (hands on training and simulation-based training) and are summarised in Table 8.1.
Hands on training
Chandran assessed the effectiveness of hands on training148 for a group of trainee doctors (residents). This was implemented over 20 full working days
(rotation period). The residents observed an instructor performing three to five circumcisions. The instructor explained the steps during the procedure using a checklist. The residents then performed three to five circumcisions under direct observation according to the training checklist and received feedback. An instructional videotape on using Gomco clamp techniques and dorsal nerve block was also given before the supervised ‘hands on’ training.
Simulation based training
Two studies provided simulation based training. Smith 2013150, designed an individual training session involving a simulation model for newborn clamp circumcision. This was supported with video learning curriculum. In this study, residents were trained independently and then performed two simulations with a mentor. The initial simulation was completed with scored feedback. The
150
Table 8.2 ROBINS-I risk of bias assessment Author Confounding Selection Classification of Reporting bias Deviations from Missing data Measuring Overall bias bias interventions interventions outcomes assessment* Chandran 2002148 3 3 3 2 2 2 3 4 Roca 2012149 3 3 3 2 3 2 3 4 Smith 2013150 3 3 3 2 2 3 3 4 * 0= No information, 1= Low, 2=Moderate, 3=Serious, 4= Critical
151 residents later performed a second simulation without feedback. The duration of the simulations was unknown. In the second study (Roca 2012149), the investigators constructed a model of neonatal genitalia using a surgical glove finger to train doctors individually. The intervention included educational materials and demonstration of both Gomco and Mogen techniques followed by supervised practice sessions using procedural checklists with life size manikins and the penile models, and feedback was provided. Residents then performed demonstration using preferred technique on the high-fidelity simulator. A checklist was used for the evaluation.
8.4.5 Duration and intensity of intervention
Chandran reported on the number of circumcisions performed by residents and the time needed to complete a procedure148. In this study, residents observed a median of 5 (range 3-7) circumcisions before the training. During the training, each resident performed a median of 6 (range 3-16) circumcisions under supervision. The median time required to complete a circumcision was 30 minutes (range 15-45)148.
8.4.6 Outcomes
Primary outcome
Our primary outcomes were all-cause morbidity and all-cause mortality. There were no studies which assessed these outcomes.
Secondary outcomes
152
The planned secondary outcomes were post-surgery presentations to clinic or hospital, knowledge improvement in health service providers, skill improvement in health service providers, health service provider satisfaction and family satisfaction and training costs.
Knowledge improvement. Two studies148,149 assessed knowledge improvement in health service providers (Chandran 2002 and Roca 2014). Chandran 2002148, found that mean test scores for overall knowledge which included understanding of analgesia for circumcision, care of the uncircumcised penis and complications increased from 5 (standard deviation [sd] 1.8) before training to 8.3 (sd 1.0) after training (p-value=<0.003). Using a score from 0 to 10,
Chandran 2002148, also reported a mean post training score of 8.5 (sd 1.0) for knowledge and attitudes towards circumcision from a pre-training mean of 5.0
(sd 1.8). Significant improvements were also noted in first year residents (5.0 to
8.8, p=0.003) and in senior residents (5.0 to 7.6, p-value=0.01). In addition, significant improvements were also reported for knowledge regarding analgesia for circumcision (p-value=<0.01) and complications of circumcision
(p-value=<0.01). The study reported by Roca 2014149 did not demonstrate any significant increase in knowledge in circumcision (mean pre-test score of 12.1 to mean post test score of 14.5 of total score of 15, p-value=0.50149).
Skill improvement. Smith 2013150, reported a significant improvement in technical competency after the second simulation (mean, 97.4) from the first simulation mean of 85.3. The trainees also achieved 96% competence by the second
153 simulation. Chandran 2002148, reported that all residents receiving training were able to successfully complete the requirements of the program and performed two circumcisions independently with adequate competency and achieved a pass rate of 100% . No statistical analyses were performed in these studies.
Health service provider satisfaction. Chandran 2002148 assessed health service provider self-report about their own competence in performing circumcision and ‘perceived competence’ pre- and post-training. No resident reported feeling competent with performing circumcision prior to the training. At post training,
90% (n=17) felt competent to perform circumcision with minimal supervision
(p-value=<0.001).
Other outcomes. No study reported on post-surgery presentations to hospital or clinic, family satisfaction or training costs or duration of the training.
Chandran reported that residents who believed that circumcision training was
“necessary” increased, from (n=8) 45% pre-training to (n=14) 74% post-training
(p=0.003)148. Chandran also reported an increase in the proportion of residents who felt that anaesthesia was required for infant male circumcision from a pre training score of 74% (n=14) to 82% (n=15) post training148. No statistical tests were performed for this study.
8.4.7 Quality of the evidence
Overall, the quality of the body of evidence for all outcomes was judged to be very low using the GRADE criteria147 (Table 8.3). The body of evidence was initially given a rating of “low” due to the non-RCT study design. This rating
154
was then reduced to “very low” because of the high risk of bias for all studies
particularly regarding confounding and measurement bias.
Table 8.3 GRADE findings
Health service provider education and/or training in infant male circumcision to improve short and long term morbidity outcomes
Patients or population: Formal health service providers, informal providers and medical students Settings: Search strategy included all settings, however only studies from the USA were eligible for inclusion Intervention: The intervention is additional education and/or training to improve health service provider skills in infant male circumcision Comparison: (i) Health service providers who have not received education and/or training; and (ii) Health service providers who have received standard education and/or training.
Outcomes1 Impact Number of participants Certainty of the evidence (Studies) (GRADE)*
Short-term morbidity No studies assessed the impact - - of education/training on this outcome.
All-cause mortality No studies assessed the impact - - of education/training on this outcome.
Knowledge improvement in The provision of 48 ⊕⊖⊖⊖ health service providers education/education increased (2 studies) Very low1 knowledge of doctors in the performance of infant male circumcision
Skill improvement Education /training could result 26 ⊕⊖⊖⊖ (competency) in health in improved technical (2 studies) Very low1 service providers competency in the provision of infant male circumcision by doctors.
Post-surgery presentations No studies assessed the impact - - to clinic or hospital of education/training on this outcome.
* GRADE Working Group grades of evidence High = This research provides a very good indication of the likely effect. The likelihood that the effect will be substantially different† is low. Moderate = This research provides a good indication of the likely effect. The likelihood that the effect will be substantially different† is moderate. Low = This research provides some indication of the likely effect. However, the likelihood that it will be substantially different† is high. Very low = This research does not provide a reliable indication of the likely effect. The likelihood that the effect will be substantially different† is very high.
1The limitations in the studies including serious risk of bias, small sample size and the lack of a standardised control group reduced the GRADE from low to very low
155
8.5 Discussion
The systematic review was not able to provide clear evidence of the effect of health service provider education and training in infant male circumcision on all-cause morbidity or mortality outcomes or the secondary outcomes of knowledge improvement, skill improvement, health service provider satisfaction, post-surgery presentations to clinic or hospital and family satisfaction. Only three prospective cohort studies with historical controls were located and the risk of bias for each study was rated as “very high”. The overall quality of the body of evidence for knowledge and skill improvements were also considered to be “very low” using the GRADE criteria. Data on the duration and intensity of training required was limited. No studies were located that assessed informal service providers, low income country settings or simple packages of training.
We found only three studies that were specific to infant males148-150 which examined the effect of training on the skills and knowledge of medical doctors and were inconclusive. In addition, since our initial search we have become aware of one other small observational study which assessed the effect of competence-based training in infant male circumcision using the Mogen clamp on knowledge scores in non-physicians151
156
In this Ugandan study, clinical officers (n=10) and midwives/nurses (n=10) received five days of training including; didactic learning (i.e good clinical practice, informed consent, screening, anatomy of the penis, contraindications of circumcision, haemostasis and suturing, postoperative care, hygiene, complications and resuscitation), interactive sessions, role plays, and case scenarios. Each of the trainees also observed five circumcisions. This study found significant improvements in knowledge 20 days after the training program. Twenty percent of clinical officers and 10% of midwives/nurses had knowledge scores over 80% before the training. Post-training, all (100%) clinical officers and 60% of midwives/nurses had knowledge scores of 80% or more (p- value=0.043). These trainees then circumcised ten neonates (aged 1-28 days) under direct supervision and were assessed for competency using a standardised checklist. Statistically significant improvements in competence were observed between the first three procedures (baseline) and the fourth procedure (endline) (clinical officers; +5.87 points, 95% CI 3.14–8.60 and midwives/nurses +7.9 points, 95% CI 5.89– 9.86, p-value= 0.043). However, pain scores were not significantly different between the two groups (78.9% for midwives/nurses versus 77.5% clinical officers, p-value=0.887). All (100%) trainees were reported to achieve full competency as measured on the checklist by the seventh procedure. No long term follow-up of skill or knowledge retention was conducted. An overall complication rate of 3.5% was reported within 24 hours post circumcision151. This study was carefully conducted and
157
had good short term follow-up. However, it only involved 20 participants and had no long term follow-up so overall it is able to contribute little to the overall body of evidence to date.
There have also been six studies69,152-156 that have compared the effect of different types of devices in infant male circumcision. However, none involved informal providers. All studies were based in a hospital setting and the circumcisions were performed by medical doctors. In Zimbabwe, Mavhu et al 2015152, implemented an RCT which compared the safety of the AccuCirc clamp
(n=100) versus the Mogen clamp (n=50) in 150 male infants aged 6-54 days.
Training was provided to four doctors using the AccuCirc and Mogen clamp devices. The training included; didactic learning, use of infant penile anatomic model, practical hands on training, written examination and practical assessments. Trainees were required to score 100% on written exams and prove competency by performing five circumcisions with each device (data on pre and post training were not reported). This study reported no significant differences in complication rates between the two groups152. Other RCTs from
Botswana69, Zambia153, Iran154 and the USA155,156 also compared different infant male circumcision devices with little effect on the incidence of complications.
Infant male circumcision is commonly performed by informal providers in low income countries2,4,5,21,157. These providers often have no formal training in
158
circumcision but learn the trade through apprenticeship from another experienced informal provider. These providers are often highly valued by families for cultural and religious reasons. The most important part of this review was to assess the effectiveness of simple training packages such as the
WHO training program1 to improve the skills of these informal or traditional providers. However, our review found no studies in this important subgroup.
We were also unable to find any studies that assessed the effectiveness of training for informal health service providers in adult male circumcision. In contrast, the effectiveness of other types of surgical and procedural education and training to non-physicians in both rural and urban areas in low income countries, using hands on training and simulation techniques is well known.
For example, Ellard et al 2015158, reported that hands on training of non- physicians improved skills in health outcomes and reduce maternal and perinatal morbidity and mortality. In this study, the health centres in 14 districts in Central and Northern Malawi were randomised to receive a non- physician training package (which included clinical skills training in obstetrics, neonatal care and leadership)-(intervention arm) or no training package
(control group). A total of 54 non-physicians were trained in the 8 intervention districts from 2011 to 2013. The study evaluation was an interrupted time series evaluation that investigated differences within the control and intervention group at four time points before the intervention and seven time points after the
159
intervention at three month intervals. This training resulted in trend improvements in maternal mortality ratio in the intervention arm over time
(effect -38.11, SE 50.30, CI -157.06 to 80.84, p-value=0.47). Neonatal mortality also reduced by 31% (from 5.8 to 4.0 per 1000 live births) in the intervention group compared to an increase of 2% (from 9.6 to 9.8 per 1000 live births) in the control group over the 3 year period158. No statistical analyses were performed to compare the change in outcomes between the intervention and control groups in this study.
Simulation-based training is also considered to be effective and cost-effective way to provide education and training to non-physicians in trauma injury management. A Myanmar trauma training program trained community health workers to evaluate and manage trauma injuries from 2000 to 2013159. A post- training evaluation of 26 of these health workers in 2013 indicated that 75%
(n=18) felt competent to evaluate and managed trauma injuries, 96% (n=25) self- reported that the training was relevant to their work, all (100% n=26) felt time spent for training was valuable and 96% (n=25) self-reported that simulation models improved skills in trauma injury management159.
Our study had a number of limitations. We found no RCTs or cluster-RCTs that assessed whether specific education and/ or training interventions can improve health service providers knowledge and skills for infant male circumcision in
160
rural areas in low income countries. While we did include all study designs, only three studies in high income settings were able to be included.
Multifaceted interventions which include circumcision training as part of a wider training program are likely to show large effects, however we excluded these programs as we wanted to assess the effectiveness of circumcision training separately. Though we included observational studies, the risk of bias of included studies was very high, which made it difficult to draw conclusions.
However, we were able to use the ROBINS-I assessment tool to assess bias in these studies and this tool is a well validated tool for observational studies.
8.6 Conclusions
Implications for policy and practice
This review found very low quality of evidence on the effectiveness of health service education and training in infant male circumcision to improve short and long term morbidity outcomes. The review also found no studies that assessed the effectiveness of training to health service providers who have had no prior training in surgery including informal providers and no information on basic training packages. Information from low income countries and on the duration required for both medical doctors and informal providers to become competent in infant male circumcision was also limited. A number of researchers4,14,48,73,157 have recommended the expansion of the training of informal and traditional
161
providers in low income countries. The findings of this review indicate that these recommendations are based on very little evidence.
Implications for research
We found no RCTs or cluster-RCTs on the effectiveness of education and training in infant male circumcision on morbidity and mortality outcomes. To address this gap, more high quality randomised trials or observational studies should be conducted that assess the effectiveness of education and training in infant male circumcision. The studies should: i) include traditional and informal circumcision providers who perform most circumcisions in low income countries; ii) test low cost simple interventions that can be scaled up across low income countries including the WHO package; iii) include novel methods to improve education and training to reduce complications among informal non- medical providers; iv) provide information about the duration required for informal non-medical providers to become competent in infant male circumcision; and v) provide data on follow-up care. Strategies for scale-up in different geographic settings, especially low income countries, must also be identified.
162
Chapter 9. Discussion, conclusions and recommendations
This chapter presents the discussion, strengths, limitations and conclusions from the research findings. Recommendations and the implications of the findings for policy and practice are also presented.
9.1 Main findings in comparison with other epidemiological studies
9.1.1. The burden and risk of morbidity associated with infant male circumcision in a community-based setting
In this population based study in rural Ghana infant male circumcision was almost universal, a total of 93% of infant males in the study area were circumcised by 20 days of age. The study indicated a high risk of morbidity
(8.1%) and two deaths following circumcision. Important determinants of morbidity included poor hygiene practices such as not washing hands, the use of uncleaned instruments and not cleaning the penile area. Circumcision performed by formal providers was associated with similar risks to those performed by informal providers. However, hygiene practices appeared poorer in formal providers. Babies born to women with low education levels were particularly at risk.
It was not surprising that the morbidities and mortality observed in this study was higher than hospital-based studies from high-income settings including
USA, UK and Israel14,77,93. However, the morbidity risk was lower than the three 163
published African studies which were conducted in large urban hospitals in
Nigeria and Botswana. These studies reported an 18% (range 10-24%) risk of morbidity 6-16 weeks post infant male circumcision27,65,69. The case definitions used in those studies were similar to this study. However, infants with high complication risks were admitted to hospital and cannot be considered representative of the whole population. This thesis appears to be the first to analyse the population level burden associated with circumcision in the African community setting.
Overall, the prevalence of circumcision in this study was similar to published rates from other parts of West and North Africa2,27,65,94. The most common morbidities in this study were: excess skin removal, excess bleeding, inadequate skin removal, infection and abnormal stream of urine. These findings are similar to morbidities reported by other similar studies27,65,69,93,95. The study reported two deaths due to excessive incisions and haemorrhage but no deaths due to infection or tetanus.
Almost 60% of infant male circumcisions in this study were performed by informal providers. Other studies report that Wanzams are the most common providers of circumcision in rural West Africa2,21. The risk of experiencing a morbidity was not significantly increased if circumcision was performed by informal providers compared to formal health service providers in this study.
164
Other studies have reported both higher27,65,95 and lower2,14,17,52 risks of morbidity after circumcision performed by informal providers. However, the findings are consistent with two reviews, the first review was published in the
Lancet in 2007 and the second review was published in the European Urology
Focus in 2016. These reviews indicated that circumcision related morbidities were common in both formal and informal providers16,96.
Overall, hygiene practices were inadequate in the study area. These findings are consistent with recent studies which report unclean and unsterile circumcision practices amongst urban and rural African service providers2,21,73. Concerningly, this research found that hygiene practices such as handwashing and cleaning of the penile area were poorer in formal health care providers compared to informal providers. Reasons for this are unclear and will be investigated in a future study.
There was a twofold increased risk of morbidity from circumcisions conducted after 20 days of life. Similar findings were reported in a recent systematic review and studies from Kenya, Pakistan and the USA14,71,74-76. Older infants are at increased risk of complications because the procedure is more difficult in older infants. That is, it is harder for primary care providers to sedate or anaesthetise older infants and older infants struggle more. Furthermore, older infants are more mobile and the wounds are more likely to get dirty and
165
infected. The reasons for the lower risk of morbidity in younger infants is considered to be due to the simpler nature of the procedure during this period and faster healing14,29,75. Infants born to mothers with no formal education had a greater risk of circumcision related morbidity. The reasons were likely to be due to the fact that women with no formal education were more likely to be from poor families who have (i) less knowledge about hygiene practices and (ii) who are less able to create a clean uncontaminated living environment.
9.1.2 Factors which influence choice of infant male circumcision provider in rural Ghana
In this population based study in rural Ghana, infant male circumcision was performed by both formal (41%) and informal (59%) circumcision providers.
Both socio-economic status and geographic access to health facilities had important and independent effects on the choice of circumcision provider. The risk of receiving a circumcision from an informal provider increased with each level of deprivation and with the distance that families lived from health facilities. This research also found that families with the lowest household income were the most likely to pay for their circumcision. Poor families were also most likely to pay additional in-kind contributions.
The relationship between socio-economic status2,52,73,100, geographic access2,48,101, and choice of informal provider for infant male circumcision has been reported
166
in many studies in low and middle income countries. However, this study is the first to report data from a rural area in Africa with high population level coverage of infant male circumcision. This is also the first study to report the double burden that circumcision places on families of low socio-economic status. In this study, poor families were more likely to receive a circumcision from an untrained informal provider and also more likely to incur a significant economic cost.
In 2008, the “Free Maternal Care Policy”102 was introduced into the Ghana
Health Insurance Scheme103. Under the policy, all pregnant women and their infants up to 90 days postpartum and all children aged 90 days to 18 years are meant to receive free care in accredited public and private healthcare facilities.
The services that are covered include antenatal care, delivery care, postnatal care, and infant male circumcision. Mothers and children just have to be registered and receive a registration card. The registration process is free and there are meant to be no out of pocket expenses. However, there have been difficulties in enrolling many families into the scheme. This has been attributed to difficulties in accessing many areas of Ghana, especially the poorest and most disadvantaged areas104,105. In 2011, close to the time of conducting this study, only 33% of Ghana’s population were registered with 4.2% coverage for the poorest105. The most recent data from 2013 indicated that the national coverage still remains limited with only 38% registered106. Inequity in health insurance
167
coverage is likely to be an important driver of the costs of circumcision incurred by poor families that was reported. This study area is located in central rural
Ghana in the Brong Ahafo region and health insurance coverage in the Brong
Ahafo region was 45.9% in 2011105. However, there are no data on coverage of health insurance in the poorest families in the study area. Antenatal care and delivery services are also meant to be free under the Ghana health insurance scheme102,103 and similar inequities are also reported for these services. There are reports of poor women being charged unofficial and non-legitimate fees for delivery and postnatal care services107,108 . Reports of poor women and their babies being forcibly kept in birthing hospitals until their bills are settled have also been published107. Poor women have also been charged unofficial fees for antenatal109, delivery, and postnatal care services108 in accredited facilities because they were unable to confront authority figures110,111. Poor women are also less likely to be insured for delivery care compared to richer women in
Ghana112,113.
Additional economic costs of circumcision include the payment of ‘in-kind’ contributions. The payment of in-kind contributions was more common with
Wanzams (informal provider) (7%) than formal providers (3%) (doctors, nurses, and medical assistants) in this study. The poorest families also paid more in- kind contributions (31%) than the highest income families (12%). Two rural
Kenyan studies have reported the payment of in-kind contributions (chickens,
168
sheep, food and medical supplies) by families for circumcision48,73. In these studies medical practitioners (49%) and informal traditional providers (51%) received similar in-kind contributions. However, these studies did not provide any information on the in-kind contributions paid by poor and richer families within the same study area.
This study also reported that families of the Muslim religion were two-fold more likely to choose an informal provider than families with other religious affiliations. The Muslim religion is a well-known determinant of use of informal providers for circumcision in urban and rural Africa2,52,114 and many Wanzams are Muslim themselves21. Approximately, 70% of Wanzams who performed circumcisions in this rural study area were Muslims. The study also indicated that mothers with no formal education were more likely to choose an informal circumcision provider compared to mothers with secondary level education.
These data are also consistent with other African studies65. There were no obvious differences in choice of circumcision provider associated with other socio-demographic characteristics in this study.
9.1.3 Health care seeking behaviour of families
There appeared to be substantial differences in healthcare seeking in families with infants experiencing circumcision related morbidities and those with other acute illnesses in the Brong Ahafo Region of central rural Ghana. A much lower proportion (15.2%) of families of infants with circumcision related morbidities
169
sought healthcare compared to families who had an infant with an acute illness in the larger Neovita population (88.1%). However, the proportion of hospital admissions in infants of families who sought health care was almost the same
(11% in the Neovita study and 9% in the circumcision study). There was some indication that mothers of low birth weight infants had poorer care seeking for circumcision related morbidities than mothers of infants with normal birth weight infants. There were no other clear determinants of healthcare seeking for acute illnesses or circumcision related morbidities in our study populations.
A study conducted in Nigeria also reported lower healthcare seeking by families of male infants with circumcision related morbidities compared to infants with other health problems46. Other studies have suggested that poor care seeking behaviour after circumcision can be due to family expectations that complications will heal without any intervention46,73, poor recognition of illness39,118,119, poor recognition of severity of illness39,120-122, and beliefs that some infant illnesses are not meant to be treated in hospitals37,39.
Interestingly, the proportion of hospital admissions post care seeking was almost the same in the Neovita and circumcision populations (11% of infants with acute morbidities whose families sought care were hospitalised and 9% of infants with circumcision related morbidities whose families sought health care were also hospitalised). This indicates that families may not be seeking care for
170
many circumcised infants in this study area with morbidities that do require hospital admission and specialised hospital care. Overall, hospital admission rates were similar to a recently published study from Ghana that reported rates of 9 to 18%123.
The determinants of care seeking for acute illness have been reported in low and middle income countries39,124,125. However, few studies have reported population based data on the care seeking patterns of families with infants who experience circumcision related morbidities and acute illness. There was some indication that mothers of low birth weight infants had poorer care seeking for circumcision related morbidities (5%) than mothers of infants with normal birth weight infants (16%) though numbers were small. These findings are similar to other studies that reported poor care seeking for acute illness91 and vaccination126 in families of low birth weight infants. The study found no evidence that other socioeconomic and demographic factors influence families care seeking behaviour for circumcision related morbidity and acute illness in the population. These findings contrast with that of other earlier studies from
Ghana39,123,126,127 and Bangladesh128 that reported higher care seeking for infants with acute illness by richer families, and mothers with higher educational levels. This indicates that health care seeking behaviour for acute illness may be improving over time in rural Ghana.
171
Overall, the study found that more families sought care for a circumcision related morbidity from formal (68.6%) than informal (13.4%) healthcare providers. Unsurprisingly, most families of infants circumcised by a formal provider also sought care from a formal provider for complications. However, fewer families of infants circumcised by an informal provider sought care from an informal provider (45%) than a formal provider (55%). These findings differ from another published study from Kenya73 which reported families of male infants who were circumcised by an informal provider were two fold less likely to seek care for a circumcision related morbidity from a formal healthcare provider73. This study was not able to assess the type of care provider for acute illness in the Neovita study as these data were not collected in the Neovita study.
Generally, healthcare seeking for acute illness in the Neovita study population was very high (88.1%). The proportion of families seeking health care in this study was higher than data reported by the Ghana Demographic and Health
Survey in 2008, close to the time of conducting the Neovita study (acute respiratory infection 51.0%, fever 51.0%, and diarrhoea 35.2%)90. Care seeking for acute illness in rural populations has also been reported to be around 61% in
Kenya122, 47% in Nigeria59 and 27% in Vietnam129. The high healthcare seeking patterns for acute illnesses seen in this study population may be due to recent communications and promotion of care seeking by the local government district
172
health management teams. However, that these communications do not appear to have influenced care seeking for circumcision related morbidities to date.
9.1.4 Effectiveness of health service provider education and/or training in infant male circumcision on morbidity or mortality outcomes in any setting
The systematic review was not able to provide clear evidence of the effect of health service provider education and/or training in infant male circumcision on all-cause morbidity or mortality outcomes or the secondary outcomes of knowledge improvement, skill improvement, health service provider satisfaction, post-surgery presentations to clinic or hospital and family satisfaction. Only three prospective cohort studies with historical controls were located and the risk of bias for each study was rated as “very high”. The overall quality of the body of evidence for knowledge and skill improvements were also considered to be “very low” using the GRADE criteria. Data on the duration and intensity of training required was limited. No studies were located that assessed informal service providers, low income country settings or simple packages of training.
The review found only three studies that were specific to infant males148-150 which examined the effect of training on the skills and knowledge of medical doctors and were inconclusive. In addition, since the initial search I have become aware of one other small observational study which assessed the effect
173
of competence-based training in infant male circumcision using the Mogen clamp on knowledge scores in non-physicians151. In this Ugandan study, clinical officers (n=10) and midwives/nurses (n=10) received five days of training including; didactic learning (i.e good clinical practice, informed consent, screening, anatomy of the penis, contraindications of circumcision, haemostasis and suturing, postoperative care, hygiene, complications and resuscitation), interactive sessions, role plays, and case scenarios. Each of the trainees also observed five circumcisions. This study found significant improvements in knowledge 20 days after the training program. Twenty percent of clinical officers and 10% of midwives/nurses had knowledge scores over 80% before the training. Post-training, all (100%) clinical officers and 60% of midwives/nurses had knowledge scores of 80% or more (p-value=0.043). These trainees then circumcised ten neonates (aged 1-28 days) under direct supervision and were assessed for competency using a standardised checklist. Statistically significant improvements in competence were observed between the first three procedures
(baseline) and the fourth procedure (endline) (clinical officers; +5.87 points, 95%
CI 3.14–8.60 and midwives/nurses +7.9 points, 95% CI 5.89– 9.86, p-value=
0.043). However, pain scores were not significantly different between the two groups (78.9% for midwives/nurses versus 77.5% clinical officers, p- value=0.887). All (100%) trainees were reported to achieve full competency as measured on the checklist by the seventh procedure. No long term follow-up of skill or knowledge retention was conducted. An overall complication rate of
174
3.5% was reported within 24 hours post circumcision151. This study was carefully conducted and had good short term follow-up. However, it only involved 20 participants and had no long term follow-up so overall it is able to contribute little to the overall body of evidence to date.
There have also been six studies69,152-156 that have compared the effect of different types of devices in infant male circumcision. However, none involved informal providers. All studies were based in a hospital setting and the circumcisions were performed by medical doctors. In Zimbabwe, Mavhu et al 2015152, implemented an RCT which compared the safety of the AccuCirc clamp (n=100) versus the Mogen clamp (n=50) in 150 male infants aged 6-54 days. Training was provided to four doctors using the AccuCirc and Mogen clamp devices.
The training included; didactic learning, use of infant penile anatomic model, practical hands on training, written examination and practical assessments.
Trainees were required to score 100% on written exams and prove competency by performing five circumcisions with each device (data on pre and post training were not reported). This study reported no significant differences in complication rates between the two groups152. Other RCTs from Botswana69,
Zambia153, Iran154 and the USA155,156 also compared different infant male circumcision devices with little effect on the incidence of complications.
Infant male circumcision is commonly performed by informal providers in low income countries2,4,5,21,157. These providers often have no formal training in
175
circumcision but learn the trade through apprenticeship from another experienced informal provider. These providers are often highly valued by families for cultural and religious reasons. The most important part of this review was to assess the effectiveness of simple training packages such as the
WHO training program1 to improve the skills of these informal or traditional providers. However, this review found no studies in this important subgroup.
The systematic review was also unable to find any studies that assessed the effectiveness of training for informal health service providers in adult male circumcision. In contrast, the effectiveness of other types of surgical and procedural education and training to non-physicians in both rural and urban areas in low income countries using, hands on training and simulation techniques is well known. For example, Ellard et al 2015158, reported that hands on training of non-physicians improved skills in health outcomes and reduce maternal and perinatal morbidity and mortality. In this study, the health centres in 14 districts in Central and Northern Malawi were randomised to receive a non-physician training package (which included clinical skills training in obstetrics, neonatal care and leadership)-(intervention arm) or no training package (control group). A total of 54 non-physicians were trained in the 8 intervention districts from 2011 to 2013. The study evaluation was an interrupted time series evaluation that investigated differences within the control and intervention group at four time points before the intervention and
176
seven time points after the intervention at three month intervals. This training resulted in trend improvements in maternal mortality ratio in the intervention arm over time (effect -38.11, SE 50.30, CI -157.06 to 80.84, p-value=0.47).
Neonatal mortality also reduced by 31% (from 5.8 to 4.0 per 1000 live births) in the intervention group compared to an increase of 2% (from 9.6 to 9.8 per 1000 live births) in the control group over the 3 year period158. No statistical analyses were performed to compare the change in outcomes between the intervention and control groups in this study.
Simulation-based training is also considered to be effective and cost-effective way to provide education and training to non-physicians in trauma injury management. A Myanmar trauma training program trained community health workers to evaluate and manage trauma injuries from 2000 to 2013159. A post- training evaluation of 26 of these health workers in 2013 indicated that 75%
(n=18) felt competent to evaluate and managed trauma injuries, 96% (n=25) self- reported that the training was relevant to their work, all (100% n=26) felt time spent for training was valuable and 96% (n=25) self-reported that simulation models improved skills in trauma injury management159.
177
9.2 Strengths and limitations of this research
9.2.1 Strengths
The main strength of this research is the use of community level and population based data on the burden of circumcision related morbidity and family level care seeking practices in rural Africa. The study findings are therefore representative of the rural population and may have external validity in similar settings particularly in low income countries.
The other strengths include the large sample size of the circumcision study (3,141 male infants) and the additional analyses performed using the 22,955 infants enrolled in the population based Neovita trial. This particular study is presented in chapter six (thesis paper three). The large sample size provided adequate power to generate estimates for the various analyses and to assess the effect of important explanatory variables and outcomes.
The quality of data collection on the burden and key determinants of morbidity associated with infant male circumcision, and care seeking was high. Visual aids were used to clarify responses, and all infants were examined by trained fieldworkers. The fieldworkers also reviewed infant health records and checked consistency with a family’s report. This improved the quality of data collected.
This study was also strengthened by the standardised questionnaires, and well- trained, experienced fieldworkers who had been involved in many large trials in
178
the study area. The fieldworkers and families were blinded to the study hypothesis therefore reporting bias was low. In addition, for the analyses of the study outcomes (circumcision related morbidity and care seeking) and the exposure of interest (circumcision), I controlled for a wide range of individual, household and community level confounders. This approach also minimised reporting bias in this study. Furthermore, in the systematic review, the ROBINS-
I assessment tool, was used to assess bias in the included studies and this tool is a well validated tool for observational studies
9.2.2 Limitations
The research has some limitations. The study would have benefitted from the availability of an in depth qualitative data on the perspectives and experiences of families and circumcision providers on infant male circumcision practices, choice of circumcision provider, circumcision related morbidities and care seeking behaviour. This study found a triple burden of high circumcision prevalence, high morbidity load and poor care seeking. However, I was unable to explore among families and circumcision providers, why infant males were more likely to be circumcised, factors which influence choice of circumcision provider, reasons for the high morbidities and poor care seeking practices. The lack of in depth qualitative data on the perspectives and experiences of families and providers limits the ability to identify interventions to reduce morbidity and improve care seeking.
179
The study designs presented in this thesis are observational and cross-sectional and do not provide proof of causation. Observational and cross-sectional studies are prone to differential misclassification and may report greater associations between potentially concerning healthcare practices (e.g. poor handwashing and lack of cleaning of the penile area and instruments) and maternal reports of morbidity. Furthermore, poor recognition of illness for example, circumcision related morbidity and the classification of the morbidity as not meant to be managed in the hospital or clinic39,118,119, poor recognition of severity of illness for example, excess incision, excess bleeding, inadequate skin removal, infection39,120-122 and poor understanding of the different types of service providers for example, description of traditional informal providers as formal health care providers73 have been reported among families in low income countries. This may have resulted in under reporting and misclassification of circumcision morbidities, circumcision providers and care seeking behaviour.
Although this study had a large sample size with adequate power for the various analyses, the small sample size in the circumcision cohort due to low level of healthcare seeking limited the ability to assess the association between key sociodemographic factors and care seeking for circumcision related morbidities. It was also not possible to statistically compare care seeking between the circumcision cohort and the larger Neovita cohort.
180
The Neovita study included in the analyses of care seeking behaviour in the study area lacked data on the type of provider care was sought from. This also limits the understanding of type of care provider sought for acute illnesses in the population.
In addition, 22% of babies were not able to be visited within a 12 week period after birth. Anecdotal information from the study area indicated that these families needed to travel more for employment and they were of lower socio economic status and educational levels. The omission of these infants reduces the generalisability of this study a little but is unlikely to have introduced any systematic bias.
9.3 Future research
There are several important unanswered questions arising out of this research. A qualitative component of this research would have been useful to explore the perspectives and experiences of families and circumcision providers (formal and informal). In particular, qualitative data to explore: i) families knowledge and experiences on infant male circumcision, circumcision related morbidities, the factors which influences choice of circumcision providers and care seeking behaviour; ii) circumcision provider knowledge and attitudes towards infant male circumcision, methods used including hygiene practices, how they ensure safety of the procedure, training received and after care procedures. These
181
valuable perspectives will further inform strategies to: i) improve training of circumcision providers (formal and informal) in hygiene practice; investment in training will improve service provider skills, reduce morbidity risk and improve safety for young infants in rural Africa; ii) support families to access high quality infant male circumcision; iii) increase community awareness of the burden and consequences of circumcision and care seeking for circumcision related morbidity.
This research reported a high risk of morbidities following infant male circumcision with key determinants being poor hygiene related practices amongst circumcision providers including not washing hands, the use of uncleaned instruments and not cleaning the penile area. These findings are consistent with published reports from other low income countries160,161. Reasons are likely to include lack of hand washing facilities and soap and poor initial training161,162. However, additional data are needed to aid the development of targeted strategies to improve hygiene practices amongst circumcision providers.
This study was not able to ascertain the level of pain experienced by the male infants. There are published reports of infants having similar responses to pain as found in adults163-165 however, specific data from rural African populations are
182
lacking. Data are needed to promote pain free circumcision in both formal and informal health care systems in rural Africa.
As indicated from systematic review of published literature, more high quality randomised controlled trials or observational studies are needed to understand the best types of infant male circumcision education and training in low and middle income countries. The studies should: i) include traditional and informal circumcision providers who perform most circumcisions in low income countries; ii) test low cost simple interventions that can be scaled up across low income countries including the WHO package; iii) include novel methods to improve education and training to reduce complications among informal non- medical providers; iv) provide information about the duration required for informal non-medical providers to become competent in infant male circumcision; v) provide data on follow-up care. Strategies for scale-up in different geographic settings, especially low income countries, must also be identified.
9.4 Recommendations and implications for policy and practice
This thesis presents the first community level population based studies of the burden and determinants of morbidity associated with infant male circumcision and care seeking in rural Africa within formal and informal health care systems.
The study found evidence of the influence of both socio-economic status and
183
geographical access to health facilities on choice of service provider. The study further provided information about care seeking patterns of families once circumcision morbidities were observed, as well as the effectiveness of health service provider circumcision education and training on reducing morbidity and improving health care seeking. Based on the findings outlined in this thesis and through the Kintampo Health Research, I will work together with the District and
Regional Health Management Teams, the Ghanaian Ministry of Health and other non-governmental organisations to implement the following recommendations to reduce morbidity and improve care seeking in rural Ghana;
i) Public educational campaigns on Ghana’s free maternal care policy
and health insurance scheme. This could help not only to dispel some
of the misconceptions surrounding the free maternal care policy and
health insurance scheme, particularly the practice of illegal fee paying
for circumcision at hospitals and clinics, but also to enhance families
understanding of the free maternal care policy and insurance scheme.
ii) Training of circumcision providers (formal and informal) in hygiene
related care practices including hand washing, cleaning of
circumcision instruments and the penile area. Investment in training
will improve provider skills, reduce the risk of circumcision related
morbidities and improve patient’s safety.
iii) Programmes to increase community awareness of the burden and
consequences of circumcision related morbidities, and encourage
184
families to seek care. This requires the involvement of community
leaders and other stakeholders and could include behaviour change
communication strategies.
iv) High quality randomised controlled trials and observational studies of
simple training packages to improve education and training of
circumcision providers, especially informal non-medical providers in
low income countries are needed.
v) Additional qualitative research to explore the perspectives and
experiences of families and circumcision providers (formal and
informal). The qualitative research should explore: i) families
knowledge and experiences on infant male circumcision, morbidities,
choice of provider and care seeking; ii) service providers knowledge
and attitudes towards infant male circumcision, methods used
including hygiene practices, how they ensure safety of the procedure,
training received and after care procedures.
9.5 Conclusions
Infant male circumcision is almost universal in Ghana and many parts of sub-
Saharan Africa. The procedure is associated with high morbidity risk and sometimes mortality and poor care seeking. Interventions to reduce morbidity and improve care seeking are needed. The development and rigorous testing of simple training packages to improve education and training of circumcision
185
providers, especially informal non-medical providers in low income countries are required to improve hygiene practices and reduce circumcision related morbidities. Qualitative research is also needed to explore the experiences of families and circumcision providers on circumcision related morbidities and care seeking patterns. As a result of this thesis, Kintampo Health Research Centre and the District and Regional Health Management Teams will work together to develop an advocacy and training package for presentation and discussions to the Ghanaian Ministry of Health.
References
1. World Health Organization, JHPIEGO. Manual for early infant male circumcision under local anaesthesia. Geneva, Switzerland: World Health Organization, 2010: http://apps.who.int/iris/bitstream/10665/44478/1/9789241500753_eng.pdf.
2. Joint United Nations Programme on HIV/AIDS (UNAIDS), World Health
Organization. Neonatal and child male circumcision: a global review. Geneva,
Switzerland: Joint United Nations Programme on HIV/AIDS (UNAIDS), 2010: http://www.who.int/hiv/pub/malecircumcision/neonatal_child__MC_UNAIDS.pdf.
3. Bailey RC, Plummer FA, Moses S. Male circumcision and HIV prevention: current knowledge and future research directions. Lancet Infect Dis 2001; 1(4): 223-31.
4. Wilcken A, Keil T, Dick B. Traditional male circumcision in eastern and southern
Africa: a systematic review of prevalence and complications. Bull World Health Organ
2010; 88(12): 907-14.
186
5. Gyan T, McAuley K, Strobel NA, et al. Determinants of morbidity associated with infant male circumcision: community-level population-based study in rural Ghana. Trop
Med Int Health 2017; 22(3): 312-22.
6. Task Force On Circumcision. Male circumcision. Pediatrics 2012; 130(3): e756-e85.
7. Morris BJ, Bailis SA, Castellsague X, Wiswell TE, Halperin DT. RACP's policy statement on infant male circumcision is ill‐conceived. Aust N Z J Public Health 2006;
30(1): 16-22.
8. Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, Puren A.
Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 Trial. PLoS Med 2005; 2(11): e298.
9. Bailey RC, Moses S, Parker CB, et al. Male circumcision for HIV prevention in young men in Kisumu, Kenya: a randomised controlled trial. Lancet 2007; 369(9562):
643-56.
10. Gray RH, Kigozi G, Serwadda D, et al. Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet 2007; 369(9562): 657-66.
11. Grund JM, Bryant TS, Jackson I, et al. Association between male circumcision and women's biomedical health outcomes: a systematic review. Lancet Glob Health 2017;
5(11): e1113-e22.
12. Morris BJ, Kennedy SE, Wodak AD, et al. Early infant male circumcision:
Systematic review, risk-benefit analysis, and progress in policy. World J Clin Pediatr
2017; 6(1): 89-102.
13. Bennett J, Breen C, Traverso H, Agha SB, Macia J, Boring J. Circumcision and neonatal tetanus: disclosure of risk and its reduction by topical antibiotics. Int J
Epidemiol 1999; 28(2): 263-6.
14. Weiss HA, Larke N, Halperin D, Schenker I. Complications of circumcision in male neonates, infants and children: a systematic review. BMC Urol 2010; 10: 2.
187
15. Wirth KE, Semo BW, Spees LP, Ntsuape C, Barnhart S, Ledikwe JH. A prospective cohort study of safety and patient satisfaction of voluntary medical male circumcision in
Botswana. PLoS One 2017; 12(11): e0185904.
16. Brook I. Infectious Complications of Circumcision and Their Prevention. Eur Urol
Focus 2016; 2(4): 453-9.
17. Edler G, Axelsson I, Barker GM, Lie S, Naumburg E. Serious complications in male infant circumcisions in Scandinavia indicate that this always be performed as a hospital‐ based procedure. Acta Paediatr 2016; 105(7): 842-50.
18. Auvert B, Taljaard D, Rech D, et al. Association of the ANRS-12126 male circumcision project with HIV levels among men in a South African township: evaluation of effectiveness using cross-sectional surveys. PLoS Med 2013; 10(9): e1001509.
19. World Health Organisation, Joint United Nations Programme on HIV/AIDS.
WHO/UNAIDS technical consultation on male circumcision and HIV prevention: research implications for policy and programming. Conclusions and recommendations.
Geneva, Switzerland: World Health Organisation, 2007.
20. UNAIDS/WHO/SACEMA Expert Group on Modelling the Impact and Cost of
Male Circumcision for HIV Prevention. Male circumcision for HIV prevention in high
HIV prevalence settings: what can mathematical modelling contribute to informed decision making? PLoS Med 2009 6(9).
21. Adu-Gyamfi S, Adjei P.O-W. Twentieth Century Wanzams among the Asante
People of Ghana: A Historical Study of the Facts on Male Circumcision. OJPM 2014;
4(9): 730-9.
22. Myers RA, Omorodion FI, Isenalumhe AE, Akenzua GI. Circumcision: its nature and practice among some ethnic groups in southern Nigeria. Soc Sci Med 1985; 21(5):
581-8.
188
23. Ahmed A, Mbibi NH, Dawam D, Kalayi GD. Complications of traditional male circumcision. Ann Trop Paediatr 1999; 19(1): 113-7.
24. Millar AJ, Roberts D. Complications of circumcision using the Plastibell device. S
Afr Med J 1987; 72(6): 438-9.
25. Manji KP. Circumcision of the young infant in a developing country using the
Plastibell. Ann Trop Paediatr 2000; 20(2): 101-4.
26. Okafor P, Orakwe J, Osuigwe A, Chianakwana G. Experience with immediate post- partum circumcision. Niger Med Pract 2005; 47(1): 9-11.
27. Okeke LI, Asinobi AA, Ikuerowo OS. Epidemiology of complications of male circumcision in Ibadan, Nigeria. BMC Urol 2006; 6: 21.
28. Lazarus J, Alexander A, Rode H. Circumcision complications associated with the
Plastibell device. S Afr Med J 2007; 97(3): 192-3.
29. Banieghbal B. Optimal time for neonatal circumcision: an observation-based study.
J Pediatr Urol 2009; 5(5): 359-62.
30. World Health Organization. Report on Neonatal, Infant and Pre-pubertal Male
Circumcision Practices in Traditional and Clinical Settings in Ghana. Final Report.
Accra, Ghana: World Health Organisation, 2008.
31. Gabrysch S, Campbell OM. Still too far to walk: literature review of the determinants of delivery service use. BMC Pregnancy Childbirth 2009; 9(1): 34.
32. Berhan Y, Berhan A. A meta-analysis of socio-demographic factors predicting birth in health facility. Ethiop J Health Sci 2014; 24 Suppl: 81-92.
33. Moyer CA, Mustafa A. Drivers and deterrents of facility delivery in sub-Saharan
Africa: a systematic review. Reprod Health 2013; 10: 40.
34. Baral YR, Lyons K, Skinner J, van Teijlingen ER. Determinants of skilled birth attendants for delivery in Nepal. Kathmandu Univ Med J (KUMJ) 2010; 8(31): 325-32.
189
35. Idris SH, Gwarzo UMD, Shehu AU. Determinants of place of delivery among women in a semi-urban settlement in Zaria, northern Nigeria. Ann Afr Med 2007; 5(2):
68-72.
36. Abor A.P, Abekah‐Nkrumah G, Sakyi K, Adjasi K.D.C. The socio‐economic determinants of maternal health care utilization in Ghana. Int J Soc Econ 2011; 38(7):
628-48.
37. Okyere E, Tawiah-Agyemang C, Manu A, Deganus S, Kirkwood B, Hill Z.
Newborn care: the effect of a traditional illness, asram, in Ghana. Ann Trop Paediatr
2010; 30(4): 321-8.
38. Hill Z, Manu A, Tawiah-Agyemang C, et al. How did formative research inform the development of a home-based neonatal care intervention in rural Ghana? J Perinatol
2008; 28: S38-S45.
39. Hill Z, Kendall C, Arthur P, Kirkwood B, Adjei E. Recognizing childhood illnesses and their traditional explanations: exploring options for care‐seeking interventions in the context of the IMCI strategy in rural Ghana. Trop Med Int Health 2003; 8(7): 668-
76.
40. Ansah Manu A, ten Asbroek A, Soremekun S, et al. Evaluating the implementation of community volunteer assessment and referral of sick babies: lessons learned from the
Ghana Newhints home visits cluster randomized controlled trial. Health Policy Plan
2014; 29 Suppl 2: ii114-27.
41. Rutherford ME, Mulholland K, Hill PC. How access to health care relates to under‐ five mortality in sub‐Saharan Africa: systematic review. Trop Med Int Health 2010;
15(5): 508-19.
42. Okawa S, Ansah EK, Nanishi K, et al. High Incidence of Neonatal Danger Signs and Its Implications for Postnatal Care in Ghana: A Cross-Sectional Study. PLoS One
2015; 10(6): e0130712.
190
43. Bazzano AN, Kirkwood BR, Tawiah-Agyemang C, Owusu-Agyei S, Adongo PB.
Beyond symptom recognition: care-seeking for ill newborns in rural Ghana. Trop Med
Int Health 2008; 13(1): 123-8.
44. Beiersmann C, Sanou A, Wladarsch E, De Allegri M, Kouyate B, Muller O. Malaria in rural Burkina Faso: local illness concepts, patterns of traditional treatment and influence on health-seeking behaviour. Malar J 2007; 8(6): 106.
45. Ahorlu CK, Koram KA, Ahorlu C, de Savigny D, Weiss MG. Socio-cultural determinants of treatment delay for childhood malaria in southern Ghana. Trop Med Int
Health 2006; 11(7): 1022-31.
46. Osifo OD, Oriaifo IA. Circumcision mishaps in Nigerian children. Ann Afr Med
2009; 8(4): 266-70.
47. Demaria J, Abdulla A, Pemberton J, Raees A, Braga LH. Are physicians performing neonatal circumcisions well-trained? Can Urol Assoc J 2013; 7(7-8): 260-4.
48. Bailey RC, Egesah O. Assessment of clinical and traditional male circumcision services in Bungoma district, Kenya: Complication Rates and Operational Needs.
Special report. Washington, DC: USAID, PSI AIDSMark 2006: http://pdf.usaid.gov/pdf_docs/Pnadg558.pdf.
49. Perera CL, Bridgewater FH, Thavaneswaran P, Maddern GJ. Safety and efficacy of nontherapeutic male circumcision: a systematic review. Ann Fam Med 2010; 8(1): 64-
72.
50. Jagannath VA, Fedorowicz Z, Sud V, Verma AK, Hajebrahimi S. Routine neonatal circumcision for the prevention of urinary tract infections in infancy. Cochrane
Database Syst Rev 2012; 11: CD009129.
51. Offord J. Restrictions concerning Circumcision under the Romans: Proceedings of the Royal Society of Medicine London UK: SAGE Publishing; 1913.
191
52. World Health Organization, Joint United Nations Programme on HIV/AIDS. Male circumcision: global trends and determinants of prevalence, safety, and acceptability.
Geneva, Switzerland: World Health Organization, Joint United Nations Programme on
HIV/AIDS., 2007: http://www.unaids.org/sites/default/files/media_asset/jc1360_male_circumcision_en_0. pdf.
53. Smith GE. The Migrations of Early Culture: A Study of the Significance of the
Geographical Distribution of the Practice of Mummification as Evidence of the
Migration of People and the Spread of Certain Customs and Beliefs. Manchester:
Manchester University Press; 1929.
54. Burger R, Guthrie TH. Why circumcision? Pediatrics 1974; 54(3): 362-4.
55. Alanis MC, Lucidi RS. Neonatal circumcision: a review of the world’s oldest and most controversial operation. Obstet Gynecol Surv 2004; 59(5): 379-95.
56. Mann TW. The book of the Torah: The narrative integrity of the Pentateuch.
Westminster, UK: Westminster John Knox Press; 1988.
57. Brian JM. What motivates parents to get their baby boy circumcised? 2016: http://www.circinfo.net/what_motivates_parents_to_get_their_baby_boy_circumcised.h tml. Accessed 24th May 2017.
58. Ghana Statistical Service (GSS), Ghana Health Service (GHS), ICF International.
Ghana Demographic and Health Survey 2014. Rockville, Maryland, USA: Ghana
Statistical Service (GSS) 2015.
59. National Population Commission (NPC) [Nigeria] and ICF Macro. 2009. Nigeria
Demographic and Health Survey 2008. Abuja, Nigeria: National Population
Commission and ICF Macro.
192
60. McQuoid-Mason DJ. Is the mass circumcision drive in KwaZulu-Natal involving neonates and children less than 16 years of age legal? What should doctors do? S Afr
Med J 2013; 103(5): 283-4.
61. Na AF, Tanny SP, Hutson JM. Circumcision: Is it worth it for 21st‐century
Australian boys? J Paediatr Child Health 2015; 51(6): 580-3.
62. Royal Australasian College of Physicians, Paediatrics & Child Health Division.
Circumcision of infant males. Sydney, Australia: Royal Australasian College of
Physicians, 2010: Available http://www.racp.edu.au/page/ policy-and advocacy/paediatrics-and child-health.
63. Richters J, Smith AM, de Visser RO, Grulich AE, Rissel CE. Circumcision in
Australia: prevalence and effects on sexual health. Int J STD AIDS 2006; 17(8): 547-54.
64. Simpson E, Carstensen J, Murphy P. Neonatal circumcision: new recommendations
& implications for practice. Mo Med 2014; 111(3): 222-30.
65. Alagoa PJ, Gbobo I. Complications of Male Infant Circumcision in a Semi-Urban
Niger Delta Town. Int J Trop Dis 2013; 3(3): 217-23.
66. El Katsha S, Labeeb S, Watts S, Younis A. Informal health providers and the transmission of hepatitis C virus: pilot study in two Egyptian villages. East Mediterr
Health J 2006; 12(6): 758-67.
67. Atikeler MK, Geçit I, Yüzgeç V, Yalçin O. Complications of circumcision performed within and outside the hospital. Int Urol Nephrol 2005; 39(1): 97-9.
68. Bailey RC, Nyaboke I, Otieno FO. What device would be best for early infant male circumcision in east and southern Africa? Provider experiences and opinions with three different devices in Kenya. PLoS One 2017; 14(12): 2.
69. Plank RM, Ndubuka NO, Wirth KE, et al. A randomized trial of Mogen clamp versus Plastibell for neonatal male circumcision in Botswana. J Acquir Immune Defic
Syndr 2013; 62(5): e131-7.
193
70. Kankaka EN, Murungi T, Kigozi G, et al. Randomised trial of early infant circumcision performed by clinical officers and registered nurse midwives using the
Mogen clamp in Rakai, Uganda. BJU Int 2017; 119(1): 164-70.
71. Young MR, Bailey RC, Odoyo-June E, et al. Safety of over Twelve Hundred Infant
Male Circumcisions Using the Mogen Clamp in Kenya. PLoS One 2012; 7(10): e47395.
72. Osuigwe A, Ikechebelu J, Okafor P. Circumcision-Related Complications in the
Male: Experience amongst the Igbo's of Southeast Nigeria. Afr J Urology 2005; 10(4):
246-51.
73. Bailey RC, Egesah O, Rosenberg S. Male circumcision for HIV prevention: a prospective study of complications in clinical and traditional settings in Bungoma,
Kenya. Bull World Health Organ 2008; 86(9): 669-77.
74. Horowitz M, Gershbein AB. Gomco circumcision: When is it safe? J Pediatr Surg
2001; 36(7): 1047-9.
75. Morris Brian J, Waskett Jake H, Banerjee Joya, et al. A'snip'in time: what is the best age to circumcise? BMC Pediatr 2012; 12(1): 20.
76. Moosa FA, Khan FW, Rao MH. Comparison of complications of circumcision by'Plastibell device technique'in male neonates and infants. J Pak Med Assoc 2010;
60(8): 664.
77. El Bcheraoui C, Zhang X, Cooper CS, Rose CE, Kilmarx PH, Chen RT. Rates of adverse events associated with male circumcision in U.S. medical settings, 2001 to
2010. JAMA Pediatr 2014; 168(7): 625-34.
78. Ekenze SO, Ezomike UO. Complications of neonatal circumcision requiring surgical intervention in a developing country. J Trop Pediatr 2013; 59(4): 292-7.
79. Amzat J, Razum O. Medical Pluralism: Traditional and Modern Health Care.
Medical Sociology in Africa. Springer International Publishing, Switzerland 2014: 207-
40.
194
80. Mboera, L.E.G., Massaga, J.J., Senkoro, K.P., Kilima, S.P., Mayala, B.K., Msovela,
J. & Shayo, E.H. (2009) Challenges and Opportunities for the involvement of
Traditional Practitioners in scaling up of Safe Male Circumcision in the Context of HIV
Prevention in Tanzania. National Institute for Medical Research, Dar es Salaam,
Tanzania. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.442.4119&rep=rep1&type=p df.
81. World Health Organization. Surgical care at the district hospital. Malta: World
Health Organization, 2003: http://www.who.int/surgery/publications/en/SCDH.pdf.
82. Deuchler S, Wagner C, Singh P, et al. Clinical Efficacy of Simulated Vitreoretinal
Surgery to Prepare Surgeons for the Upcoming Intervention in the Operating Room.
PLoS One 2016; 11(3): e0150690.
83. Opiyo N, English M. In-service training for health professionals to improve care of the seriously ill newborn or child in low and middle-income countries (Review).
Cochrane Database Syst Rev 2010; 14(4): CD007071.
84. Nagendran M, Gurusamy KS, Aggarwal R, Loizidou M, Davidson BR. Virtual reality training for surgical trainees in laparoscopic surgery. Cochrane Database Syst
Rev 2013; 27(8): CD006575.
85. Kirkwood BR, Manu A, Tawiah-Agyemang C, et al. NEWHINTS cluster randomised trial to evaluate the impact on neonatal mortality in rural Ghana of routine home visits to provide a package of essential newborn care interventions in the third trimester of pregnancy and the first week of life: trial protocol. Trials 2010; 11: 58.
86. Edmond KM, Newton S, Shannon C, et al. Effect of early neonatal vitamin A supplementation on mortality during infancy in Ghana (Neovita): a randomised, double- blind, placebo-controlled trial. Lancet 2015; 385(9975): 1315-23.
195
87. Kirkwood BR, Hurt L, Amenga-Etego S, et al. Effect of vitamin A supplementation in women of reproductive age on maternal survival in Ghana (ObaapaVitA): a cluster- randomised, placebo-controlled trial. Lancet 2010; 375(9726): 1640-9.
88. Ghana Health Service. Annual Report. Brong Ahafo Region. Sunyani: Ghana Health
Service, 2010.
89. UNAIDS, Ghana AIDS Commission. Ghana country AIDS progress report.
Reporting period January 2010-December 2011 Geneva, Switzerland: UNAIDS, 2012.
Available: http://www.unaids.org/en/dataanalysis/knowyourresponse/countryprogressreports/2012 countries/ce_GH_Narrative_Report[1].pdf
90. Ghana Statistical Service (GSS), Ghana Health Service (GHS), and ICF Macro.
2009. Ghana Demographic and Health Survey 2008. Accra, Ghana: GSS, GHS, and ICF
Macro.
91. O'Leary M, Edmond K, Floyd S, Newton S, Thomas G, Thomas SL. A cohort study of low birth weight and health outcomes in the first year of life, Ghana. Bull World
Health Organ 2017; 95(8): 574-83.
92. Kirkwood BR, Sterne JAC. Essential medical statistics second edition.
Massachesetts USA: Blackwell pulishing; 2006.
93. Ben Chaim J, Livne PM, Binyamini J, Hardak B, Ben-Meir D, Mor Y.
Complications of circumcision in Israel: a one year multicenter survey. Isr Med Assoc J
2005; 7(6): 368-70.
94. Lerman SE, Liao JC. Neonatal circumcision. Pediatr Clin North Am 2001; 48(6):
1539-57.
95. Yegane RA, Kheirollahi AR, Salehi NA, Bashashati M, Khoshdel JA, Ahmadi M.
Late complications of circumcision in Iran. Pediatr Surg Int 2006; 22(5): 442-5.
196
96. Sawires SR, Dworkin SL, Fiamma A, Peacock D, Szekeres G, Coates TJ. Male circumcision and HIV/AIDS: challenges and opportunities. Lancet 2007; 369(9562):
708-13.
97. Gyan T, McAuley K, Strobel N, Newton S, Owusu-Agyei S, Edmond K. The influence of socioeconomic factors on choice of infant male circumcision provider in rural Ghana; a community level population based study. BMC Pediatr 2017; 17(1): 185.
98. World Bank. Socio-Economic Differences in Health, Nutrition, and Population within developing countries Ghana,Washington, D.C: World Bank, 2007.
99. Ghana Statistical service. Ghana Living Standards Survey Round 6 (GLSSR6):
Poverty Profile in Ghana 2005-2013, GLSSR6 Poverty Report 2014. Available: http://catalog.ihsn.org/index.php/catalog/5350. Accessed 25th August 2016. Accra:
Ghana Statistical service, 2014.
100. Douglas M, Hongoro C. The Consideration of Socioeconomic Determinants in
Prevention of Traditional Male Circumcision Deaths and Complications. Am J Mens
Health 2016; 18(1): 1-11.
101. Weiss HA, Plummer ML, Changalucha J, et al. Circumcision among adolescent boys in rural northwestern Tanzania. Trop Med Int Health 2008; 13(8): 1054-61.
102. Ministry of Health Ghana. Implementation guidelines for financing free maternal health care. Accra: Ministry of Health Ghana, 2008.
103. National Health Insurance Authority. Annual Report Accra, Ghana: National
Health Insurance Authority, 2009.
104. Ghana Statistical Service. Ghana Multiple Indicator Cluster Survey with an
Enhanced Malaria Module and Biomarker, Final Report. Accra, Ghana: Ghana
Statistical Service, 2011.
105. National Health Insurance Authority. Annual Report. Accra, Ghana: National
Health Insurance Authority, 2011. 197
106. National Health Insurance Authority. Annual Report. Accra, Ghana: National
Health Insurance Authority, 2013.
107. Barimah KB, Mensah J. Ghana's National Health Insurance Scheme: insights from members, administrators and health care providers. J Health Care Poor Underserved
2013; 24(3): 1378-90.
108. Ghana News Agency. Ghana’s free maternal healthcare policy not working –
Research. 2016. https://www.ghanabusinessnews.com/2016/04/23/ghanas-free- maternal-healthcare-policy-not-working-research/ (accessed 1st April 2017).
109. Ghana News Agency. Free maternal health policy. Is it really working? 2012. http://vibeghana.com/2012/01/18/free-maternal-health-policy-is-it-really-working/
(accessed 1st April 2017).
110. Kabeer N. Can the MDGS provide a pathway to social justice? The challenge of intersecting inequalities. Institute of Development Studies and MDG Achievement
Fund, 2010. www.ids.ac.uk/files/dmfile/MDGreportwebsiteu2WC.pdf. Accessed 6th
September 2016.
111. Mumtaz Z, Salway S, Bhatti A, McIntyre L. Addressing invisibility, inferiority, and powerlessness to achieve gains in maternal health for ultra-poor women. Lancet
2014; 383(9922): 1095-7.
112. Ganle JK, Parker M, Fitzpatrick R, Otupiri E. Inequities in accessibility to and utilisation of maternal health services in Ghana after user-fee exemption: a descriptive study. Int J Equity Health 2014; 13: 89.
113. Houweling TA, Ronsmans C, Campbell OM, Kunst AE. Huge poor-rich inequalities in maternity care: an international comparative study of maternity and child care in developing countries. Bull World Health Organ 2007; 85(10): 745-54.
114. Rizvi SA, Naqvi SA, Hussain M, Hasan AS. Religious circumcision: a Muslim view. BJU Int 1999; 83 Suppl 1: 13-6.
198
115. Gyan T, McAuley K, O'Leary M, Strobel N.A, Edmond K.M. Healthcare seeking patterns of families of infants with circumcision-related morbidities from two population-based cohort studies in Ghana. BMJ Open 2017; 7(9): 1-12.
116. Gove S. Integrated management of childhood illness by outpatient health workers: technical basis and overview. The WHO Working Group on Guidelines for Integrated
Management of the Sick Child. Bull World Health Organ 1997; 75 Suppl 1: 7-24.
117. Lambrechts T, Bryce J, Orinda V. Integrated management of childhood illness: a summary of first experiences. Bull World Health Organ 1999; 77(7): 582-94.
118. Geldsetzer P, Williams TC, Kirolos A, et al. The recognition of and care seeking behaviour for childhood illness in developing countries: a systematic review. PLoS One
2014; 9(4): e93427.
119. Nabiwemba EL, Atuyambe L, Criel B, Kolsteren P, Orach CG. Recognition and home care of low birth weight neonates: a qualitative study of knowledge, beliefs and practices of mothers in Iganga-Mayuge Health and Demographic Surveillance Site,
Uganda. BMC Public Health 2014; 14: 546.
120. Paudel M. Male circumcision: care practices and attitudes in a Muslim community of western Nepal. IJPH 2011; 8(1).
121. Muula AS, Prozesky HW, Mataya RH, Ikechebelu JI. Prevalence of complications of male circumcision in Anglophone Africa: a systematic review. BMC Urol 2007; 7: 4.
122. Taffa N, Chepngeno G. Determinants of health care seeking for childhood illnesses in Nairobi slums. Trop Med Int Health 2005; 10(3): 240-5.
123. Manu A, Hill Z, Ten Asbroek AH, et al. Increasing access to care for sick newborns: evidence from the Ghana Newhints cluster-randomised controlled trial. BMJ
Open 2016; 6(6): e008107.
124. Colvin CJ, Smith HJ, Swartz A, et al. Understanding careseeking for child illness in sub-Saharan Africa: a systematic review and conceptual framework based on
199
qualitative research of household recognition and response to child diarrhoea, pneumonia and malaria. Soc Sci Med 2013; 86: 66-78.
125. Mohan P, Iyengar SD, Martines J, Cousens S, Sen K. Impact of counselling on careseeking behaviour in families with sick children: cluster randomised trial in rural
India. BMJ 2004; 329(7460): 266.
126. O'Leary M, Edmond K, Floyd S, et al. Neonatal vaccination of low birthweight infants in Ghana. Arch Dis Child 2017; 102(2): 145-51.
127. Kirkwood BR, Manu A, ten Asbroek AH, et al. Effect of the Newhints home-visits intervention on neonatal mortality rate and care practices in Ghana: a cluster randomised controlled trial. Lancet 2013; 381(9884): 2184-92.
128. Shah R, Mullany LC, Darmstadt GL, et al. Determinants and pattern of care seeking for preterm newborns in a rural Bangladeshi cohort. BMC Health Serv Res
2014; 14: 417.
129. Lee HY, Van Huy N, Choi S. Determinants of early childhood morbidity and proper treatment responses in Vietnam: results from the Multiple Indicator Cluster
Surveys, 2000-2011. Glob Health Action 2016; 9: 29304.
130. Gyan T, Strobel N, McAuley K, et al. Health service provider education and/or training in infant male circumcision to improve short-and long-term morbidity outcomes: protocol for systematic review. Syst Rev 2016; 5(1): 41.
131. Shamseer L, Moher D, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation.
BMJ 2015; 350: g7647.
132. World Health Organisation. International Clinical Trials Registry Platform
(ICTRP): Available: www.who.int/ictrp/en/ (2015). Accessed 20 Nov 2015. World
Health Organisation; 2015.
200
133. United Kingdom Clinical Research Collaboration. UK Clinical Research Network:
Portfolio Database (UKCRN study portfolio). Available: http://www.ukcrc.org/research-infrastructure/clinical-research-networks/uk-clinical- research-network-ukcrn. Accessed 20 November 2015. United Kingdom Clinical
Research Collaboration; 2015.
134. Higgins J, Green S. (editors). Cochrane Handbook for Systematic Reviews of
Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011.
Available from http://handbook.cochrane.org/.
135. Green S. Cochrane handbook for systematic reviews of interventions version 5.1. 0
[updated March 2011]. The Cochrane Collaboration 2011.
136. Guyatt G, Oxman AD, Akl EA, et al. GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol 2011; 64(4): 383-
94.
137. Gyan T, Strobel N, McAuley K, Shannon C, et al. Health service provider education and/or training in infant male circumcision to improve short and long term morbidity outcomes: A systematic review. Submitted to PLoS One, 27th September
2017.
138. World Health Organization (WHO), Joint United Nations Programme on
HIV/AIDS (UNAIDS). WHO and UNAIDS announce recommendations from expert consultation on male circumcision for HIV prevention. Paris/Geneva: World Health
Organization, 2007: http://apps.who.int/iris/bitstream/10665/73924/1/PR_10_eng.pdf.
139. Ericsson KA. The influence of experience and deliberate practice on the development of superior expert performance. The Cambridge handbook of expertise and expert performance 2006; 38: 685-705.
140. Martens FM, op 't Root JM. Practical medical education in general practice. Med
Educ 1992; 26(3): 213-7.
201
141. Ringsted C, Schroeder TV, Henriksen J, et al. Medical students' experience in practical skills is far from stakeholders' expectations. Med Teach 2001; 23(4): 412-6.
142. Kneebone R. Simulation in surgical training: educational issues and practical implications. Med Educ 2003; 37(3): 267-77.
143. McGaghie WC, Issenberg SB, Petrusa ER, Scalese RJ. A critical review of simulation-based medical education research: 2003-2009. Med Educ 2010; 44(1): 50-
63.
144. Nousiainen M, Brydges R, Backstein D, Dubrowski A. Comparison of expert instruction and computer-based video training in teaching fundamental surgical skills to medical students. Surgery 2008; 143(4): 539-44.
145. Sterne JA, Hernan MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 2016; 355: i4919.
146. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008; 336(7650): 924-6.
147. The Cochrane Collaboration. Effective Practice and Organisation of Care (EPOC).
EPOC worksheets for preparing a Summary of Findings (SoF) table using GRADE.
EPOC Resources for review authors. Oslo: Norwegian Knowledge Centre for the
Health Services, 2013. Available at: http://epoc.cochrane.org/epoc-specific-resources- review-authors. Accessed 15th April 2017.
148. Chandran L, Latorre P. Neonatal circumcisions performed by pediatric residents: implementation of a training program. Ambul Pediatr 2002; 2(6): 470-4.
149. Roca P, Alvarado C, Stausmire J.M, Farooq S, Hill-Engstler E.A. Effectiveness of a simulated training model for procedural skill demonstration in neonatal circumcision.
Simul Healthc 2012; 7(6): 362-73.
150. Smith A, Maizels M, Korets R, et al. A novel method of teaching surgical techniques to residents--computerized enhanced visual learning (CEVL) with
202
simulation to certify mastery of training: a model using newborn clamp circumcision. J
Pediatr Urol 2013; 9(6 Pt B): 1210-3.
151. Kankaka EN, Kigozi G, Kayiwa D, et al. Efficacy of knowledge and competence- based training of non-physicians in the provision of early infant male circumcision using the Mogen clamp in Rakai, Uganda. BJU Int 2017; 119(4): 631-7.
152. Mavhu W, Larke N, Hatzold K, et al. Implementation and Operational Research:
A Randomized Noninferiority Trial of AccuCirc Device Versus Mogen Clamp for Early
Infant Male Circumcision in Zimbabwe. J Acquir Immune Defic Syndr 2015; 69(5): e156-63.
153. Bowa K, Li MS, Mugisa B, et al. A controlled trial of three methods for neonatal circumcision in Lusaka, Zambia. J Acquir Immune Defic Syndr 2013; 62(1): e1-6.
154. Fanaei SA, Musavi NS, Mehrvarz SH. Plastibell and conventional circumcision in infants: a randomized clinical trial. J Rafsanjan Univ Med Sci 2003; 2(2): 68-73.
155. Kurtis PS, DeSilva HN, Bernstein BA, Malakh L, Schechter NL. A comparison of the Mogen and Gomco clamps in combination with dorsal penile nerve block in minimizing the pain of neonatal circumcision. Pediatrics 1999; 103(2): E23.
156. Taeusch HW, Martinez AM, Partridge JC, Sniderman S, Armstrong-Wells J,
Fuentes-Afflick E. Pain during Mogen or PlastiBell circumcision. J Perinatol 2002;
22(3): 214-8.
157. World Health Organisation. Traditional male circumcision among young people: a public health perspective in the context of HIV prevention. Geneva, Switzerland: World
Health Organisation, 2009.
158. Ellard D.R, Chimwaza W, Davies D, et al. Up-skilling associate clinicians in
Malawi in emergency obstetric, neonatal care and clinical leadership: the ETATMBA cluster randomised controlled trial. BMJ Glob Health 2016; 1(1): 1-13.
203
159. Washington C.H, Tyler F.J, Davis J, et al. Trauma training course: innovative teaching models and methods for training health workers in active conflict zones of
Eastern Myanmar. Int J Emerg Med 2014; 7(1): 46-12.
160. Pittet D, Boyce JM. Hand hygiene and patient care: pursuing the Semmelweis legacy. Lancet Infect Dis 2001; 1: 9-20.
161. World Health Organization, Patient Safety: A World Alliance for Safer Health
Care. Save Lives, Clean Your Hands: Hand Hygiene Technical Reference Manual.
Geneva, Switzerland: World Health Organization, 2009: http://apps.who.int/iris/bitstream/10665/44196/1/9789241598606_eng.pdf.
162. World Health Organization. Guidelines on hand hygiene in health care: first global patient safety challenge Clean Care is Safer Care. Geneva, Switzerland: World Health
Organization, 2009: http://apps.who.int/iris/bitstream/10665/44102/1/9789241597906_eng.pdf.
163. Paix BR, Peterson SE. Circumcision of neonates and children without appropriate anaesthesia is unacceptable practice. Anaesth Intensive Care 2012; 40(3): 511.
164. Kaya Z, Suren M, Arici S, Karaman S, Tapar H, Erdemir F. Prospective, randomized, double-blinded comparison of the effects of caudally administered levobupivacaine 0.25% and bupivacaine 0.25% on pain and motor block in children undergoing circumcision surgery. Eur Rev Med Pharmacol Sci 2012; 16(14): 2014-20.
165. Bellieni CV, Alagna MG, Buonocore G. Analgesia for infants' circumcision. Ital J
Pediatr 2013; 39: 38.
204
Appendix A. Circumcision data collection form
Appendix B. PRISMA-P (Preferred Reporting Items for Systematic review and
Meta-Analysis Protocols). Protocol checklist 2015.
Appendix C. MEDLINE search strategy
Appendix D. PRISMA-P (Preferred Reporting Items for Systematic review and
Meta-Analysis). Final publication checklist 2017.
Appendix E. MEDLINE search strategy used for the systematic review
Appendix F. PhD publications
205