WHO South-East Asia Journal of Public Health ISSN 2224-3151 Volume 1, Issue 4, October—December 2012, 359-486

Volume 1, Issue 4, October–December 2012, 359-486 Volume 1, Issue 4, October–December 2012, 359-486 Inside Editorial Research brief Success of tuberculosis and HIV collaboration Evidence of HPV subtypes linked with cervical Iyanthi Abeyewickreme, Neelamanie cancer in Punchihewa, Amaya Maw-Naing 359 Chop L Bhusal, Sulochana Manandhar, Meeta Singh, Aarati Shah, Perspective Sushharma Neupane, Dibesh Karmacharya, Janani Suraksha Yojana: the conditional cash Kate Cuschieri, Heather Cubie, transfer scheme to reduce maternal mortality in Duncan C Gilbert, Sameer M Dixit 441 – a need for reassessment Rajesh Kumar Rai, Policy and practice Prashant Kumar Singh 362 Institutionalizing district level infant death review: an experience from southern India Pandemic influenza preparedness planning: Sanjeev Upadhyaya, Sudeep Shetty, lessons from Cambodia Selva S Kumar, Amol Dongre, Anthony C Huszar, Tom Drake, Teng Srey, Pradeep Deshmukh 446 Sok Touch, Richard J Coker 369 Health systems responsiveness and its correlates: Review evidence from family planning service A review of Japanese encephalitis in Uttar provisionin Sri Lanka WHO Pradesh, India W L S P Perera, Lillian Mwanri, Roop Kumari and Pyare L Joshi 374 Rohini de A Seneviratne, Thushara Fernando 457 Original research A study on delay in treatment of kala-azar Challenges faced by skilled birth attendants in patients in Bangladesh providing antenatal and intrapartum care in South-East Asia Syed M Arif, Ariful Basher, Mohammad R selected rural areas of Myanmar Rahman, Mohammad A Faiz, 396 Kyaw Oo, Le Le Win,Saw Saw,Myo Myo Mon, WHO South-East Asia Journal of Public Health Yin Thet Nu Oo, Thae Maung Maung, Factors associated with high prevalence of Su Latt Tun Myint, Theingi Myint 467 pulmonary tuberculosis in HIV-infected people Public health classic Journal of visiting for assessment of eligibility for highly active antiretroviral therapy in Kathmandu, Epidemiologic investigation of excess maternal Nepal and neonatal deaths and evidence-based Bishnu R Tiwari, Surendra Karki, Prakash low-cost public health interventions – Ghimire, Bimala Sharma, Sarala Malla 404 Ignaz Semmelweis: the etiology, concept and prophylaxis of child bed fever Public Health Compliance of off-premise alcohol retailers with S D Gupta, MD 477 the minimum purchase age law Recent WHO Publications 485 Areekul Puangsuwan, Kannapon Phakdeesettakun, Thaksaphon Thamarangsi, Surasak Chaiyasong 412

External quality assessment in blood group serology in the World Health Organization South-East Asia Region Patravee Soisangwan 423

Prevalence of group A genotype human rotavirus among children with diarrhoea in Nepal, 2009–2011 Jeevan B Sherchand, W William Schluter, Jatan B Sherchan, Sarmila Tandukar, Jyoti R Dhakwa, Ganga R Choudhary, Chandeshwar Mahaseth 432

ISSN 2224315-1

World Health House Indraprastha Estate, Mahatma Gandhi Marg, New Delhi-110002, India www.searo.who.int/who-seajph www.searo.who.int/ publications/journals/seajph WHO South-East Asia Journal of Public Health WHO South-East Asia Journal of Public Health Guidelines for contributors The WHO South-East Asia Journal of Public Health (WHO-SEAJPH) is a Volume 1, Issue 4, October–December 2012, 359-486 peer-reviewed, indexed (IMSEAR), open access quarterly publication Original research articles on public health, primary health Text Advisory Board care, epidemiology, health administration, health systems, of the World Health Organization, Regional Office for South-East Asia. The main text of the paper should begin with title and an health economics, health promotion, public health nutrition, Samlee Plianbangchang abstract which should be structured under the following The Journal provides an avenue to scientists for publication of original communicable and non-communicable diseases, maternal and (Chair) headings: title, background, methods, results, conclusions, research work so as to facilitate use of research for public health child health, occupational and environmental health, social Richard Horton and 5-8 key words for indexing (http://www.ncbi.nlm.nih. and preventive medicine are invited which have potential action. Hooman Momen gov/entrez/query.fcgi?db=mesh). It should not exceed 250 to promote public health in the South-East Asia Region. We David Sanders words. Follow the IMRD format, i.e. Introduction, Methods, also publish editorial commentaries, perspectives, state of Editorial Process Fran Baum Results and Discussion. Manuscript must be written in a the art reviews, research briefs, report from the field, policy manner, which is easy to understand, and should be restricted All manuscripts are initially screened by editorial panel for scope, relevance and Lalit M Nath and practice, letter to the editor and book reviews etc. scientific quality. Suitable manuscripts are sent for peer review anonymously. Chitr Sitthi-amorn to the topic being presented. The discussion should end in Submitted articles are peer reviewed anonymously conclusion statement. Each table and figure should be on Recommendations of at least two reviewers are considered by the editorial Editorial Board and confidentially. The manuscript should be original which separate page at the end of the manuscript. panel for making a decision on a manuscript. Accepted manuscripts are edited has neither been published nor is under consideration for for language, style, length etc. before publication. Authors must seek permission Poonam K Singh publication in any substantial form in any other publication. Acknowledgement from the copyright holders for use of copyright material in their manuscripts. (Chair) All published manuscripts are the property of the World Place it as the last element of the text before references. Submissions are promptly acknowledged and a decision to publish is usually Sattar Yoosuf Health Organization. communicated within three months. Nyoman Kumara Rai Written permission of person/agency acknowledged should Aditya P Dash Submit article to the Chief Editor, as an electronic copy be provided. through CD ROM or email [email protected], World Disclaimer Sangay Thinley Health Organization, Regional Office for South-East Asia, Athula Kahandaliyanage References The designations employed and the presentation of the material in this Indraprastha Estate, Mahatma Gandhi Road, New Delhi, 110 Quazi Munirul Islam publication do not imply the expression of any opinion whatsoever on the part 002, India. Tel. 91-11-23309309, Fax. 91-11-23370197. Only original references should be included. Signed permission of the World Health Organization concerning the legal status of any country, Nata Menabde is required for use of data from persons cited in personal territory, city or area or its authorities, or concerning the delimitation of its Maureen E Birmingham communication. ANSI standard style adapted by the National Manuscript preparation frontiers or boundaries. Dotted lines on maps represent approximate border Mahmudur Rahman Library of Medicine should be followed (http://www.nlm.nih. lines for which there may not yet be full agreement. Chencho Dorji Type the article in double space (including references) gov/bsd/uniform_requirements.html). References should be Hasbullah Thabrany using 12 point font with at least one inch margin on all numbered and listed consecutively in the order in which they The mention of specific companies or of certain manufacturers’ products Nay Soe Maung sides. Uniform Requirements for Manuscripts submitted to are first cited in the text and should be identified in the text, does not imply that they are endorsed or recommended by the World Health Biomedical Journals should be consulted before submission tables and legends by Arabic numerals as superscripts. The Suniti Acharya Organization in preference to others of a similar nature that are not mentioned. of the manuscript (http://www.icmje.org ). All articles should full list of references placed at the end of the main text of the Rohini de A Seneviratne All reasonable precautions have been taken by the World Health Organization mention how human and animal ethical aspects of the study paper should include surnames and initials of all authors up to verify the information contained in this publication. However, the published Phitaya Charupoonphol have been addressed. When reporting experiment on human to six (if more than six, only the first six are given followed material is being distributed without warranty of any kind, either expressed or Mariam Cleason subjects indicate whether the procedures followed were in by et al.); the title of the paper, the journal title abbreviation implied. The responsibility for the interpretation and use of the material lies Kenneth Earhart accordance with the Helsinki Declaration (http://www.wma. according to the style of Index Medicus, year of publication; with the reader. In no event shall the World Health Organization be liable for Jacques Jeugmans net). volume number; first and last page number. Reference of books should give the surnames and initials of the authors, the damages arising from its use. The named authors alone are responsible for Guidelines for preparation of manuscripts for various sections Editorial Team book title, place of publication, publisher and year; citation of the views expressed in this publication. of WHO SEAJPH are available on website (http:www.searo. chapters in a multiple author book should include the surname Aditya P Dash who.int/who-seajph) and initials of the authors, chapter title, and names and (Chief Editor) Permissions initials of editors, book title, place of publication, publisher Charles Raby © World Health Organization 2013. All rights reserved. and year and first and last page numbers of the chapter. Title page (Coordinator) For citing website references, mention author surname and Requests for permission to reproduce or translate WHO publications – whether Rajesh Bhatia This should contain the title of the manuscript (not more than initials, title of the article, the website address, and the date for sale or for non-commercial distribution should be addressed to Information (Associate Editor) 150 characters including spaces), the name of all authors on which website was accessed. Management and Dissemination Unit, World Health Organization, Regional Nyoman Kumara Rai (first name, middle initials, and surname), a short title (not Office for South-East Asia. We encourage other users to link to articles on our Sunil Senanayake more than 40 characters including spaces), name(s) of the Tables and figures web site, provided they do not frame our content; there is no need to seek our Sudhansh Malhotra institution(s) where the work has been carried out, and the address of the corresponding author including telephone, fax permission. Richard Brown The tables/figures must be self explanatory and must not and e-mail. One of the authors should be identified as the Nihal Abeysinghe duplicate information in the text. Each table and figure must corresponding author and guarantor of the paper who will have a title and should be numbered with Arabic numerals. Prakin Suchaxaya Editorial Office take responsibility of the article as a whole. Briefly mention Each table should be typed in double space on a separate WHO South-East Asia Journal of Public Health Roderico Ofrin the contribution of each author in a multi-author article. Also sheet of paper. Figures should be of the highest quality, i.e. World Health Organization, Regional Office for South-East Asia, Renu Garg mention conflict of interest, the source of support in the form glossy photographs or drawn by artist or prepared using Indraprastha Estate, Mahatma Gandhi Road, New Delhi 110 002, India. of grants, equipment, drugs etc. Word count of the abstract standard computer software. A descriptive legend must Tel. 91-11-23309309, Fax. 91-11-23370197 and main text, number of references, figures and table should accompany each figure which should define all abbreviations e-mail: [email protected] also be mentioned on the bottom of title page. used. All tables and figures should be cited in the text. Web site: www.searo.who.int/publications/journals/seajph WHO South-East Asia Journal of Public Health

Volume 1, Issue 4, October—December 2012, 359-486

Contents

Editorial Success of tuberculosis and HIV collaboration Iyanthi Abeyewickreme, Neelamanie Punchihewa, Amaya Maw-Naing 359 Perspective Janani Suraksha Yojana: the conditional cash transfer scheme to reduce maternal mortality in India – a need for reassessment Rajesh Kumar Rai, Prashant Kumar Singh 362

Pandemic influenza preparedness planning: lessons from Cambodia Anthony C Huszar, Tom Drake, Teng Srey, Sok Touch, Richard J Coker 369 Review A review of Japanese encephalitis in , India Roop Kumari, Pyare L Joshi 374 Original research A study on delay in treatment of kala-azar patients in Bangladesh Syed M Arif, Ariful Basher, Mohammad R Rahman, Mohammad A Faiz 396

Factors associated with high prevalence of pulmonary tuberculosis in HIV-infected people visiting for assessment of eligibility for highly active antiretroviral therapy in Kathmandu, Nepal Bishnu R Tiwari, Surendra Karki, Prakash Ghimire, Bimala Sharma, Sarala Malla 404

Compliance of off-premise alcohol retailers with the minimum purchase age law Areekul Puangsuwan, Kannapon Phakdeesettakun, Thaksaphon Thamarangsi, Surasak Chaiyasong 412

External quality assessment in blood group serology in the World Health Organization South-East Asia Region Patravee Soisangwan 423

Prevalence of group A genotype human rotavirus among children with diarrhoea in Nepal, 2009—2011 Jeevan B Sherchand, W William Schluter, Jatan B Sherchan, Sarmila Tandukar, Jyoti R Dhakwa, Ganga R Choudhary, Chandeshwar Mahaseth 432 Research brief Evidence of HPV subtypes linked with cervical cancer in Nepal Chop L Bhusal, Sulochana Manandhar, Meeta Singh, Aarati Shah, Sushharma Neupane, Dibesh Karmacharya, Kate Cuschieri, Heather Cubie, Duncan C Gilbert, Sameer M Dixit 441

i Policy and practice Institutionalizing district level infant death review: an experience from southern India Sanjeev Upadhyaya, Sudeep Shetty, Selva S Kumar, Amol Dongre, Pradeep Deshmukh 446

Health systems responsiveness and its correlates: evidence from family planning service provision in Sri Lanka W L S P Perera, Lillian Mwanri, Rohini de A Seneviratne, Thushara Fernando 457

Challenges faced by skilled birth attendants in providing antenatal and intrapartum care in selected rural areas of Myanmar Kyaw Oo, Le Le Win,Saw Saw,Myo Myo Mon, Yin Thet Nu Oo, Thae Maung Maung, Su Latt Tun Myint, Theingi Myint 467 Public health classic Epidemiologic investigation of excess maternal and neonatal deaths and evidence-based low-cost public health interventions – Ignaz Semmelweis: the etiology, concept and prophylaxis of child bed fever S D Gupta 477

Recent WHO publications 485

ii Editorial

Success of tuberculosis and HIV collaboration Iyanthi Abeyewickremea, Neelamanie Punchihewab, Amaya Maw-Naingca

Almost 30 years since it was first reported HIV-associated TB disease.3 The world had in 1983, the HIV epidemic remains a serious reached the Millennium Development Goal public health issue globally, and particularly in (MDG) targets of halting and reversing the TB countries of the WHO South-East Asia Region epidemic in many high prevalent states,3 but (SEAR). The overwhelming global efforts to HIV–TB coinfection is a challenge to reduce the combat HIV/AIDS over the past 10 years have global target of 50% deaths due to TB by 2015, led to fewer HIV epidemics. New infections The target is linked to MDG and endorsed by have declined by over 25% in 29 countries the stop TB partnership/UNAIDS in 2006 and worldwide, including four of the five high HIV recommitted by the Member States at the burden countries in SEAR (India, Myanmar, United Nations high level meeting on AIDS in Nepal and Thailand but not Indonesia) since 2011.1,3 While the HIV and TB epidemics are the epidemic peaked in 2001.1 declining in almost all high-burden countries, HIV-related TB continues to remain a serious More importantly, the survival of people challenge for the health sector response by living with HIV (PLHIV) and their ability national governments and threatens the to enjoy productive lives have improved survival of PLHIV.1,3 significantly over the past decade due to the scaled-up provision of life-saving antiretroviral Of the estimated 34 million people living therapy (ART). Moreover, ART is not only a with HIV globally, about one third are thought priority life-saving intervention; but it can also to have concomitant latent TB. In 2011, of the reduce the incidence of tuberculosis (TB) by over 8.7 million incident TB cases worldwide, up to 90%.1 1 million were among PLHIV, The WHO South- East Asia Region accounts for nearly 15% of HIV is the strongest risk factor for the global burden of new HIV/TB infections. developing active TB disease and PLHIV are 20 Five countries of the Region are among the to 30 times more likely to develop active TB 22 highest TB burden countries in the world than HIV negative people.2 In some countries (Bangladesh, India, Indonesia, Myanmar and up to 82% of people with TB have HIV.1 TB Thailand). Among these, four countries (India, is a major killer among PLHIV and accounts Indonesia, Myanmar and Thailand) have a for more than 25% deaths among this group high burden of TB and HIV. While Nepal is worldwide.2 In 2011, 430 000 out of 1.7 million not among the countries with the highest TB AIDS-related deaths (25%) were caused by

aWorld Health Organization, Regional Office for South-East Asia, New Delhi, India. Correspondence to Amaya Maw-Naing (e-mail: [email protected])

WHO South-East Asia Journal of Public Health 2012;1(4):359-361 359 Success of tuberculosis and HIV collaboration Iyanthi Abeyewickreme et al.

burden, the number of PLHIV is high1,3 which and counselling to anyone presenting with poses a potential threat of HIV–TB coinfections signs and symptoms of TB and people with in the country. confirmed TB. ART also must be made available for all PLHIV who have active TB In 2011, nearly 5 million people were living irrespective of their immune status or CD4 with TB in SEAR, and approximately 140 000 cell count. Simultaneously, HIV programmes of the estimated 3.5 million people who newly should refer PLHIV attending treatment and developed TB were HIV-positive.3 care facilities for early TB assessment, so that Collaborative TB–HIV activities between active TB can be detected early and treatment national TB and HIV programmes are vital initiated without delay. to prevent, diagnose and treat TB among While the number of TB patients identified PLHIV, and HIV among individuals with TB, as HIV-positive and enrolled on ART has grown to achieve the goal of reducing TB-related steadily in the Region as well as globally,3 deaths. World Health Organization (WHO) necessary testing systems appear to be still has provided clear policy recommendations inadequate In 2011 worldwide 40% of TB on interventions needed to achieve this patients had a documented HIV test result collectively known as collaborative TB–HIV while on average only 32% of TB patients activities.4,5 These include: (i) establishing knew their HIV status in SEAR.3 This means mechanisms for collaboration between HIV and that two out of three TB patients were not TB programmes (joint planning, coordinating aware of their HIV status in the Region. bodies, surveillance, and monitoring and Lowest HIV testing rates were recorded in evaluation); (ii) reducing the HIV burden Myanmar and Indonesia while India and among TB patients through HIV testing Thailand recorded high rates.3 Perhaps, stigma and counselling, provision of ART and co- associated with HIV prevents people with TB trimoxasole preventive therapy (CPT) to TB from attending HIV care services for an HIV patients living with HIV, and HIV prevention, test, and clinicians in TB clinics appear to be care and support services for TB patients and still reluctant to offer HIV testing (known as (iii) reducing the burden of TB among PLHIV provider-initiated counselling and testing) to with the Three I’s for HIV/TB :– intensified TB patients under their care.1,3 About 59% of case-finding; isoniazid preventive therapy; the TB patients diagnosed with TB–HIV co- and infection control in health-care and infection were provided with lifesaving ART in congregate settings.3–5 The progress made SEAR compared with 48% globally though the towards achieving these goals in countries of proportion varied widely across the Region, the Region has been variable.1,3 Myanmar had the highest with 80% ART Highly active antiretroviral therapy coverage and Indonesia lowest 42% among (HAART) is recognized as a great achievement the high TB/HIV burden countries in SEAR.3 To in public health as it has not only prevented reach the Global Plan’s target of providing ART deaths among PLHIV, but has also reduced to all TB patients known to be living with HIV common opportunistic infections, especially by 2015 the coverage of ART for HIV-positive Mycobacterium tuberculosis, in people living TB patients needs to be improved. with HIV(6) For ART to have the maximum In 2011 about 400 000 PLHIV in the benefit, it is important to detect people living Region were assessed for TB during their last with TB–HIV co-infection early National TB visit, of whom the largest number was in India programmes should provide HIV testing compared to 2.3 million globally,3 India and

360 WHO South-East Asia Journal of Public Health 2012;1(4):359-361 Iyanthi Abeyewickreme et al. Success of tuberculosis and HIV collaboration

Myanmar have assessed and recorded the TB status of almost all HIV patients enrolled for References HIV care at the last visit.1,3 1. Joint United Nations Programme on HIV/AIDS. Global report: UNAIDs report on the global AIDS epidemic HIV-associated TB deaths can be further 2012 Geneva: UNAIDS, 2012 -http://www.unaids. reduced by offering co-trimoxazole preventive org/en/media/unaids/contentassets/documents/ therapy to TB patients living with HIV. Over 89% epidemiology/2012/gr2012/20121120_UNAIDS_ patients in this group in SEAR have benefitted Global_Report_2012_en.pdf - accessed 17 April 2013. from this measure, since many national programmes have offered CPT and intensified 2. World Health Organization. WHO global plan to stop TB 2011-2015: transforming the fight towards TB screening of PLHIV at HIV treatment elimination of tuberculosis. Geneva, 2010. and care service points.3 It is heartening to note that TB–HIV collaborative activities are 3. World Health Organization. Global tuberculosis report 2012. Geneva: WHO, 2012. progressing steadily in the Region. India and Thailand have fully integrated nationwide 4. World Health Organization. Interim policy on collaborative TB/HIV activities. Geneva: WHO, 2004. implementation of this collaboration, other Document No. WHO/HTM/TB/2004.330; WHO/HTM/ high-burden countries like Indonesia and HIV/2004.1. Myanmar, as well as low prevalence countries 5. World Health Organization. WHO policy on collaborative such as Bangladesh, Maldives and Sri Lanka TB/HIV activities: guidelines for national programmes have started scaling-up services. However, and other stakeholders. Geneva: WHO, 2012. Nepal is yet to institute strong TB–HIV Document No. WHO/ M/TB/2012.1 collaborative programmes. 6. World Health Organization. Antiretroviral therapy The global target of halving TB–HIV deaths for HIV infection in adults and adolescents: recommendations for a public health approach. 2010 by 2015 can be realized only if services are revision. Geneva: WHO, 2010 - http://whqlibdoc. scaled up through concerted and joint efforts who.int/publications/2010/9789241599764_eng. of National AIDS and TB programme to detect pdf - accessed 17 April 2013. HIV–TB coinfected people early and ensure them treatment.2

WHO South-East Asia Journal of Public Health 2012;1(4):359-361 361 Perspective

Janani Suraksha Yojana: the conditional cash transfer scheme to reduce maternal mortality in India – a need for reassessment Rajesh Kumar Raia, Prashant Kumar Singhb

Alongside endorsing Millennium Development Goal 5 in 2000, India launched its National Population Policy in 2000 and the National Health Policy in 2002. However, these have failed thus far to reduce the maternal mortality ratio (MMR) by the targeted 5.5% per annum. Under the banner of the National Rural Health Mission, the launched a national conditional cash transfer (CCT) scheme in 2005 called Janani Suraksha Yojana (JSY), aimed to encourage women to give birth in health facilities which, in turn, should reduce maternal deaths. Poor prenatal care in general, and postnatal care in particular, could be considered the causes of the high number of maternal deaths in India (the highest in the world). Undoubtedly, institutional delivery in India has increased and MMR has reduced over time as a result of socioeconomic development coupled with advancement in health care including improved women’s education, awareness and availability of health services. However, in the light of its performance, we argue that the JSY scheme was not well enough designed to be considered as an effective pathway to reduce MMR. We propose that the service-based CCT is not the solution to avoid/reduce maternal deaths and that policy-makers and programme managers should reconsider the ‘package’ of continuum of care and maternal health services to ensure that they start from adolescence and the pre-pregnancy period, and extend to delivery, postnatal and continued maternal health care.

Key words: Menternal Mortality ratio, continuum of care, cash incentive.

Background The high incidence of maternal deaths, 2000, endorsed by leaders from 190 countries, especially in developing countries, has been of re-emphasized the importance of improving growing concern to programme managers and maternal health in MDG5, and set a target of a policy-makers. To reduce maternal deaths, 75% reduction in the maternal mortality ratio global leaders promised to extend every (MMR) from 1990 by 2015.1 India is a signatory possible effort in a series of protocols, starting to the MDGs and managed to reduce the MMR from the Safe Motherhood Initiative in 1987, by 66%, that is, from an estimated 600 to 200 followed by the International Conference on maternal deaths per 100 000 live births, which Population and Development in 1994. The is an average annual decrease of 5.2% during Millennium Development Goals (MDGs) in the period 1990–2010.2 Despite this turning

aTata Institute of Social Sciences, Mumbai, India. bInternational Institute for Population Sciences, Mumbai, India. Correspondence to Rajesh Kumar Rai (e-mail: [email protected])

362 WHO South-East Asia Journal of Public Health 2012;1(4):362-368 Rajesh Kumar Rai et al. The conditional cash transfer scheme to reduce maternal mortality in India – a need for reassessment

point, the MMR is still incongruously high and if they had a government-issued below-the- fails to meet MDG5. poverty-line card or were from a Scheduled (low) Caste or Tribe.4 Against this background, India launched its National Population we argue that the purpose of JSY was not well Policy (NPP) in 2000 and the National Health designed and that it cannot be considered Policy (NHP) in 2002. The NPP set ambitious as an effective pathway to reduce maternal goals to be achieved by 2010, including deaths. We have proposed an alternative way reducing the MMR to 100 per 100 000 live forward to reduce maternal mortality. births. The NHP reiterated its commitment to achieving this goal and to increase institutional deliveries to 80%.3 The failure of the NPP and Causes of maternal the NHP to reduce the MMR pointed to the mortality: where JSY stands need for health policies and programmes to Among the direct causes of maternal deaths be given a new direction, and this lead to the in India, haemorrhage is the leading cause establishment of the National Rural Health (38%), followed by sepsis (11%), hypertensive Mission (NRHM) in 2005. Under the broad disorders (5%), obstructed labour (5%), ambit of NRHM, the Government of India abortion (8%), and other conditions (33%).6 launched a broad conditional cash transfer The conceptual framework of delays scheme called Janani Suraksha Yojana (JSY) proposed by Thaddeus and Maine focuses in April 2005, to encourage women of low on socioeconomic/cultural factors (women’s socioeconomic status to give birth in health status in household and society, educational facilities. This, in turn, was intended to help and economic status of women, etc.), access reduce the MMR. to facilities (distance, transportation, etc.) The scheme was guided by the previous and availability of quality of care (staff and National Maternity Benefit Scheme,4 under equipment in health centre) as the crucial which a provision was made for the payment factors behind maternal morbidity and of Rs 500 (US$ 9.41) per pregnancy to mortality.7 However, this model was debated women from poor households for prenatal because of its oversimplification of each level and postnatal maternity care up to the of care.8 first two live births.5 In the 10 focus states Haemorrhage remains one of the top (Assam, , Chhattisgarh, Jammu and three preventable obstetrics-related causes Kashmir, Jharkhand, Madhya Pradesh, Orissa, of maternal mortality worldwide, with most Rajasthan, Uttarakhand, and Uttar Pradesh), deaths occurring within 24–48 hours of the JSY promised cash incentives to a woman delivery.9 Indeed, in 2010 the World Health if the delivery was conducted in a government Organization estimated that 77% of maternal or accredited private health facility.4 According deaths in developing countries occurred within to JSY’s guidelines, after delivery in one of two days of delivery.10 The latest nationally these facilities, the eligible woman would representative District Level Household receive Rs 600 (US$ 11.30) in urban areas and Facility Survey 2007–2008 (DLHS-3)11 and Rs 700 (US$ 13.18) in rural areas.4 The revealed that in India, where most maternal cash incentive was higher in the 10 focus deaths occur due to postpartum haemorrhage states – Rs 1000 (US$ 18.83) in urban areas (PPH),6 around 48% of women aged 15–49 and Rs 1400 (US$ 26.36) in rural areas. In the years reported receiving postnatal care non-focus states, women were eligible for the within 48 hours of delivery. Maternal death cash benefit for their first two livebirths, and

WHO South-East Asia Journal of Public Health 2012;1(4):362-368 363 The conditional cash transfer scheme to reduce Rajesh Kumar Rai et al. maternal mortality in India – a need for reassessment

due to PPH is preventable through risk factor should be given equal importance to avoid identification, rapid diagnosis and timely maternal deaths. According to 2010 estimates, management of postnatal care;9 therefore, only 19% women received full antenatal care poor performance in postnatal care in India (at least three visits for antenatal check-up, remains a daunting challenge for programme one tetanus toxoid injection and 100+ iron managers and policy-makers. folate acid tablets/syrup),11 indicating the possibility of poor demand or/and provision On the other hand, as per the estimates of antenatal care. In addition, 26% of women from DLHS-3, almost 50% of women aged aged 15–49 did not undergo any antenatal 15–49 years had at least three antenatal care care, and almost half (48%) of the eligible visits; the lowest performance was 22% in women did not receive safe delivery care Uttar Pradesh. Only 47% of women reported (either institutional delivery or home delivery that they underwent institutional delivery, and assisted by skilled person).11 Moreover, only 13% of eligible women had received JSY while 37% women experienced post-delivery assistance.11 complications, nearly 45% of them did not The experiment of introducing incentives seek any post-delivery treatment at national for physicians to perform delivery to reduce level. The reported post-delivery complications maternal mortality was deemed fruitless. For were nearly 50% in states such as Bihar, instance, the Government of Gujarat (western Jharkhand, and Uttar Pradesh.11 India) launched a public–private partnership scheme called ‘Chiranjeevi Yojana’ or “plan for The way forward a long life” to bridge the gap in the availability of quality maternal health-care services for The statistics documented above lead us to below-the-poverty-line families in rural areas believe that incentives attached to institutional by collaborating with private practitioners in delivery are insufficient to defer maternal small towns.12 The state government paid deaths. In the last few years, there has been private gynaecologists Rs 1795 (US$ 33.75) a gradual decline in the country’s MMR overall. per delivery – including Rs 200 (US$ 3.76) This has been the outcome of efforts to reduce to the patient for transportation costs to delays in seeking medical help, timely use of the place of delivery and Rs 50 (US$ 0.94) medical facilities and provision of adequate to the person accompanying the patient to care. Moreover, behavioural factors such as compensate for loss of wages.12 But, contrary improved women’s education, as well as mass to expectation, there was no an appreciable media exposure, awareness and use of health progress in the overall reduction of MMR in services, availability of, and access to health Gujarat in 2004–2006 (160/100 000 live facilities, are some of the key factors that have births)6 or 2007–2009 (148/100 000 live led to increased maternal health-care use. In births).13 In contrast, the reduction was more addition, it is clear that institutional delivery 4 impressive in Uttarakhand/Uttar Pradesh, and has increased due to JSY. However, what even in Bihar/Jharkhand, which are recognized should be prioritized to help attain MDG5? We as underdeveloped states. propose a two-fold approach.

The argument for this is that the promotion First, the priority areas of the maternal of institutional delivery through JSY is not health-care programme should be reoriented the only way to reduce maternal deaths; to postnatal care. The proposition of offering antenatal, delivery as well as postnatal care a cash incentive, especially within 48 hours

364 WHO South-East Asia Journal of Public Health 2012;1(4):362-368 Rajesh Kumar Rai et al. The conditional cash transfer scheme to reduce maternal mortality in India – a need for reassessment

of delivery, needs to be evaluated and the • expanded antenatal care (treatment effectiveness of the programme checked on and management of maternal diabetes, a pilot basis. Delivery in an institution must management of HIV, anti-platelet agents in be encouraged as well as supervision by high-risk pregnancies, antihypertensives qualified physicians as long as is necessary, for mild to moderate hypertension i.e. the programme should not be limited during pregnancy, magnesium sulphate to discharging women automatically from for treatment of pregnancy-induced the institution within 48 hours of delivery. hypertension/eclampsia, influenza and Clearly, this needs physical facilities and pneumococcal vaccination); human resources to enhance the currently • community-based interventions (emer- overburdened public health infrastructure. gency transport funds, cash transfers, Second, the focus should be on the supportive care during childbirth, etc.); continuum of maternal health care as well as on • childbirth care (basic obstetrics, emergency “packages”, which include integrated services obstetrics [lower segment caesarean for potential mothers from adolescence section, active induction of post-term and pre-pregnancy period to delivery and pregnancy], treatment of pre-term pre- immediate postnatal and maternal health labour rupture of membranes, training care.14 Such care is provided by families and traditional birth attendants in clean communities through outpatient services, delivery and referral); clinics and other health facilities. As far as • postnatal care (birth spacing/family the package is concerned, a recent review of planning, postnatal visits). ongoing research15 has proposed packages of care for delivery at the community, health- It is encouraging that the High Level centre or hospital level, where the package Expert Group Report on Universal Health of maternal care includes: Coverage for India, instituted by the Planning • general supportive care (cessation of Commission, Government of India, suggested smoking during pregnancy, prevention of that the ‘National Health Package’ cover intimate partner violence, prevention of essential health at primary, secondary, as well maternal drug abuse during pregnancy, as tertiary level care for all citizens of India 16 recognition and treatment of postnatal by 2022. Here, it is imperative to suggest depression, and family planning); a comprehensive “package” for maternal healthcare that follows a continuum of care maternal nutritional (supplementation of • to avert maternal deaths in India. The role of multiple micronutrients, balanced protein the NRHM in promoting such a comprehensive energy, periconceptual folic acid and iron- package cannot be underestimated. Since folic acid during pregnancy, and calcium its launch in 2005, it has provided the key supplementation in pregnant women with strategy to India’s current public health low/inadequate intake); system, has influenced the development of • improved quality of basic antenatal care (at new guidelines to strengthening prenatal, least four visits of focused antenatal care, natal and postnatal service delivery, and screening and management of sexually promoted the extended use of available health transmitted infections and prevention and infrastructure, particularly human resources, management of malaria in pregnancy); at facility and non-facility levels. As far as the facility level is concerned, mandatory

WHO South-East Asia Journal of Public Health 2012;1(4):362-368 365 The conditional cash transfer scheme to reduce Rajesh Kumar Rai et al. maternal mortality in India – a need for reassessment

counselling on the benefits of postnatal expanded its focus to include maternal and care could help in raising awareness on newborn health strategies, with newborns post-delivery health. However, non-facility serving as a bridge between child and initiatives under NRHM could reach eligible maternal health interventions and strategies. women through community health workers Since 2004, Ethiopia has developed an in- to provide appropriate counselling to women country coordination process, RMNCH survival and other household members, particularly strategies,21 and has incorporated these husbands, on the importance of postnatal care strategies into its major policy initiatives, like and on the adverse consequences of ignoring the Poverty Reduction Strategy Programme. postnatal health check-ups. Likewise in Mexico, a national health Although the continuum of care has recently insurance programme called Seguro Popular, been highlighted as central for reproductive, introduced in 2003, provides access to a maternal, newborn, and child health (RMNCH), package of comprehensive health services and a means to reduce the burden of maternal with financial protection for more than 50 and child deaths globally,17 the main barrier to million people previously excluded from increased coverage of the integrated package insurance.22 Since its inception, there is in most developing countries is inadequate evidence that Seguro Popular has improved operational management, especially at district access to health services and reduced the and community level.18 Since the concept prevalence of catastrophic and impoverishing of continuum of care in RMNCH is quite health expenditure, especially for the poor. recent, there is little evidence of its effective The reduction of maternal and child deaths, integration in more global programmes. and coverage of maternal and child health- care services, have been pragmatic during 23 Neighbour initiatives the past few years. These efforts can be adapted and translated for countries seeking Two recent initiatives from Ethiopia (Health to provide universal coverage against threats Service Extension Package) and Mexico to the health security of individuals and (Seguro Popular Program) could be helpful for populations. Learning from this experience – policy-makers and programme managers in its successes and its challenges – will not only India to initiate an integrated RMNCH approach. help to improve health conditions and financial In 2003, the Government of Ethiopia and its protection for all people in Mexico, but will partners analysed selected maternal and child contribute to the global movement towards health interventions and the country’s Health universal health coverage.24 Sector Development Programme.19 The results of the analysis indicated the need to shift from a facility-based model to an approach that Conclusions and extended access to basic health services to recommendations communities. In response, the Government Given our analysis of the institutional delivery of Ethiopia developed a new Health Service of cash through JSY, it is proposed to rethink Extension Package (HSEP), which posted two the cash incentive system to reduce maternal female health workers in every community, mortality and reconsider maternal health supported by health centres and several care as a continuum of care. According to thousand new medical officers.20 During the the budgetary provision in the Twelfth Five following year, the Child Survival Partnership Year Plan, the Government of India allocated

366 WHO South-East Asia Journal of Public Health 2012;1(4):362-368 Rajesh Kumar Rai et al. The conditional cash transfer scheme to reduce maternal mortality in India – a need for reassessment

nearly 2% of its Gross Domestic Product to 8. Bhutta ZA. Seeing the unseen: targeting neonatal healthcare. Some of this could be used to mortality in rural Vietnam. Global Health Action. design an effective package of maternal care. 2011; 4: 6360. Reduction in maternal mortality and associated 9. Haeri S, Dildy GA. Maternal mortality from issues should be high on the political agenda, hemorrhage. Seminar in Perinatalogy. 2012; 36(1): 48-55. especially of a developing country like India as a clearly rising power in the global economic 10. World Health Organization. WHO technical situation. consultation on postpartum and postnatal care. Geneva: WHO, 2010.

11. International Institute for Population Sciences. Acknowledgements District level household and facility survey, 2007- The authors are indebted to Dr Faujdar Ram, 08. Mumbai: International Institute for Population Sciences, 2010. Director & Senior Professor, International Institute for Population Sciences, Mumbai, 12. United Nations Children’s Fund. India: ‘Chiranjeevi India, for his valuable insights. Yojana’ (Plan for a long life): public-private partnership to reduce maternal deaths in Gujarat, India. UNICEF, 2009 - http://www.unicef.org/ devpro/46000_47108.html - 16 April 2013.

References 13. Government of India, Ministry of Home Affairs. 1. Paxton A, Wardlaw T. Are we making progress in Maternal mortality in India 2007-09: special bulletin. maternal mortality? The New England Journal of New Delhi, India: sample registration system, New Medicine. 2011; 364(21): 1990-1993. Delhi: Office of the Registrar General, 2011.

2. World Health Organization. Trends in maternal 14. The Partnership for Maternal, Newborn and Child mortality: 1990 to 2010: WHO, UNICEF, UNFPA, and Health. Analyzing commitments to advance the World Bank estimates. Geneva: WHO, 2012. the global strategy for women’s and children’s 3. Rai RK, Tulchinsky TH. Addressing the sluggish health: the PMNCH 2011 report. Geneva: WHO, progress in reducing maternal mortality in India. 2011 - http://www.who.int/pmnch/topics/part_ Asia Pacific Journal of Public Health. 2012; doi: publications/2012_pmnch_report_full_publicaton. 10.1177/1010539512436883. pdf - accessed 16 April 2013.

4. Lim SS, Dandona L, Hoisington JA, James SL, Hogan 15. Family Care International. A systematic review of MC. India’s Janani Suraksha Yojana, a conditional cash the interconnections between maternal & newborn transfer program to increase births in health facilities: health. New York: FCI, 2011 – http://www. an impact evaluation. The Lancet. 2010; 375(9730): familycareintl.org/en/resources/publications/94 2009–2023. - accessed 16 April 2013.

5. Devadasan N, Elias MA, John D, Grahacharya S, 16. Planning Commission of India. High level expert Ralte L. A conditional cash assistance programme for group report on universal health coverage for India. promoting institutional deliveries among the poor in New Delhi, 2011 - http://planningcommission.nic. India: process evaluation results. In: Richard F, Witter in/reports/genrep/rep_uhc0812.pdf - accessed 16 S, De Brouwere V, eds. Reducing financial barriers April 2013. to obstetric care in low-income countries. Antwerp. 17. Bhutta ZA, Cabral S, Chan CW, Keenan WJ. Reducing Belgium: ITG Press, 2008. pp. 257-273. maternal, newborn, and infant mortality globally: 6. India: Ministry of Home Affairs, Government of India. an integrated action agenda. International Journal Maternal Mortality in India: 1997-2003, Trend Causes of Gynecology and Obstetrics. 2012; 119 (Suppl 1): and Risk Factors. New Delhi: Registrar General of S13-S77. India, 2006. 18. Kerber KJ, de Graft-Johnson JE, Bhutta ZA, Okong P, 7. Thaddeus S, Maine D. Too far to walk: maternal Starrs A, Lawn JE. Continuum of care for maternal, mortality in context. Social Science and Medicine. newborn, and child health: from slogan to service 1994; 38(8): 1091–1110. delivery. The Lancet. 2007; 370(9595): 1358- 1369.

WHO South-East Asia Journal of Public Health 2012;1(4):362-368 367 The conditional cash transfer scheme to reduce Rajesh Kumar Rai et al. maternal mortality in India – a need for reassessment

19. Federal Ministry of Health. Ethiopia health sector 23. Lozano R, Soliz P, Gakidou E, Abbott-Klafter J, Feehan development program, HSDPIII, 2005/06 - 2010/11, DM, Vidal C, Ortiz JP, Murray CJ. Benchmarking mid-term review, vol. I. Component report. Addis of performance of Mexican states with effective Ababa, 2008. coverage. The Lancet. 2006; 368 (9548): 1729– 1741. 20. Negusse H, McAuliffe E, MacLachlan M. Initial community perspectives on the Health Extension 24. 24. Knaul FM, González-Pier E, Gómez-Dantés O, Program in Welkait, Ethiopia. Human Resource for García-Junco D, Arreola-Ornelas H, Barraza-Lloréns Health. 2007; 5: 21. M, Sandoval R, Caballero F, Hernández-Avila M, Juan M, Kershenobich D, Nigenda G, Ruelas E, Sepúlveda 21. Sebastian MS, Lemma H. Efficiency of the health J, Tapia R, Soberón G, Chertorivski S, Frenk J. The extension program in Tigray, Ethiopia: a data quest for universal health coverage: achieving envelopment analysis. BMC International Health and social protection for all in Mexico. The Lancet. 2012; Human Rights. 2010; 10: 16. 380(9849): 1259-1279. 22. Frenk J. Bridging the divide: global lessons from evidence-based health policy in Mexico. The Lancet. 2006; 368(9539): 954–961.

368 WHO South-East Asia Journal of Public Health 2012;1(4):362-368 Perspective

Pandemic influenza preparedness planning: lessons from Cambodia Anthony C Huszara, Tom Drakea, Teng Sreyb, Sok Touchb, Richard J Cokera,c

Introduction Experts have estimated that a pandemic and there is now early evidence that certain influenza outbreak could cost more than 60 immunemodulatory pharmaceuticals, including million lives worldwide.1 When considered off-patent statins, may be effective in in combination with recent research reducing influenza-related mortality.8 Non- demonstrating human-to-human transmission pharmaceutical approaches are varied, but potential of the highly pathogenic strain of include investing in improved surveillance avian influenza A(H5N1),2 this highlights the by way of sentinel sites and a network of growing need for public health officials to approved laboratories, social distancing by review their pandemic influenza preparedness risk communication, the closure of schools and strategy. This is especially true for South- workplaces, and the use of protective personal East Asia which is predicted to be the likely equipment (PPE). epicentre of future pandemics.3 One drawback of these guidance documents Following the 2009 H1N1 swine flu is that they hold a neutral perspective on which pandemic, many governments drew upon policy options might be the most beneficial, guidance from the World Health Organization and furthermore do not give any indication on (WHO) pandemic influenza preparedness how best to allocate resources. It should be a framework4 to revise their national pandemic priority to ensure that the US$ 2.7 billion that influenza plans. Cambodia was no different has been invested globally in the past seven and the government updated its National years into pandemic preparedness9 is spent Comprehensive Avian and Human Influenza effectively. Plan by producing three further pandemic As part of a wider project evaluating influenza strategies.5–7 pandemic influenza mitigation investment Interventions can be separated into options in Cambodia, we conducted semi- pharmaceutical and non-pharmaceutical structured interviews with members of categories. Pharmaceutical options include Cambodia’s pandemic influenza national antiviral stockpiling, the use of vaccines, coordinating committees to assess current which would take six months to produce, beliefs. Informants were chosen from the

aCommunicable Disease Policy Research Group, London School of Hygiene and Tropical Medicine, Bangkok, Thailand. bDepartment of Communicable Disease Control, Ministry of Health, Phnom Penh, Cambodia. cSaw Swee Hock School of Public Health, National University of Singapore, Singapore. Correspondence to Anthony C Huszar (e-mail: [email protected])

WHO South-East Asia Journal of Public Health 2012;1(4):369-373 369 Pandemic influenza preparedness planning Anthony C Huszar et al.

National Pandemic Planning and Response Contrasting concerns Committee, National Committee for Disaster Despite the overarching importance placed Management, National Pandemic Vaccine upon all of the WHO-backed interventions Committee, and the Joint Partners Interface in the first instance, some informants raised Group. Of the 33 identified members, 14 concerns that were not held consistently (42%) responded and took part. Informants across the group. This was observed with were first asked to score the 10 pandemic both pharmaceutical and non-pharmaceutical policy areas, as listed above, based on approaches. importance and their willingness to invest resources, on a scale of one to five (with five meaning ‘very important’). Participants were Pharmaceutical interventions permitted to use each number more than once Many informants spoke of the need to in order to gain a more reflective estimate stockpile antivirals, with the majority of of how the different interventions compared informants opening with a line similar to this with each other. They were then invited to participant: comment on how they reached their decision “It is very critical to have a stockpile for for each area. Using basic statistical methods Tamiflu.” and Grounded Theory,10 two main themes emerged from the analysis of the transcripts, Yet in contrast to this stated need for i.e. difficulties in prioritizing the different Tamiflu, some informants raised concerns pandemic preparedness interventions, and a about the associated costs of antiviral lack of consensus on certain interventions. We stockpiling, including storage and distribution, believe these problems are also likely to exist as well as the assumptions made about the in other South-East Asian countries that share drug’s efficacy. These concerns are illustrated similar attributes.3 by the following comments:

“[We] need to spend a lot of money and Difficulties in prioritization the duration of use is short thus stockpiling Given that the ability to prioritize different is not important... if we stock it, we waste policy areas or interventions is a key part of money… It is not necessary for a developing the health planning process,11 it was surprising country like us.” that all but one of the policy areas scored “[We make an] assumption that antivirals between four and five, which we interpreted as are good...we assume that the virus is ‘important’ to ‘very important’. This highlighted sensitive to antivirals… and that antivirals an inability to differentiate adequately between decrease mortality, and decrease the period policy areas. The single policy area that did of transmissibility.” not receive all-round acknowledgement of its importance was, unsurprisingly, generic Vaccination strategies also appeared to immunemodulatory pharmaceuticals, due to invoke mixed concerns among the informants. the current paucity of evidence. The results All informants commented on at least one of are shown in Table 1. the following issues: speed and capacity of vaccine manufacturing, unknown efficacy, and cost. Yet despite these legitimate concerns, investment for this policy option remained popular among respondents.

370 WHO South-East Asia Journal of Public Health 2012;1(4):369-373 Anthony C Huszar et al. Pandemic influenza preparedness planning

Non-pharmaceutical in less cost-effective measures. This could in interventions effect result in unnecessary loss of life in the All 14 respondents agreed on the importance event of a future pandemic. of governance, diagnostic capacity, and human resources, and that all medical staff must What next? receive suitable PPE. Justifications included In order to support the decision-making the International Health Regulations and process in pandemic influenza policy, we healthworkers’ faith in the equipment. The suggest need for better research on two general impression of PPE among informants counts. First, there needs to be better was captured by this informant: evidence on the effectiveness of interventions, “without [PPE] we can’t do anything.” especially in low-income countries. Second, appropriate economic evaluation needs to In contrast, another stated: complement these studies in order to guide “Our country is poor and the PPE is decisions on implementation. While we expensive and... if we wait for PPE we can’t recognize the inherent difficulties in pandemic do anything else.” influenza modelling due to its uncertain timing, and viral virulence and response to The policy option for critical care equipment interventions, having a better grasp of the generated even more polarized results. Some relative differences between the interventions informants firmly stated the importance of described in the WHO pandemic influenza having life-saving critical care equipment, preparedness framework would allow for more perhaps over-stating their efficacy, while intelligent decision-making by public health others made comments about their being officials. costly and potentially non-cost-effective. Social distancing raised similar mixed responses: In order to facilitate this process of four informants stated that it was a valid policy generating better evidence, we make two option, while five others raised concerns about recommendations. First, pre-approved clinical efficacy and difficulties with implementation. trials, where ethical approval and funding would be confirmed ahead of the next pandemic, may be able to generate more Why does this matter? reliable strain-specific data that could be These interviews highlight the difficulties that used as events unfold. Second, improved policy-makers face when allocating resources communication could enable researchers to to pandemic influenza preparedness in help policy-makers interpret their results Cambodia, and may also reflect difficulties in the context of pandemic uncertainty, faced by their counterparts throughout and conversely, it is crucial that in such a South-East Asia. Our findings also show that challenging area of decision-making, policy- drawing concrete conclusions from the current makers should liaise with researchers to help evidence base is challenging, possibly because direct their work. the evidence has largely been generated in high-income settings. While high-income countries are able to allay the uncertainties Acknowledgements by investing across the board, countries with We would like to thank the informants for their limited resources may inadvertently invest time in conducting these interviews. This work was supported through Deutsche Gesellschaft

WHO South-East Asia Journal of Public Health 2012;1(4):369-373 371 Pandemic influenza preparedness planning Anthony C Huszar et al.

2.8 2.8 4.1 4.1 4.4 4.4 4.2 4.2 4.4 4.4 4.7 4.7 4.3 4.3 4.9 4.9 4.0 4.0 4.8 4.8 CI] Mean Mean score [95% [95% [1.8-3.8] [3.1-4.9] [3.4-4.9] [3.2-4.7] [3.4-4.9] [3.7-5.0] [3.3-4.7] [3.9-5.0] [3.0-4.5] [3.8-5.0] 4 4 5 5 5 com - ment 3 No 5 4 4 5 5 No No 4.5 com - ment score score 5 3 4 4 3 No 4 5 4 No No 4.5 com - ment score 4 5 5 5 5 No com - ment 10 11 12 13 14 4 5 5 No 3.5 3.5 3.5 4.5 com - ment 9 5 3 3 2 4 4 4 4 4 4 4 3 4 5 5 5 5 No com - ment score 3 No 5 5 5 5 4.5 2.5 4.5 com - ment 2 3 3 3 No 1.5 4.5 4.5 4.5 4.5 2.5 Key informant 6 7 8 1.5 3.75 score 5 4 No 4 5 5 4 5 5 5 5 5 3.5 willingness to invest willingness to com - ment 3 4 4 4 4 4 4 1 No 5 1 2 3 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 No 4.5 com - ment Human resources Human resources doctors, nurses, (e.g. technicians) laboratory Antiviral stockpiling Antiviral Personal protective protective Personal equipment (e.g. masks, gloves) Critical care equipment equipment care Critical (e.g. and training ventilators) Influenza diagnostic Influenza diagnostic capacity Personal protective protective Personal equipment stockpiling masks, gloves) (e.g. Policy area Policy Health Health facility resources Stockpiling Vaccination (assuming an (assuming an Vaccination be produced can vaccine effective the 6 months after approximately is identified) pandemic strain Generic immune-modulators - Generic immune-modulators - to support framework a developing immune- other or statins the use of modulators Governance (including strategies for for (including strategies Governance frameworks) legal and cooperation Measures to temporarily reduce reduce temporarily Measures to social contact (including school or work closure) Table 1: Fourteen key pandemic influenza informants in Cambodia were asked to score each pandemic influenza policy based on importance and their policy based on importance each pandemic influenza to score asked were in Cambodia informants pandemic influenza key 1: Fourteen Table Scores were given from one to five (with five being the most important). ‘No comment’ was used when the informant had insufficient knowledge of being the most important). ‘No comment’ (with five from one to five Scores were given provided. was score no but expressed was opinion an when used was score’ ‘No score. a give to area policy

372 WHO South-East Asia Journal of Public Health 2012;1(4):369-373 Anthony C Huszar et al. Pandemic influenza preparedness planning

für Internationale Zusammenarbeit (GIZ) 5. Cambodia, Ministry of Health. Health sector pandemic GmbH on behalf of the German Federal Ministry influenza response strategy 2011-2015. Centers for for Economic Cooperation and Development Disease Control and Prevention, 2010. and the Rockefeller Foundation. 6. Cambodia, Ministry of Health. The Cambodia pandemic influenza communication strategy 2011- 2015. Centers for Disease Control and Prevention, 2010. References 7. Cambodia, Ministry of Health. The Cambodia national 1. Murray CJ, Lopez AD, Chin B, Feehan D, Hill KH. rapid containment strategy 2011-2015. Centers for Estimation of potential global pandemic influenza Disease Control and Prevention, 2010. mortality on the basis of vital registry data from the 1918–20 pandemic: a quantitative analysis. The 8. Vandermeer ML, Thomas AR, Kamimoto L, Reingold Lancet. 2006 Dec; 368(9554): 2211–8. A, Gershman K, Meek J, et al. Association Between Use of Statins and Mortality Among Patients 2. Enserink M, Cohen J. One H5N1 Paper Finally Goes Hospitalized With Laboratory-Confirmed Influenza to Press; Second Greenlighted. Science. 2012 May Virus Infections: A Multistate Study. J Infect Dis. 4; 336(6081): 529–30. 2012 Jan 1; 205(1): 13–9.

3. Coker RJ, Hunter BM, Rudge JW, Liverani M, 9. UNSIC. 2010 Annual report. 2010 - http://un- Hanvoravongchai P. Emerging infectious diseases in influenza.org/files/UNSIC2010Report.pdf - accessed southeast Asia: regional challenges to control. The 17 April 2013. Lancet. 2011 Feb 12; 377(9765): 599–609. 10. Strauss A, Corbin J. Basics of Qualitative Research 4. World Health Organization. Pandemic influenza techniques and procedures for developing grounded preparedness framework . Geneva: WHO, 2012 theory. 2nd ed. London: Sage Publications, 1998. - http://www.who.int/influenza/resources/pip_ framework/en/index.html - accessed 17 April 11. Buse K, Mays N, Walt G. Making health policy. 2nd 2013. edn. Milton Keynes: Open University Press, 2012.

WHO South-East Asia Journal of Public Health 2012;1(4):369-373 373 Review

A review of Japanese encephalitis in Uttar Pradesh, India

Roop Kumaria, Pyare L Joshib

Background: Japanese encephalitis (JE) is a major public health problem in India. When the first case was reported in 1955, the disease was restricted to south India. The disease spread to north India in 1978 from where extensive and recurrent outbreaks of JE have been reported ever since. An attempt has been made to review the epidemiology of JE over the past 30 years and suggestions made for its prevention and control.

Methods: An epidemiological profile of JE (1978–2009) has been compiled and analysed to understand the trend and status of the disease.

Results: In India, while 24 states are endemic for JE, Uttar Pradesh contributed more than 75% of cases during the recent past. Over the years, the seasonal trend has changed and the epidemic peak of the disease has advanced by one month.

Conclusion: JE is closely associated with the pattern of precipitation, flooding and rice production systems. Analysis of trends and influencing factors will help in designing suitable strategies for the prevention and control of JE in the country. Continuous monitoring of vector populations and JE virus infection rates in vector mosquitoes will help in predicting an outbreak and in taking effective intervention measures.

Key words: Japanese encephalitis, epidemiology, Uttar Pradesh, India, vector, high-risk districts

Introduction Japanese encephalitis (JE), a vector-borne and the People’s Republic of China has steadily viral disease, is endemic to large parts of declined. In contrast, new epidemic foci of JE Asia and the Pacific regions.1 An estimated were reported in the parts of tropical south- 3 billion people are at risk, and the disease has eastern Asia as late as 1969.3 recently spread to new territories globally.2 JE In India, the first human case was reported is a major public health challenge due to its from North Arcot district of Tamil Nadu in high epidemic potential, high case–fatality and 1955.4 Until 1973, the disease was confined to neuropsychiatric sequelae among survivors. JE southern parts of India, with low prevalence; was first recognized in Japan in 1924. Since subsequently the disease spread to various the late 1960s, the size of epidemics in Japan other parts of India. The first outbreak of

a National Centre for Disease Control, Delhi, India. b National Institutes of Health and Family Welfare, Delhi, India. Correspondence to Roop Kumari (e-mail: [email protected]).

374 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

JE was recorded in 1973 from Burdwan and rivers originating in the Nepali hills. Regular Bankura districts of West Bengal. In 1978, flooding is devastating for the crops, life and suspected outbreaks of JE were reported from property of the farming community, and to a 18 states and 24 states and Union Territories large extent determines their socioeconomic have reported suspected JE cases till recent and political lives.8 past.

The State of Uttar Pradesh (UP) is highly Domestic animals endemic; since the first report of a JE epidemic The common domestic animals include cows, in 1978,5,6 extensive and recurrent outbreaks buffaloes, goats, pigs, dogs and horses. have been reported from the eastern parts The pig is known to be the amplifier host of of the State. An attempt has been made in the JE virus (JEV). Despite this knowledge, this article to review the epidemiology of JE unorganized piggeries are common in most since 1978, particularly in UP, and to make districts in the region. recommendations for its prevention and control. Methodology

Study area Surveillance and case definitions Clinical diagnosis is made by the respective Uttar Pradesh, 241 000 km2, is the most medical officers of the health centres/ populous State in India with over 199 million hospitals who attend the JE cases in outpatient people as of 2011.7 The majority of the or inpatient departments, using standard population is engaged in agriculture, which case definitions as per national guidelines. contributes to 41% of the State’s economy. Serum and cerebrospinal fluid samples of Since its division in 2000 into two states, suspected cases are tested by IgM enzyme- UP has 71 districts, 106 thousand villages, linked immunosorbent assay (ELISA) for 3640 primary health centres and 18565 sub- confirmation of JE. Previously, laboratory centres. and treatment facilities were only available Climatically, while the State experiences at Baba Raghav Das (BRD) Medical College, tropical monsoons, weather conditions ; King George Medical College, change significantly as per the location and (now Shahuji Maharaj Medical onset of the season. UP has three broad University) and Sanjay Gandhi Post Graduate seasons – winter from October to February Institute of Medical Sciences, Lucknow. Since (minimum temperatures of 3–4 °C), summer 2005, district hospitals are being strengthened from March to mid-June (temperatures up to for diagnosis of JE cases. Eastern districts of 45 °C), and the rainy season from June to UP are highly endemic, and most JE patients September (85% of average annual rainfall from these areas have been treated at BRD of 99cm and temperatures of 30–45 °C). The Medical College, along with JE patients from entire region is prone to annual flooding due border districts of Bihar and even neighbouring to its topography. Five districts in Eastern Nepal. UP (Gorakhpur, , Maharajganj, In 2006, the National Vector Borne Sant Kabir Nagar and Siddarth Nagar), a Disease Control Programme (NVBDCP) region surrounded by embankments and modified the case definition of JE. Since drainage structures, are the most affected then, epidemiological surveillance for acute by floods as they are traversed by major encephalitis syndrome (AES) was initiated

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 375 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

and suspected JE cases are now reported as Epidemiological data were divided into three follows:“Clinically, a case of AES is defined decadal periods, i.e. 1978–1987, 1988–1997 as a person of any age, at any time of year and 1998–2009 to see the trend of the disease with the acute onset of fever and a change in and its association with ecological change, if mental status (including symptoms such as any. confusion, disorientation, coma, or inability to talk) AND/OR new onset of seizures (excluding Results simple febrile seizures). Other early clinical findings may include an increase in irritability, As compared with the total number of JE somnolence or abnormal behaviour greater cases reported in the country, UP contributed than that seen with usual febrile illness”.9 JE a fifth of the disease burden (20.4% of cases surveillance therefore aims to identify patients and 18.7% of deaths) during 1978–1987 with AES and thereafter confirm JEV infection (Figures 1A and1B). Its contribution increased using IgM ELISA test. between 1988 and 1997 to 24.3% cases and 20.9% deaths, and represented well over half of the total cases and deaths reported in the Methods country between 1998 and 2009. While the Epidemiological data for JE over the past proportion of JE cases reported from Uttar 32 years (1978–2009) as reported by State Pradesh in 1997 was only 14%, it started Health Authorities and from published and increasing dramatically from 1998 and in unpublished documents were collected and 2005, a major outbreak in UP contributed over analysed for the purpose of this study. The 90% of all suspected JE cases (6061) – and burden of JE morbidity, annual incidence over 89% deaths (1501) – in the country. rate (number of cases per 100 000 risk During the four following years, 2006–2009, area populations) and case–fatality rate 80.8%, 73.6%, 78.5% and 77.0% of cases, (proportion of deaths against the total number respectively, were contributed by the State of of cases reported per year) were estimated. Uttar Pradesh (Figure 1C).

Figure 1A: Proportion of Japanese encephalitis cases contributed by Uttar Pradesh in national figures by decade, 1978–2007

100.00% 90.00% 80.00% 43.32% 70.00% 60.00% 79.60% 75.74% 50.00% 40.00% 30.00% 56.68% 20.00% 10.00% 20.40% 24.26% 0.00%

1978-1987 1988-1997 1998-2007

Uttar Pradesh India

376 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

Figure 1B: Proportion of Japanese encephalitis deaths contributed by Uttar Pradesh in national figures by decade, 1978–2007

100.00% 90.00% 80.00% 47.60% 70.00% 60.00% 81.30% 79.14% 50.00% 40.00% 30.00% 52.40% 20.00% 10.00% 18.70% 20.86% 0.00%

1978-1987 1988-1997 1998-2007

Uttar Pradesh India

Figure 1C: Contribution of Japanese encephalitis cases by Uttar Pradesh in national figures by year, 1997–2009

100% 90%

80% 70% 60%

50% 90.1 40% 80.8 73.6 78.5 77.0 30% 60.1 49.6 48.8 20% 40.0 45.1 43.8 34.2 10% 14.0 0% 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

UttarPradesh India

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 377 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

Trend of JE cases and deaths in at 18 760 cases and 4189 deaths, reported Uttar Pradesh from 40 districts in the State. Figure 2 shows that, despite year-to-year fluctuation, there was no definite trend of JE Case–fatality rate cases and deaths in the State since 1978. In The annual case–fatality rate (CFR) of reported the first decade (1978–1987), a total 9299 suspected cases shown in Figure 2 was higher reported JE cases, and 3103 deaths attributed in the beginning (1978–1987), ranging from to suspected JE cases, were reported from 31.2% to 48.0%,before declining in 1988– 46 districts. The first outbreak of suspected 1997 (range 21.0–39.9%) and 1998–2007 JE was recorded in UP in 1978, during which (18.6–24.9%). In the 2005 JE outbreak, the 3550 cases and 1117 deaths were reported overall CFR was 24.9% as compared with earlier from 40 districts. While in 1979 only 150 cases outbreaks in 1978 (31.5%), 1985 (34.5%) and were reported, this number again increased in 1988 (31.5%). This indicates an improvement 1980 to 1604 cases reported from 15 districts. in the management of JE cases over time. Thereafter, the incidence of JE declined until 1985. In the second decade (1988–1997), the Seasonal variation total number of reported cases and deaths were 10 064 and 3194, respectively, from 25 Monthly reported JE cases from 1998 to districts. During 1998–2009, the burden of JE 2007 indicate that sporadic cases occurred was much higher than the previous decades, in the month of June and reached a peak in September before declining thereafter.

Figure 2: Trend of suspected Japanese encephalitis: cases, deaths and case– fatality rate (%) in Uttar Pradesh, 1978–2009

7000 60.0

6000 50.0

5000 40.0

ate (%) 4000

yR 30.0 3000

Cases&Deaths 20.0

2000 Case Fatalit

10.0 1000

0 0.0

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Cases Deaths CFR

378 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

Seasonal peaks of JE cases have occurred subsiding in November. The entire course of during the months of August to October in the the disease during the 1985 outbreak was of State; data also indicate that maximum cases 8 to 9 weeks’ duration from mid September to were concentrated in the period from July to mid November. However, the 1978 outbreak October, coinciding with rainy and post-rainy reported cases reported from mid October seasons. With the onset of winter, JE incidence until November. Thus, over three decades, declined substantially. the peak of the epidemic has advanced by one month approximately. Average AES/JE cases in recent years compared with data of 1988 (Figure 3) show some variation in seasonal trends and cyclic Districts with repeated patterns of the disease. For example, cases have occurrence of JE cases started to increase from early July, peaking in Figure 4 shows the districts where JE cases October as opposed to September (cases also occurred repeatedly from 1978 to 2007. peaked in October in the 1970s and 1980s). Of the total 72 districts in the State, 17 During the outbreak in UP in 2005, weekly districts have reported only sporadic cases data indicated that cases started occurring of suspected JE (not laboratory confirmed) in July with a sharp peak in end August/early at one time in 30 years. Of these districts, September. In contrast, the 1988 outbreak in 7 (Bijnor, Bulandshahr, Etah, Kaushambi, Gorakhpur and Deoria districts witnessed JE Deharadun, Pauri and ) reported cases starting in September and reaching a only 1 or 2 cases of suspected JE, 6 districts peak in the second week of October, before (, , Hamirpur, Badaun, Ormoi

Figure 3: Evolution of average number of Japanese encephalitis cases from 1988 to 2009 showing shift of seasonal peak in Uttar Pradesh

1600

1400

1200

1000

800

Cases

600

400

200

0 Jan Feb. March April May June July Aug. Sept. Oct. Nov. Dec.

1988 1998 to 2005 2006-2009

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 379 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

Figure 4: Map showing frequency of Japanese encephalitis cases reported annually from each district in Uttar Pradesh, 1978–2007

and Lalitpur) reported 3–10 cases and the During the second decade under review remaining 4 districts (, Mirzapur, (1987–1997), JE emerged for the first time as Rampur and ) reported 11–26 cases. a problem in Siddharth Nagar district in 1989, However, 15 districts were reported once only in Maharajganj in 1990 and in Kushinagar in 1978 and JEV infection in these districts was in 1997. Since then, these districts have not confirmed serologically. regularly reported JE cases every year. An in- depth review of the origins of these districts Among the remaining affected districts, indicates that Siddharth Nagar was created nine reported suspected JE cases for only 2–4 in 1988 out of ; Maharajganj years since 1978, the number of cases ranging was carved out from Gorakhpur districtin from 4–51. Three districts reported suspected 1989, and Kushinagar emerged in 1994 from JE cases only twice in a period of 30 years. . Thus, the high incidence of JE Figure 4 also reveals that 15 districts reported reported in the newly created districts was due JE 5–9 times during last three decades. JE to their geographic affinity with their highly JE occurred almost every year in four districts endemic parent districts since 1978. However, (, Basti, Deoria and Gorakhpur). in 1995 and 1996, no case was reported by Since 1978, repeated occurrence of JE cases the state, therefore, there was less number in each year was recorded in Deoria and of districts affected. Gorakhpur districts in eastern UP, while Basti district JE was reported 26 times during the The three new districts that emerged in the 30-year period in question. decade 1998–2007 – , Sant Kabir Nagar and – have also reported JE

380 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

cases since their creation. Sant Kabir Nagar (Lakhimpur, , Gonda, Gazipur, emerged in September 1997 from Basti Lucknow, and Sultanpur) showed a district, while Balrampur was created from sporadic distribution of cases from 0.5 to 1.0 in 1997. Shravasti became per 100 000 population. a new district in 2000 carved out from the From 1988 to 1997, like the earlier decade, districts of Bahraich and Gonda. All these the highest incidence rate was reported from parent districts were highly endemic for JE. Deoria (11.90 cases per 100 000 population) Figure 5 shows the increasing number of followed by Gorakhpur (11.59). It is important districts affected by JE in Uttar Pradesh. to highlight that during this decade, a serious JE problem emerged for the first time in the Disease burden by district two newly created districts of Maharajganj Figures 6A–C indicate the distribution of cases (847 cases and an attack rate of 3.84 per in different districts from 1978 to 2007. The 100 000 population) and Siddarth Nagar (274 incidence rates appear to be higher in eastern cases with an attack rate of 1.38 per 100 000). districts of UP. The district-wise estimation Since then, these districts have reported JE of JE cases per 100 000 population (annual cases every year. incidence rate) from 1978 to 1987 ranged First time JE incidence was reported from 0.01 to 9.87. Deoria district had the from immediately after maximum average annual incidence rate its creation, which has since emerged as the (9.87) followed by Gorakhpur (9.40), most problematic district in the State with (3.14), Azamgarh (2.01), (1.66) the highest average annual incidence rate and (1.19). The remaining districts (29.90 per 100 000 population), followed by

Figure 5: Shifting scenario of Japanese encephalitis in Uttar Pradesh by year showing annual number of affected districts and trends in incidence rate

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 381 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

Figure 6: Distribution of reported Japanese encephalitis cases in Uttar Pradesh 6A: Japanese encephalitis incidence rates, 1978–1987

6B: Japanese encephalitis incidence rates,1988–1997

382 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

6C: Japanese encephalitis incidence rates, 1998–2007

Maharajganj (21.80), Gorakhpur (10.01), Kushinagar district has become the most high- Siddharth Nagar (7.60), Deoria (6.59) and risk district in the State with an average annual Bahraich (3.62). incidence rate over 10 years of 29.90 per 100 000 population, followed by Maharajganj During the period from 1998 to 2007, (21.80). The incidence rate was comparatively Kushinagar, Maharajganj, Gorakhpur, Siddharth less in Gorakhpur (10.01) and Deoria (6.59) Nagar, Deoria and Bahraich were the most districts, where a high proportion of JE had JE problematic districts in the State. Data been reported repeatedly since 1978. indicate that Azamgarh, Ballia, Kheri and Pilibhit districts had been highly endemic from 1978 to 1988, after which time the disease Age distribution of JE cases burden declined. Analysis of age distribution of JE cases from in outbreaks in 1985, Besides increased disease burdens in some 1988, 2005 and 2009 (Figure 7) show that districts, new epidemic foci of JE have also children were more susceptible than adults. been emerging in Uttar Pradesh. During the The average age of affected children in eastern second decade of study from 1988 to 1997 UP was 6.8 years between 2004 and 2006. The (Figure 6B) first-time JE problems emerged highest number of cases occurred in children in the new districts of Maharajganj (1990), aged between 0 and 10 years: 47% in 1985, Siddharth Nagar (1989), and Kushinagar 60% in 1988, 88% in 2004, and 94% in 2006 (1997). Since these districts were part of outbreaks, respectively. the old endemic districts, they also reported JE cases every year since their creation.

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 383 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

Figure 7: Distribution of Japanese encephalitis cases by age and sex, Gorakhpur division, Uttar Pradesh 7A: Distribution of JE cases by age/sex, Gorakhpur, 2009

80 70 60 50 40

Cases 30 20 10 0 0to1 2to5 6to1011to14 above 15

Years

Male Female

7B: Distribution of JE cases by age/sex, Gorakhpur, 2005

300

250

200

150

Cases 100

50

0 0to1 2to3 4to6 7to1011to14 above 15

Years

Male Female

384 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

7C: Distribution of JE cases by age/sex, Gorakhpur, 1988

800 700 600 500

Cases 400 300 200 100 0 1to2 3to1011to1819&above Year

Male Female

7D: Distribution of JE cases by age/sex, Gorakhpur, 1985

120

100

80

60

Cases

40

20

0 0to4 5to9 10 to 19 19 above Year

Male Female

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 385 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

Analysis of 3887 JE cases reported in 2005 Nagar) reported coverage of above 95% of the from Gorakhpur and Basti divisions indicated estimated target. This success led to the JE that the maximum cases (58%) were reported vaccination programme being extended in a in the age group of4 to 6 years, while 18% of phased manner to the other endemic districts. cases were reported in the age group 15 years During 2007, an additional seven endemic and above. Age distribution of JE seropositive districts (Bahraich, Balrampur, Basti, Gonda, cases reported by BRD Medical College during , and Shravasti) were 2009 shows that 75.1% of confirmed JE cases covered. A further 16 districts were included were reported in the age group below 15 in 2008 and 2009. Sustained high coverage of years. In earlier outbreaks in 1985 and 1988, JE vaccination is clearly required to bring down the proportion of adults was higher – 25% and the disease burden. The percentage coverage 22%, respectively – which then declined by of JE vaccination in Uttar Pradesh from 2006 3% in 2004. After JE vaccination in children to 2009 is given in Table 1. up to 15 years, the proportion of cases in Although the reported coverage was very adults slightly increased in 2009 (24.87%) high, an independent evaluation by UNICEF in in 2006. 2008 showed that only 51% of eligible children received JE vaccination during the campaign. Sex distribution of JE cases However, it may be said that the proportion The 2005 JE outbreak in UP reported a case of JE sero positive cases in Gorakhpur did not breakdown of 61% males and 39% females cease to occur but there was decline, from (deaths attributed to disease were 59% in 38.8% in 2005 to 14.2% in 2006, 12% in males and 41% in females). However, in 2007, 8% in 2008 and 10.8% in 2009 after Gorakhpur division, a line listing of 1087 cases introduction of JE vaccination in 2006 in high showed a ratio of 56.0% males to 43.8% endemic districts (Report of the National females. A year later in 2006, a line listing of Institute of Virology Field Unit, Gorakhpur). 249 cases revealed 155 (62.3%) males against Figure 8 shows that JE cases usually 94 (37.7%) females. From 2007 to 2009, line appear in the third month after onset of the listings showed steady average of male to monsoon. For example, in 2005 JE cases female cases of JE (68 (57.1%) males and were reported from August following the 51 (42.9%) females, and 122 (59.5%) males monsoon in May, while in 2006 and 2007, the and 83 (40.5% females, respectively). monsoon started in April and cases started being reported in July. Vector mosquitoes JE vaccination proliferate profusely subsequent to a monsoon, In 2006, the Government of India launched stimulating an increase of virus activity in a JE vaccination campaign for children from mosquitoes and human–mosquito contacts 1 to 15 years of age. This was followed by and thus transmission, followed by JE cases. immunization of new cohorts as an integral This preparatory phase of 2–3 months should component of the Universal Immunization be used to prepare for the prevention of an Programme with single dose of live attenuated outbreak. The reason for the 2005 outbreak JE vaccine (SA–14-14-2) in 11 highly endemic may have been caused by prolonged rainfall: districts of four states (Assam, Karnataka, Uttar as seen in Figure 8, the pattern of rainfall was Pradesh and West Bengal). Of the 11 districts, different in that year, which had two peaks of 7 (Deoria, Gorakhpur, Kheri, Kushinagar, rainfall – one in July and another in October Maharajganj, Sant Kabir Nagar and Siddharth – with a prolonged rainy season from May to

386 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

Table 1: Coverage of Japanese encephalitis vaccine (SA 14-14-A) in Uttar Pradesh

Target (1–15 year Total vaccinated Sl No. Year District Total coverage (%) olds) children 1 2006 Gorakhpur 1 390 307 1 349 047 97.03 2 Deoria 1 074 219 1 072 683 99.86 3 Kushinagar 1 095 877 1 085 055 99.01 4 Maharajganj 776 500 806 996 103.93 5 Kheri 1 183 481 1 218 364 102.95 6 Sant Kabir Nagar 542 062 511 417 94.35 7 Siddarth Nagar 775 934 792 944 102.19 Total 6 838 380 6 836 506 99.97 1 2007 Ambedkar Nagar 764 068 741 354 97.03 2 Bahraich 990 327 992 254 100.19 3 Balrampur 623 020 622 963 99.99 4 Barabanki 1 074 154 1 063 815 99.04 5 Basti 774 322 750 262 96.89 6 Gonda 1 040 501 1 045 957 100.52 7 Mau 719 800 691 341 96.05 8 Raebareli 1 058 987 1 029 154 97.18 9 Saharanpur 1 056 185 923 246 87.41 10 1 385 606 1 312 326 94.71 11 Shravasti 331 903 326 485 98.37 Total 9 818 873 9 499 157 96.74 1 2008 Azamgarh 1 400 000 1 329 471 94.96 2 Ballia 1 067 337 1 040 948 97.53 3 Bareilly 1 337 392 1 399 552 104.65 4 649 634 641 736 98.78 5 Hardoi 1 270 372 1 296 593 102.06 6 Lucknow 1 506 695 1 477 852 98.09 7 Muzzafar Nagar 1 340 232 1 243 694 92.80 8 Sultanpur 1 234 068 1 229 295 99.61 9 1 030 843 1 049 252 101.79 Total 10 836 573 10 708 393 98.82 1 2009 1 953 904 1 673 687 85.66 2 Pratapgarh 1 028 331 984 230 95.71 3 Nagar 1 660 800 1 227 209 73.89 4 953 078 887 267 93.09 5 Fatehpur 913 749 681 420 74.57 6 Jaunpur 1 546 931 1 370 141 88.57 7 1 202 355 1 007 125 83.76 Total 9 259 148 7831079 84.58

Source: NVBDCP

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 387 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

Figure 8: Correlation of monthly rainfall and Japanese encephalitis cases in Uttar Pradesh

Cases Rainfall 3500 600

3000 500

2500 400

2000

300

1500 2004-cases 2005 200 1000 2006 2007 100 2004-Rainfall 500 2005 2006 0 0 2007 Jan Feb. March April May June July Aug. Sept. Oct. Nov. Dec.

November. Other years had only one peak of to floods is attributed, among other things, to rainfall. heavy rainfall, low and flat topography and the silting of river beds which make river levels Discussion to rise and have impact on temperature and humidity. Flood problems of eastern UP were In India, 24 States/UTs were affected with JE due to reduction in the absorptive capacity of but the State of UP contributed more than 75 the soil. This `reduction’ has been magnified per cent of JE cases to the country (Figure 1). with the development of intensive canal In Uttar Pradesh, out of total of 71 districts, irrigation during the past five decades. Due 7 eastern districts were highly affected due to construction of unlined canal networks and to JE and contributed about 78.27% of the embankment; ground water table increased in total cases of the State during the recent this area resulting more flood and more water decade. Gorakhpur, Deoria and Basti have stagnations after the monsoon rains. As in been found to be high risk districts for JE Eastern UP, Bihar’s flood-prone area too tripled since 1978, and other four districts which from 2.5 m ha in 1954 to 6.8 m ha in 1994. were carved out from these three districts Flood proneness and water logging hit the namely Maharajganj, Kushinagar, Siddhath lives and livelihoods of people in several ways Nagar and Sant Kabir Nagar also remained including the growth of the vector mosquito high JE endemic since their inception. The population. Between 1951 and 1981, the area eastern districts are the most prone to JE than cultivated in kharif which fell from 214 000 western districts. Eastern district’s proneness ha to 68 000 ha due to annual flooding and

388 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

surface water logging; as a result, in these Although the burden of disease in Uttar areas, farmers clung to traditional mixed-crop Pradesh has increased, CFRs of JE cases have farming technologies.10 The tradition of animal declined over the years from 31.5% in 1978 husbandry too has been undergoing change to 18.0% in 2009. This is as a result of early due to waterlogging; as grazing lands remain detection, better case management and early submerged in water for long periods, large referral of patients to treatment centres. In bovines have declined in population. Marginal contrast in Nepal, overall mortality from JE farmers and landless have increasingly taken was 9.8% in the year 2000 and 20.9%in to piggery. All these factors contributes to 2003.14 A similar trend in CFR (7–30%) was increase of JE incidence in the eastern districts reported in Philippines from 1972 to 1985.15 during the past decades.10 An in-depth review of various outbreaks Studies from peninsular and eastern reported from Uttar Pradesh shows that in parts of India indicate that pigs are the main the 1985 outbreak of JE,16 the attack rate was vertebrate host of JEV and the major reservoir 0.08 and 0.16 per thousand population in of the infection.11 Because infected pigs act as Gorakhpur and Deoria districts, respectively. amplifying hosts, domestic pig rearing is an The average number of cases per village was important risk factor in the transmission of JE 1.1 in Gorakhpur and 1.2 in Deoria. Three to humans. In different parts of the country, years later in the 1988 outbreak, these 12–44% of the pig population have been found averages rose to 1.6 cases per affected to be positive for JE antibodies, particularly in village in Deoria,17 and 2.4 in Gorakhpur. This endemic areas. Pigs play an important role in difference may be attributed to variance in the natural cycle and serve as an amplifier host the population size and demographic profile since they allow manifold virus multiplication of the villages. without suffering from disease and maintain Eight UP districts were affected by the prolonged viraemia. Thus, mosquitoes biting 1988 outbreak: Gorakhpur, Deoria, Azamgarh, pigs can be dangerous for humans. In India, Basti, Gonda, Bahraich, Ballia and Faizabad. the population of pigs is reported to be more Nearly half of the cases and deaths were than 135 million, 31.7% (42.84 million) of recorded from Gorakhpur alone, which was the which are in Uttar Pradesh.12 As per the last worst affected district in the entire country.17 17th livestock census, UP has the largest This outbreak became the largest epidemic pig population in the country. Unorganized recorded in any state in India to date, with piggeries are plenty: the rural population who 4544 cases and 1413 deaths. The next major rear pigs virtually co-exist with these animals. JE outbreak to affect eastern was in 2005 and Hence, the chance of human infection is high.13 was the longest and most severe epidemic in The probability of vector mosquito species three decades: 6061 persons were affected, getting infected with JEV is higher when 1501 of whom died. The etiologic agent was infected mosquito populations dramatically confirmed to be Japanese Encephalitis virus in increase and the human biting rate increases. 2005 outbreak of viral encephalitis in UP.18 During the transmission season, when vector density is abundant, infected mosquitoes Analysis of epidemic peaks of JE outbreaks transmit infection to human beings as shows that the duration of outbreaks is incidental, dead-end hosts in the transmission getting longer. Uttar Pradesh experienced its cycle due to low and short-lived viraemia. first outbreak in 1978 from mid October to November,5,19 i.e. a duration of 6–7 weeks.

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 389 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

In 1985, an outbreak16 started in the second months. Relative to the 2005 outbreak, the and third week of September with an explosive duration of rainfall was longer compared with rise in incidence in the subsequent 4 weeks other years, starting in May, then peaking in followed by a decline over the next 2–3 weeks. July and August before its decline (Figure 8). The entire course of the epidemic was of 8 to However, a second peak of rain occurred in 9 weeks duration. Thus, while the duration of October, which explains the longest duration epidemics in the 1970s and 1980s was 6–9 of JE outbreak (August to November) in that weeks,16 the 2005 JE epidemic lasted for about year: incidence increased rapidly in August and 17 weeks. Note that the first JE outbreak in reached a peak in the first week of September, Karnal, Haryana in 1990 continued for about but the epidemic peak was not until November 10 weeks.20 due to the prolonged rainy season.

In all outbreaks until 1988, seasonality Thus, seasonality of the disease and a of the disease was recorded from September shift in epidemic peak may be attributed to to November, the post-monsoon season, the pattern of monsoons, floods and change which coincides with paddy cultivation and of weather in the State.24 It is also observed the period of peak mosquito density.11,23 The that JE cases usually appear in the third month maximum number of cases was generally after onset of the monsoon; thus the pattern reported in October.22 However, monthly JE of rainfall influences vector breeding and cases for the decade 1997 to 2007 indicate occurrence of JE cases. As per studies carried that sporadic JE incidence was reported in the out from 1990 to 1996 in Gorakhpur, vector month of June, gradually increasing from July density increased after onset of the monsoon and peaking in September before declining and reached its peak in September.25 However, (Figure 3). Maximum cases were concentrated the peak of JE cases was reported in October in the period August to October, coinciding when the vector population, particularly Culex with the rainy and post-rainy season. With tritaeniorhynchus, was on the decline.25 Such onset of winter, JE incidence subsided. It was a relationship between the epidemic and also reported that JE cases reached a peak declining vector population has also been during August/September in the Nepalese observed in Japan, where this type of pattern districts bordering UP in 2004.14 The results was attributed to the differences in total and reveal that between 1978 and 2009 in UP, the infected populations of vector species.26 peak of the epidemic season moved forward The present analysis shows that, though by approximately one month from October both sexes were affected, males usually to September, and since 1997, seasonality outnumbered females. The female to male of the disease has started in July instead of ratio in 2007 was 1:1.3. The 2005 outbreak September. showed female to male ratios of 1:1.28 and About 90% of the rainfall occurs during 1:1.58 while it was 1:1.7 in 198817 and 1:1.5 the south-west monsoon, lasting from about in 1985.16 Male predominance was also seen in June to September. Flooding usually occurs the first JE outbreak in 1978. Previous reports from July to August during the rainy season and recent analyses clearly indicate that in and is a recurrent problem that creates an general men are more prone to JE infection excellent breeding ground for JE vectors. than women, which could be attributed to Comparison of rainfall data with JE incidence more outdoor activity by the male population indicates vector density, and subsequently JE during the peak biting time of exophilic vectors cases increase one month after the peak rainy and more exposed body parts, especially the

390 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

lower extremities, to mosquito bites, while exposure and subclinical infections in such women tend to be fully covered by dresses. individuals. The estimated population at risk in endemic areas is 45.1%.24 However; in Data also reveal that, in the first JE relatively non-endemic areas where the virus epidemic in 1978 in UP, while no age group is introduced for the first time, individuals of was spared more cases (42.5%) occurred all age groups are susceptible. in children below 10 years of age.19 A high number of cases in children was also After vaccination in 2006, the proportion reported in the 1988 outbreak (adults 22%, of JE laboratory-confirmed positive cases children >10 years,60% of total cases).17 In declined in the Uttar Pradesh (NVBDCP contrast, the number of adult cases was higher Report). However, the virus remains a major in the 1985 and 1987 outbreaks. Similar and consistent cause of outbreaks in the observations were made in the Indian states eastern region of the State, accounting for of Assam, Bihar and West Bengal. An analysis approximately 10%–15% of total AES cases of age distribution in1990–1996 revealed that annually.24,32 the maximum number of cases occurred in Culexvishnui subgroup mosquitoes, children aged between 1 and 15 years, with comprising Cx. tritaeniorhynchus, Cx. vishnui the highest occurrence in the age group 6–10 and Cx. pseudovishnui, have been implicated years in Gorakhpur region.25 In 2004, of the as major vectors of JE in India33 as well as in total 115 cases analysed from Gorakhpur, 105 many countries of south-east Asia.34,35 These (91.3%) were among children below 15 years mosquitoes are usually found in rural rice- of age.29 In 1988, 22% cases occurred in growing and pig-farming regions of Asia, but adult age groups: This figure declined to 16% can also be found at the outskirts of cities in in 2005, largely because most adults in the close proximity to human populations. They area were immune to JEV, and the majority of prefer to breed in rice fields, and outbreaks cases were seen in children from 3 months to of JE are commonly associated with intensive 15 years of age.18,27 Since introduction of JE rice cultivation.23,36,37 In Uttar Pradesh, SA14-14-2 vaccine in 2006, JE cases reported based on the elevated density and infection in upper age groups, including from Uttar with JEV, Cx.tritaeniorhynchusis considered Pradesh, have increased from 3% in 2006 responsible for causing epidemics in the to 25% in 2009. Similarly in Japan, JE has area. Cx. pseudovishnui, Cx. whitmorei, become a disease of older age groups since Cx. gelidus, Cx. epidesmus, Anopheles 1967, the CFR being higher in the elderly. subpictus, An. Peditaeniatus and Mansonia A gradual decrease of herd immunity in older uniformis are suspected to play some role in persons has been taking place, creating a the epidemiology of JE in the region.25 potential for future disease outbreaks.30 In Gorakhpur area, JEV infection was Over the years, children in UP have detected in female Cx. tritaeniorhynchus and become a more susceptible group to JE. Cx. epidesmus tested by ELISA. But JEV could However, those with debilitating chronic illness be isolated only from Cx. tritaeniorhynchus or immunosuppression are at greater risk of by the Toxorhynchites splendens inoculation- disease, with symptoms ranging from malaise indirect immunofluorescence (Toxo-IFA) to meningo encephalitis with seizures and insect bioassay.38 Isolation of JEV from death.31 The attack rate decreases with age Cx. epidesmushas been reported in Bankura, in endemic populations due to the presence of West Bengal.39 During the 1991 outbreak in neutralizing antibodies as a result of natural

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 391 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

Gorakhpur, JEV was isolated from wild caught irrigation, may check proliferation of breeding Cx. tritaeniorhynchus.40 However, JEV has since of Cx. tritaeniorhynchus and other JE vectors been detected by ELISA42 at the National Centre and even contain the vector population during for Disease Control from adult Cx. gelidus, JE transmission season. In addition, rice Cx. pseudovishnui, Cx.bitaeniorhynchus, fields support waterfowl, especially egrets in addition to Cx. tritaeniorhynchus reared and herons, which are the reservoirs of JEV. from wild caught immature mosquitoes from Surveillance should be strengthened during Gorakhpur region (2009). JEV was detected the lean season to map breeding sites of in adult mosquitoes raised from wild caught vector mosquitoes in order to plan appropriate immatures and males of the following vector control measures. species: Cx. tritaeniorhynchus, Cx. vishnui, Rice fields being the most productive Cx. infula, Cx. gelidus, Cx. fuscocephala and breeding sites of Cx. tritaeniorhynchus, its Ma. indiana. Vertical transmission of JEV population dynamics is closely associated occurred in both hot and cool seasons, and with paddy cultivation.25,43 In Gorakhpur, a is thus regularly maintained in nature during single paddy crop is grown per year and the the non-transmission season as well.25 The majority of JE vectors show one peak a year, presence of a wide variety of vector species, i.e. in September. The occurrence of JE in the their different bionomics and vectorial capacity, region has therefore been closely associated therefore needs to be studied. The monitoring with this peak. Interestingly, two peaks of JE of virus infection in vector mosquitoes should vectors were recorded when double paddy form an essential component of a surveillance crops were cultivated in Tamil Nadu.44 system to assess monthly infection rates of vectors to predict disease outbreaks and to The steady spread of JE confirmed cases develop an early warning system in order to to new areas in India is a serious public take remedial measures in the State of Uttar health concern. Eastern districts of Uttar Pradesh. Pradesh have been highly endemic since 1978; in 1990, the disease spread to the The larval habitat of Cx. tritaeniorhynchus west, affecting neighbouring state Haryana is primarily low lying flooded areas containing for the first time,20 followed by Saharanpur grasses and rice paddies. Eastern Uttar and Muzaffarpur in in Pradesh is mainly a paddy growing area, 2003 and 2005, respectively. Climate change with clay soil and a very high water table. may also influence the migration patterns of The ecosystem comprising rivers, lakes, birds, which may result in the introduction of irrigation canals, ponds, reservoirs and rice JEV to new areas. Since little is known about fields favour vector breeding. Surface water reservoir bird migration patterns, this is an bodies are perennial breeding sources in UP important issue to investigate.2,45,46 Climate and provide a wintering and staging ground change is also affecting agricultural patterns, for a number of migratory waterfowls and a and the development of agriculture in the breeding ground for resident birds.42 They region has resulted in the vast expansion of also act as mother foci for vector mosquitoes. water bodies and irrigation systems which After the monsoon, vectors spread to other support mosquito breeding,25,38 and in turn water stagnation areas and rice fields. Thus JEV. breeding control with appropriate larvicides or using larvivorous fishin all permanent water Most of the JEV strains isolated in India bodies, before the start of monsoon and paddy belong to genotype III.18 However, JEV

392 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

genotype I has recently been reported early detection of JE outbreaks, and high in human patients from Gorakhpur.47Due coverage of eligible children with JE vaccine. to the evolution of new viral strains and/ Continuous monitoring of vector populations or re-emergence of older strains, children and virus infection rates in vector mosquitoes lack protective immunity. Although health would help to predict outbreaks more in management facilities have improved in UP advance and to implement proper control (CFR reduced), concrete steps should be taken measures. To this end, a combination of to combat JE, including the development of strategies involving other sectors and the more efficient surveillance in human, vector community would be required to instil the and vertebrate hosts and strengthening of sense of urgency needed to accelerate JE vaccination programme. In addition, first effective control of JE in the country. and secondary referral health facilities need to be reinforced for better management of JE Acknowledgements cases. We thank the State Health Authorities of Uttar Pradesh, India and BRD Medical College, Conclusion Gorakhpur for providing detailed information Japanese encephalitis is still a major health on Japanese encephalitis cases. The technical problem in eastern Uttar Pradesh, north India. help of Mr Ajay Negi, Consultant (Information No antiviral drug against JEV is available.32 Technology) at NVBDCP in the preparation Chemical control of vector populations of maps is gratefully acknowledged. We are with insecticides plays a marginal role in especially grateful to Shri P Gopalan, former control of the disease due to their exophilic Technician at NVBDCP, who helped to compile behaviour. The use of larvicides is limited as the data. vector mosquitoes prefer to breed in large water bodies. In some situations, e.g. during References outbreaks of JE, space-spraying may interrupt 1. Solomon T. Control of Japanese encephalitis–within the transmission cycle for a short time. our grasp? N Engl J Med. 2006; 355: 869–71. Wider issues, including current agricultural practices, water management systems, 2. Tobias E. Erlanger, Svenja Weiss, Jennifer Keiser, Jürg Utzinger, Karin Wiedenmayer. Past, present, and and human behaviour patterns, need to be future of Japanese Encephalitis 2009. Emerg Infect investigated. The most important long term Dis. 2009; 15(1): 1–7. strategy to fight flood-proneness is of rapid 3. Okuno T. An epidemiological review of JE. WH Stat increase in groundwater irrigation which will Q. 1978; 31(2): 120-33. not only lower water tables but also help 4. Webb J KG, Pereira SM. Clinical diagnosis of arthropod reduce the intensity of floods which will further borne type viral encephalitis in children in North impact on public health and to improve the Arcot district, Madras start. India. Indian J Med Sci. socioeconomic status of people in JE-affected 1956; 10: 572. areas. Awareness needs to be developed on 5. Bhardwaj M, et al. Serological study of JE in Deoria personal protection against mosquito bites and district of Uttar Pradesh. J Commun Dis. 1981; 13: the importance of early referral of cases. 96-101.

In summary, JE could be controlled with 6. Khare JB, Current status of Japanese encephalitis in Uttar Pradesh: Proceedings of the National effective surveillance systems, segregation of Conference on Japanese encephalitis. New Delhi: pigs, an integrated vector control approach, ICMR, 1982; 22-24.

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 393 Japanese encephalitis in Uttar Pradesh, India Roop Kumari et al.

7. Census of India. Table 1 Distribution of on Japanese encephalitis outbreak in Uttar Prdaesh population, sex ratio, density and decadal growth during 1988. J Com Dis. 1988; 20: 263-275. rate of population : 2011. New Delhi, 2011. 18. Parida M, Dash PK, Tripathi NK, Ambuj, Sannarangaiah http://www.censusindia.gov.in/2011-prov-results/ S, Saxena P, et al. Japanese encephalitis outbreak, data_files/india/table_1.pdf - accessed 23 April India, 2005. Emerg Infect Dis. 2006; 12: 1427-30. 2013. 19. Mathur A, Chaturvedi UC, Tandon HO, Agarwal AK, 8. Diverse Farming System in Flood Affected Areas of Mathur GP, Nag D, et al. Japanese Encephalitis Eastern Uttar Pradesh. http://agropedia.iitk.ac.in/ epidemic in Uttar Pradesh, India during 1978. Indian content/diverse-farming-system-flood-affected- J Med Res. 1982; 75 161-9. areas-eastern-uttar-pradesh - accessed 19 April 2013. 20. Kar NJ, Saxena VK. Some epidemiological characteristics of Japanese encephalitis in Haryana 9. National Vector Borne Diseases Control Programme. state of northern India. J Commun Dis. 1998; 30: Guidelines for surveillance of acute encephalitis 129–131. syndrome (With special reference to Japanese encephalitis). New Delhi: Directorate General of 21. Kar NJ, Bora D, Sharma RC, Bhattacharjee J, Datta Health Services, Ministry of Health and Family Welfare, KK and Sharma RS. Epidemiolgical profile of Japanese 2006. http://nvbdcp.gov.in/Doc/AESguidelines.pdf - encephalitis in , Uttar Pradesh, accessed 15 April 2013. 1982-1988. J Commun Dis. 1992; 24(3): 145-149.

10. Saxena NC. Wells and Welfare in the Ganga Basin: Essay 22. Karve JB. Current status of Japanese encephalitis in on Public Policy and Private Initiative: Tushaar Shah. Uttar Pradesh. Proceeding of the National Conference New Delhi: Planning Commission, Government of on Japanese encephalitis. New Delhi: ICMR, 1982. India, 2009. http://planningcommission.nic.in/ pp. 22-24. reports/articles/ncsxna/index.php?repts=wells. 23. Srivastava VK, Sinha NK, Singh A. Chandra Ramesh.. htm#null - accessed 19 April 2013. Japanese Encephalitis situation in Gorakhpur Division, 11. Kumar R, Mathur A, Kumar A, Sharma S, Saksena PN, UP.J Commun Dis. 2003; 35(1): 56-58. Chaturvedi UC. Japanese encephalitis an important 24. Saxena SK, Niraj Mishra, RakhiSaxena, Maneesh cause of acute childhood encephalopathy in Lucknow, Singh, AshaMathur, Niraj Mishra1, Rakhi Saxena, India. Postgraduate Med J. 1988; 64: 18-22. Maneesh Singh, AshaMathur.Trend of Japanese 12. Khandelwal KK, Chopra JK. Agriculture. In: Unique encephalitis in North India. J Infect Dev Ctries. 2009; quintessence of general studies. New Delhi: Unique 3(7): 517-530. Publishers 2006; 248-76. 25. Kanojia PC, Shetty PS, Geevarghese G. A long-term 13. Endy TP, Nisalak A. Japanese encephalitis virus: studyon vector abundance & seasonal prevalence in ecology and epidemiology. Cur Top Microbiol relation to the occurrence of Japanese Encephalitisin Immunol. 2002; 267: 11-48. Gorakhpur district, Uttar Pradesh. Indian J Med Res. 2003; 117: 104-10. 14. Joshi AB, Banjara MR, Bhatta LR,Wierzba T. Status and trend of Japanese encephalitis epidemics in 26. Buescher EL, Scherer WF, Rosenberg MZ, Gresser I, Nepal: a Five-Year retrospective review. Journal of Hardy JL, Bullock HR.Ecologic studies of Japanese Nepal Health Research Council. 2004; 2: 59-64. encephalitis virus in Japan.II.Mosquito infection. Am J Trop Med Hyg. 1959; 8: 651-64. 15. Barzaga NG. A review of JE cases in the Philippines (1972-1985) Southeast Asian. J Trop Med Public 27. Roy AR, Tandon SK, Agarwal, Banerjee G. Health. 1989; 20(4): 587-92. Seroprevalence of Japanese encephalitis virus infection in Lucknow. Uttar Pradesh. Indian J Med 16. Chakrabarty S, Sacena VK,Bhardwaj Mohan. Res. 2006; 124: 211-212. Epidemiological investigation of Japanese Encephalitis outbreak in Gorakhpur and Deoria districts of Uttar 28. Thakare JP, Gore MM, Risbud KR, Banerjee K, Ghosh Pradesh 1985. J Com Dis. 1986; 18 (2): 103-108. SN. Detection of virus specific IgG subclasses in Japanese encephalitis patients. Indian J Med Res. 17. Narshimam, MVVL, Rao Krishna C, Bendle MS, Yadava 1991; 93: 271–276. RL, Johri YC, Pandey RS. Epidemiological investigation

394 WHO South-East Asia Journal of Public Health 2012;1(4):374-395 Roop Kumari et al. Japanese encephalitis in Uttar Pradesh, India

29. Neeru Gupta, Shah Hussain, Ramesh Lal, Das BP, 39. Banerjee K, Mahadev PVM, Ilkal MA, Mishra AC, Ventatesh S, Chatterjee K. Epidemiological Profile of Dhanda V, Modi GB, et al. Isolation of Japanese Japanese Encephalitis outbreak in Gorakhpur, UP in encephalitis virus from mosquitoes collected in 2004. J CommunDiseases. 37(2): 145-149. Bankura district, West Bengal, during October 1974 to December 1975. Indian J Med Res. 1979; 69: 30. Umenai RT, Krzysko TA, Bektimirov FA. Japanese 201-5. encephalitis: current worldwide status. Bulletin of the World Health Organization. 1985; 63 (4): 625-631. 40. Pant U, Ilkal MA, Soman RS, Shetty PS, Kanojia PC, Kaul HN. First isolation of Japanese encephalitis virus 31. King N J, Getts D R, Getts M T, Rana S, Shrestha B from the mosquito, Culex tritaeniorhynchus Giles, and Kesson A M 2007 Immunopathology of flavivirus 1901 (Diptera: Culicidae) in Gorakhpur district, Uttar infections; Immunol. Cell Biol. 85 33–42. Pradesh. Indian J Med Res. 1994; 99: 149-51. 32. Saxena V, Dhole TN. Preventive strategies for frequent 41. Gajanana A, Rajendran R, Thenmozhi V, Samuel PP, outbreaks of Japanese encephalitis in Northern India. Tsai TF, Reuben R. Comparative evaluation of bioassay Biosc. 2008; 33: 505–514. and ELISA for detection of Japanese encephalitis virus 33. Reuben R, Tewari SC, Hiriyan J, Akiyama J. Illustrated in field collected mosquitoes. Southeast Asian J Trop keys to species of Culex (Culex) associated with Med Public Health. 1995; 26: 91-7. Japanese encephalitis in Southeast Asia (Diptera: 42. Tripathy SC. Can Bakhira bird sanctuary safeguard Culicidae). Mosquito Systematics. 1994; 26: 75- the purple moorhens? Curr Sci. 2004; 86: 367-8. 96. 43. Sunish IP & Reuben R. Factors influencing the 34. Leake CJ, Ussery MA, Nisalak A, Hoke CH, Andre RG, abundance of Japanese encephalitis vectors in rice BurkeDS. Virus isolations from mosquitoes collected fields in India. Medical and Veterinary Entomology. during the 1982 Japanese encephalitis epidemic in 2001; 15: 381-392. northern Thailand. Trans R Soc Trop Med Hyg. 1986; 80: 831-7. 44. Gajanana A, Rajendran R, Samuel PP, Thenmozhi V, Tsai TF, Kimura - Kuroda J,et al.Japanese encephalitis 35. Amersinghe FP, Peiris JS, Karunaratne SH, Amersinghe in South Arcot district, Tamil Nadu, India : a three PH. Epidemiology of the 1987 Japanese encephalitis – year longitudinal study of vector abundance and outbreak in Anuradhapura area.II. Entomological infection frequency. J Med Entomol. 1997; 34: aspects. Proc Sri Lanka Med Assoc. 1988; 101: 651-9. 22-3. 45. Hsu SM, Yen AM, Chen TH. The impact of climate 36. Reuben R. Studies on the mosquitoes of North Arcot on Japanese encephalitis. Epidemiol Infect. 2008; District, Madras state, India. Part 5. Breeding places 136: 980–7. of the Culex vishnui group of species. Journal of Medical Entomology. 1971; 8: 363–366. 46. Yasuoka J, Levins R. Ecology of vector mosquitoes in Sri Lanka–suggestions for future mosquito control in 37. Service MW. Problems of vector-borne disease and rice ecosystems. Southeast Asian J Trop Med Public irrigation projects. Insect Sci Appl. 1984; 5: 227- Health. 2007; 38: 646 -57. 231. 47. Fulmali PV, Sapkal GN, Athawale S, Gore MM, Mishra 38. Arunachalam N, Samuel PP, Paramasivan R, AC, Bondre VP. Introduction of Japanese encephalitis BalasubramanianA, Tyagi BK. Japanese encephalitis virus genotype I, India. Emerg Infect Dis. 2011, 17, in Gorakhpur Division, Uttar Pradesh. Indian Journal 319-321. of Medical Research. 2008; 128(6): 775-7.

WHO South-East Asia Journal of Public Health 2012;1(4):374-395 395 Original research

A study on delay in treatment of kala-azar patients in Bangladesh

Syed M Arifa, Ariful Basherb, Mohammad R Rahmanc, Mohammad A Faizd,e

Visceral leishmaniasis (kala-azar) continues to be a major rural public health problem in Bangladesh. A cross-sectional study was carried out in two subdistricts of Mymensingh district from January 2006 to June 2007 to evaluate the delay kala-azar treatment. Suspected patients who attended to out patient department (OPD) were subjected to a dipstick test (RK39) for kala-azar. Sixty five from Bhaluka and 60 positive patients from Gafargaon subdistrict were enrolled. Most of the patients (80%) first visited nonqualified private practitioners, while only 15.2% consulted registered doctors. Fifty per cent were referred to the Upazilla health complex (UZHC) by the family members or relatives. About 49% and 43% patients required third and second health-care providers for kala-azar treatment, respectively. Patient delay ranged from 2 to 30 days; median 4 (IQR 3 to 7 days), the system delay ranged from 0 days to 225 days; median 54 (IQR 40–66 days). Residential status (p value <0.05) had impact on patient delay. Educational status and number of treatment providers had impact on system delay (p<0.05). System delay rather than patient delay is the important weakness of the kala-azar control programme in Bangladesh. Residence in rural areas, low educational background and treatment providers are associated with these delays. A proper educational programme may reduce the delay.

Key words: Visceral leishmaniasis, sand fly bite, health-seeking behaviour

Introduction The Indian subcontinent has experienced a case–fatality rates in excess of 10% are major resurgence of kala-azar (KA) which not uncommon.7 It is associated with about accounts for 60% of all reported cases of 2.4 million disability–adjusted life years and kala azar in the world.1,2 Kala-azar is caused around 70 000 deaths per year.8 by infection in the reticulo endothelial system Kala-azar is one of the major public health by the protozoan Leishmania donovani problems in Bangladesh, and the disease acquired through sand fly bites.3,4,5 It carries has been endemic for many decades. In the a high mortality ranging from 80% to 100% late 1970s kala-azar emerged sporadically in untreated cases.6 Even with treatment, in Bangladesh. During 1981–1985 only

a Medicine Department, Dhaka Medical College Hospital, Dhaka, Bangladesh b Surya Kanta Kala Azar Research Centre, Mymensingh, Bangladesh c Medicine Department, Shahid Sarwardi Medical College Hospital, Dhaka, Bangladesh d Director General of Health Services, Dhaka, Bangladesh e Dev Care Foundation, Dhaka, Bangladesh. Correspondence to Ariful Basher (e-mail: [email protected])

396 WHO South-East Asia Journal of Public Health 2012;1(4):396-403 Syed M Arif et al. A study on delay in treatment of kala-azar patients in Bangladesh

8 upazilas (subdistrict) reported kala-azar for kala azar could be divided into “patient cases, which was increased to 105 upazilas delay” (time from symptoms recognition to in 2004.9 In the past few years, the kala-azar initial medical consultation) and “system situation had assumed epidemic proportions delay” (time from first medical consultation with the number of reported cases increasing to final treatment).11 from 3978 in 1993 to 8505 in 2005.9 But, in We approached 125 new patients of kala 2010, the number of cases reported was less azar attending at the UZHC during the study than 4000 and there was only 1 death.10 The period. The sample size was determined actual number of cases might be a little higher with a single population proportion formula due to incomplete reporting systems and late by assuming that 50% of the patients will detection of PKDL cases. The hyperendemic come early for kala azar treatment at a UZHC zones in Bangladesh for kala-azar are mostly (to obtain minimum sample size) with 95% in Mymensingh district. confidence interval. Social behaviour and economic Consecutive confirmed 65 patients from consequences of kala-azar patients are still not Bhaluka and 60patients from Gafargaon well known, despite long standing occurrence upazilla were enrolled in this study. Patients of the disease and the implementation of attending to the out patient department various control programmes. This is due (OPD) or in-patient facility with history of to inadequate information and the delay in fever (Temp ≥38 °C) for at least for 2 weeks diagnosis, social and community factors that with one or more of the followings criteria: influence care-seeking behaviour and the (a) anaemia (5

WHO South-East Asia Journal of Public Health 2012;1(4):396-403 397 A study on delay in treatment of kala-azar patients in Bangladesh Syed M Arif et al.

Ethical clearance was obtained from the of a sandfly. About 30.0% had family history ethical committee of Bangladesh Medical of kala-azar. Most patients (85.6%) heard Research Council (National Ethics Committee). about the disease from hospital. According to Prior to study enrollment, written informed the interviewees, about 60.8% always used consent was obtained from each participating mosquito nets at bedtime and 1.6% never patient. Detailed study-related information used one. Approximately 96.8% fell asleep was read out and explained in local language. on a bed rather than the floor, and 56% slept Research assistants supplied a copy of the during daytime (Table 2). handout to the patient/his/her guardian. Signed written informed consent was obtained Type of facility first visited from the patient/guardian. Finger impressions One hundred and twenty five patients were obtained from participants, who could not consulted with various types of health- sign. For subjects below 18 years, consent was care providers as the first place for seeking taken from the eligible guardian. All patients help with most visiting private health-care were ensured treatment with Miltefosine providers. About half (49.6%) went to village tablets. doctors, 27.2% to pharmacy personnel, 3.25% to the ‘Kabiraj’ (traditional health provider) Results and 2.4% to a medical assistant. Only 15.2% visited registered doctors and 2.4% directly General characteristics of patients reported to the Upazilla Health Complex About 60% study subjects were less then 20 (UZHC). Majority (85.6%) attended at UZHC years of age with only of the 15.1% above as their third or fourth order treatment 45 years (range 4–60 years). A majority location. Many patients were referred to 64% (80) were male and almost all (98.4%) the UZHC by immediate family members or were from the Muslim community. Most relatives (49.6%) for definite treatment. MBBS patients were from rural areas (80.8%) and (registered) doctors referred 21% of patients half (42.2%) were married; 34.4% were to UZHC. farmers and 32.8% students. About half had monthly income < US$ 50, while 8% had Determinants of pattern of delay income >US$ 100.00. Twenty four per cent had no formal education. Those educated Patient delay ranged from 2 to 30 days with up to the fifth grade were 52.8%, and only a median of 4 (IQR 3 to 7) days. The system 23.2% were graduates. Splenomegaly was delay ranged from 0 to 225 days, with a present in 87.2% patients, 8.8% patients were median duration of delay 54 (IQR 40 to 66) moderately anaemic while the rest were mildly days. The median total delay was 55 (IQR 35 anaemic (Table 1). to 68) days. Seventy three per cent of cases had patient delay below 7 days (73%). System delay over 2 months were experienced by Knowledge about the disease 28% of patients (Table 5). Residential status Only 15.2 % had no idea about the disease had significant statistical association with (Table 2). Almost all the respondents (84.8%) patient delay. Educational status, number had heard the name kala-azar, and most of treatment providers and first healthcare (92.8%) knew that the disease was completely provider had statistical association with curable but only one person knew that it is System delay (p<0.05). (Table 3) an infectious disease transmitted by the bite

398 WHO South-East Asia Journal of Public Health 2012;1(4):396-403 Syed M Arif et al. A study on delay in treatment of kala-azar patients in Bangladesh

Table 1: General characteristics of the study population

Characteristics (n=125) Catagories Number Percentage Sex Male 80 64 Female 45 36 Age (years) 1–10 44 35.2 11–20 32 25.6 21–30 18 14.4 31–40 20 16.0 41–50 7 5.6 51–60 4 3.2 Educational status Illiterate 30 24 Up to 5th classes 66 52.8 Up to 12th classes 29 23.2 Religion Muslim 123 98.4 Hindu 02 1.6 Occupation Agriculture worker 43 34.4 Business 7 5.6 Daily labour 6 4.8 Skilled labour 7 5.6 Housewife 10 8.0 Service 3 2.4 Student 41 32.8 Unemployed 8 6.4

Household income per month (US$) <38.46 64 51.2 <64.10 47 37.6 <128.21 14 12.2 Residential status Urban 1 0.8 Semiurban 23 18.4 Rural 101 80.8 Marital status Married 23 42.2 Single 39 32.2 Divorced 1 0.08 NA 32 25.6 Past family history of kala-azar Yes 38 30.4 No 87 69.6 Clinical anaemia Moderate anaemia 11 8.8 (Hb-6–10g/dl) Mild anaemia 113 90.4 (Hb-10–12g/dl) No anaemia 1 0.8 Splenomegaly Yes 109 87.2 No 16 12.8

WHO South-East Asia Journal of Public Health 2012;1(4):396-403 399 A study on delay in treatment of kala-azar patients in Bangladesh Syed M Arif et al.

Table 2: Knowledge of respondents about kala-azar in Baluka and Gafargaon

Items Respondents Number Percentage Heard the name kala-azar Yes 106 84.8 No 19 15.2 Kala-azar is an infectious disease, Yes 1 0.8 transmitted by sand fly bite No 124 99.2 Complete cure of the disease is possible Yes 116 92.8 No 1 0.8 Not known 8 6.4 How kala-azar is known about FM/Neighbour/ 3 2.4 Relative Village doctor 1 0.8 MBBS doctor 14 11.2 Hospital 107 85.6 Frequency of using mosquito net Always 76 60.8 Often 32 25.6 Never 2 1.6 Always 76 60.8 Sleeping in a bed Bed 121 96.8 Floor 4 3.2 Habit of day-time sleeping Regular 2 1.6 Occasional 38 30.4 Rarely 70 56.0 Not at all 15 12.0

Table 3: Relationship between sociodemographic factors and patient delay and system delay (multiple linear regression analysis)

Factor Patient delay System delay (95% confidence (95% confidence β-coefficient (p value) β-coefficient (p value) interval) interval) Age –0.086 (–0.918 – 0.345) 0.347 0.045 (–0.053–0.127) 0.420 Sex 0.065 (–0.031–0.028) 0.475 0.008 (–0.003–0.003) 0.932 Education 0.002 (–0.140–0.142) 0.986 –0.200 (–0.043–(–0.003) 0.028 Religion –0.052 (–0.007–0.004) 0.572 0.017 (0.000–0.001) 0.850 Marital status –0.049 (–0.066–0.038) 0.594 –0.050 (–0.010–0.005) 0.585 Monthly –0.048 (–96.963–55.661) 0.593 –0.146 (–20.026–2.060) 0.110 income Residential –0.269 (–0.045–(–0.010) 0.003 0.029 (–0.002–0.003) 0.743 status No. treatment NA –0.277 (0.003–0.014) 0.002 providers First care NA 0.248 (0.005–0.023) 0.004 provider

NA – not applicable

400 WHO South-East Asia Journal of Public Health 2012;1(4):396-403 Syed M Arif et al. A study on delay in treatment of kala-azar patients in Bangladesh

Discussion views of inadequacy of the health system. Only 2.4% rated public health facilities as Understanding the treatment-seeking their first choice. The care-seeking pattern for behaviour of communities regarding kala-azar fever were more likely to be managed initially can be the key to the success of an elimination at low cost. Multiple care-seeking events and programme launched by the Government as switching different types of providers were not Bangladesh along with India and Nepal have uncommon. More than 50% patients required set the target of reducing the cases below three or four health-care providers for their 1 per 10 000 by the year 2015.1 definitive treatment. Our data showed that most patients were Sociodemographic factors have long been below 40 years (82.2%) which is consistent associated with delay in management of with the studies.12,13 Awareness about the kala-azar patients. The results indicate the signs and symptoms of a disease can prompt obvious existence of delay in diagnosis and patients to seek early treatment. However, in treatment of kala-azar patients. System delay our study population, knowledge about the was the most frequent type of delay and the signs and symptoms of the disease was very greatest contributor to total delay. Distance poor despite the fact that the disease has from patients’ home to a health facility is also been endemic for such a long time. Fever and important as it affects health-care seeking splenomegaly were present in almost every and follow-up of diagnostic procedures. As kala-azar patient in Bangladesh. The majority most patients were from rural areas, distance of subjects lacked knowledge about the between their homes to the health-care service involvement of humans in the transmissibility had contributed to the delay in the diagnosis. and the infectious nature of the disease. There This is consistent with the studies done in TB was an almost complete lack of knowledge patients in Uganda and Ethiopia.18,19 Patients about the transmitting vector (99.2%), who tried their first consultation at rural areas which is a matter of concern for adaptation experienced a longer patient delay than those of preventive measures against the disease. who first sought care in urban facilities. Our Although most people were aware of the data suggested that the first health-care protective role of mosquito nets (Table 2), provider plays an important role in determining and many of households used bednets during the duration of system delay. A major finding sleeping, their acceptability and efficacy of this study was that about 50% of the against kala-azar need more study. study subjects who soughted final treatment The main health-seeking behaviour with at UZHC were referred by their immediate the onset of fever was to visit a local health- family members or relatives. Village doctors care provider. People in the study community referred only 5.6% patient although they generally consult traditional healers due to play a significant role as first-care providers. a hierarchical system in their households. This exposes a very weak point in our health The majority of patients seeked care at the care system. Therefore, by raising awareness village doctor or personnel at the pharmacy among health-care providers and field staff, (76.8%) or private doctor (MBBS) (15.2%). we can strengthen the referral system and This indicates that the traditional healers are thus shorten system delay. Eighty per cent of considered important in the treatment of kala- patients soughted medical care from village azar. Only a small number preferred public doctors, pharmacies and ‘kabiraj’, so health health facilities for treatment because of their campaigns at the community level should be

WHO South-East Asia Journal of Public Health 2012;1(4):396-403 401 A study on delay in treatment of kala-azar patients in Bangladesh Syed M Arif et al.

conducted to raise awareness and to facilitate to be substantially different in other parts, early referral. Educational status also had because the study area is the oldest focus of significant relationship with system delay kala-azar in Bangladesh and the population is (p<0.05). This may be due to fact that most expected to have maximum awareness about of the patients of this study reside in rural the disease. Even after such a prolonged and areas (80.8%) (Table 1). They were illiterate incessant disease transmission in the area, and had to travel a long distance to consult this lack of knowledge, indifferent attitude, with a qualified doctor to seek medical care and incorrect practices are pointers about the and most rural people had no easy access to poor commitment of health–policy planners Government facilities. However educational for the disease. For optimal utilization, public status had no effect on delay to initiation of health facilities should be fully equipped to treatment (patient delay) as established in deal with cases of kala-azar. other studies.20,21 Other factors such as age, sex, marital status, economic status and religion were not associated with patient delay and system delay (Table 3), although other Reference studies had confirmed prolonged delays with 1. Bangladesh, Ministry of Health and Family Welfare. initiation of treatment.18-23 Kala azar elimination programme, MIS: technical report Dhala: Disease Control Unit, DGHS, 2010.

Further in this study, poor knowledge of 2. World Health Organization, Regional Office for South the study subjects about symptoms, infectious East Asia. Status of Kala-azar in Bangladesh, Bhutan, nature, and mode of transmission emphasizes India and Nepal: a regional review update. New Delhi: the need for health education campaigns. WHO-SEARO. http://www.searo.who.int/LinkFiles/ Kala_azar_kala-status2008Webpagefeb2009.pdf - A system delay in treatment of kala- accessed 18 April 2013.

azar appears to be the major determinant 3. Todd WTA. Infection and immune failure; of a treatment delay in the selected highest Leishmaniasis. In: Haslett C, Chilvers ER, Boon kala-azar prevalent area of Bangladesh. NA, and Colledge NR, eds. Davidson’s principle and The current National Kala-azar Elimination practice of medicine. 19th edn. London: Churchil Programme should address this issue in Livingstone, 2002. pp. 66-8. attempts to improve patients’ access to early 4. Bryceson ADM. Leishmaniasis. In: Weatherall DJ, treatment. Residence in rural areas and Ledingham GG, Warrell DA, eds. Oxford text book of medicine. 3rd edn. New York: Oxford University treatment providers have contributed to both Press, 1996. pp. 899-907. the patient delay and the system delay. The elimination programme should emphasize 5. Sundar S, Jha TK, Thakur CP, Engel J, Sindermann H, Fischer C, et al. Oral miltefosine for Indian visceral training and continued education program for leishmaniasis. N Engl J Med. 2002; 347: 1739-46. the First level health-care providers, many of 6. Bern C, Hightower AW, Chowdhury R, Ali M, Amann whom are private health-care providers, in J, Wagatsuma Y, et al. Risk factors for Kala azar in addition to the general BCC (behaviour change Bangladesh. Emerg Infect Dis. 2005, May; 11(5): communication) campaign, to effectively 1-15. reduce the treatment delay in the early case 7. Indu BA, Bern C, Costa C, Akter T, Chowdhury R, Ali management of kala-azar. M, et al. Visceral leishmaniasis: consequences of a neglected disease in a Bangladeshi community. Am Although it is difficult to extrapolate J Trop Med Hyg. 2003; 69(6): 624-8. these findings, the level of KAP (knowledge– attitude–practice) about Kala-azar is not likely

402 WHO South-East Asia Journal of Public Health 2012;1(4):396-403 Syed M Arif et al. A study on delay in treatment of kala-azar patients in Bangladesh

8. Desjeux P. Leishmaniasis: current situation and new 16. Thongsuksai P, Chongsuvivatwong V, Sriplung H. perspectives. Comp Immunol Microbiol Infect Dis. Delay in breast cancer care: A study in Thai women. 2004; 27: 305-18. Med Care. 2000 Jan; 38(1): 108-114.

9. Bangladesh, Ministry of Health and Family Welfare. 17. Benjamin SCU, Obinna EO. Socioeconomic differences Kala azar elimination programme. National guideline and health seeking behaviour for the diagnosis and and training module for kala azar elimination in treatment of malaria: a case study of four local Bangladesh. 1st edn. Dhaka: Disease Control Unit, government areas operating the Bamako initiative DGHS, 2008. programme in south-east Nigeria. International Journal for Equity in Health. 2004; 3:6. 10. World Health Organization, Regional Office for South East Asia. Regional strategic framework for 18. Mpungu SK, Karamagi C, Mayanja KH. Patient and elimination of Kala Azar from the South-East Asia Health service delay in pulmonary tuberculosis Region (2005-2015). New Delhi: WHO-SEARO, patients attending a referral hospital: a cross- 2012. sectional study. BMC Public Health. 2005; 5:122.

11. Plan T. Giao, Peter J. de Varies, Tran Q. Binh, Nguyen 19. Meaza D, Bernt L, Yemane B. Patient and Health V. Nam, Piet A. Kager. Early diagnosis and treatment service delay in the diagnosis of pulmonary of uncomplicated malaria and patterns of health tuberculosis in Ethiopia. BMC Public Health. 2002; seeking in Vietnam. Tropical Medicine & International 2: 23. Health. 2005 September; 10(9): 919-25 20. Lienhardt C, Rowley J, Manneh K, Lahai G, Needham 12. Caryn B, Anand BJ, Shambhu NJ, Muralal LD, Allen H D, Milligan P, et al: Factors affecting time delay to et al. Factors associated with visceral leishmaniasis treatment in a tuberculosis control programme in in Nepal: Bed-Net use is strongly protective. Am. J. a sub-Saharan African country: the experience of Trop. Med. Hyg. 2000; 63(3,4): 184-188. The Gambia. Int Tuberc Lung Dis. 2001; 5(3): 233- 239. 13. Jharna Bhattachryya, A K Hati. A selection of Essays: The Adverse affects of Kala azar (Visceral 21. Pronyk PM, Makhubele MB, Hargreaves JR, Tollman Leishmaniasis) in women. IDRC Books free online. SM, Hausler HP: Assessing health seeking behaviour among tuberculosis patients in rural South Africa. Int 14. Ahluwalia IB, Bern C, Costa C, Akter T, Chowdhury Journ Tuberc Lung Dis 2001, 5(7): 619-627. R, Ali M, Alam D, Kenah E, Amann J, Islam M, Wagatsuma Y, Haque R, Breiman RF, Maguire JH. 22. Oola J. Factors influencing delayed diagnosis of Visceral leishmaniasis: Consequences of a neglected tuberculosis in Mukono district Uganda. Institute of disease in a Bangladeshi community. Am J Trop Med Public Health, 2001. Hyg. 2004; 69: 624–628 23. Lawn SD, Afful B, Acheampong JW: Pulmonary 15. Shri P.S, Reddy DCS , Mishra R N, SUNDAR S. tuberculosis: diagnostic delay in Ghanaian adults. Int Knowledge, Attitude, and Practices elated to Kala J Tuberc Lung Dis. 1998, 2(8): 635-640. Azar in a rural area of Bihar state , India Am. J. Trop. Med. Hyg. 75(3), 2006; 505–508

WHO South-East Asia Journal of Public Health 2012;1(4):396-403 403 Original Research

Factors associated with high prevalence of pulmonary tuberculosis in HIV-infected people visiting for assessment of eligibility for highly active antiretroviral therapy in Kathmandu, Nepal

Bishnu R Tiwaria, Surendra Karkib, Prakash Ghimirec, Bimala Sharmad, Sarala Mallae

Background: Tuberculosis is the leading cause of deaths among HIV patients. In this study, we estimated the prevalence of pulmonary tuberculosis (PTB) and identified the factors/co-morbidities associated with active PTB in HIV-infected people visiting the national public health laboratory to assess their eligibility to receive highly active antiretroviral therapy.

Methods: A cross-sectional study was conducted to measure the prevalence of pulmonary tuberculosis. Data on probable risk factors in patients with and without PTB were compared, calculating the odds ratio as a measure of association. Factors showing significant association in univariate analyses were included in a stepwise backward logistic regression model to adjust for confounding.

Results: The prevalence of pulmonary tuberculosis was 32.4 % (95% confidence interval (CI) 30.25–34.56). In the univariate analysis, patients with PTB were more likely to be older, married, and have a longer duration since the diagnosis of HIV, diarrhoea, parasitic infection, lower CD4 T-cell counts, and lower CD4/CD8 ratio. However, the backward stepwise logistic regression revealed that only the CD4 T-cell count < 200/µL (AOR 11.69, 95% CI 6.23–21.94), CD4 T-cell count 200–350/µL (AOR 2.52, 95% CI 1.30–4.89), diarrhoea (AOR 2.77, 95% CI 1.78–4.31), parasitic infection (AOR 3.34, 95% CI 2.02–5.50) and ‘sex with partner’ as probable modes of transmission (AOR 0.44, 95% CI 0.20–0.93) were independently associated with pulmonary tuberculosis.

Conclusion: A high prevalence of pulmonary tuberculosis was observed. Participants with tuberculosis were significantly more likely to have lower CD4 counts, diarrhoea, and parasitic infections. HIV treatment programmes should consider these factors for better outcomes.

Key words: HIV, CD4, pulmonary tuberculosis, diarrhoea, Nepal

a School of Health and Allied Sciences, Pokhara University, Kaski, Nepal. b Department of Epidemiology and Preventive Medicine, Infectious Disease Epidemiology Unit, Monash University, Melbourne, Australia. c Central Department of Microbiology, Tribhuvan University, Kathmandu, Nepal. d Institute of Medicine, Department of Public Health, Tribhuvan University, Kathmandu, Nepal. e National Public Health Laboratory, Kathmandu, Nepal. Correspondence to Surendra Karki (e-mail: [email protected]).

404 WHO South-East Asia Journal of Public Health 2012;1(4):404-411 Bishnu R Tiwari et al. Factors associated with high prevalence of pulmonary tuberculosis in HIV-infected people in Kathmandu

Background is the largest referral centre in the country, equipped with an automated CD4 T-cell count Tuberculosis (TB) is caused by bacteria facility. Although the facility is located in belonging to Mycobacterium tuberculosis the capital city, the participants came from complex. About one third of the world’s various parts of the country as a CD4 T-cell population is latently infected with TB bacteria. count service was not available outside of HIV infection leads to immune deterioration, Kathmandu until the middle of 2008. Nepal thus providing a platform for the activation of is a landlocked country in South Asia with a latently infected TB.1 poorly functioning health system. During our TB is the leading cause of deaths in HIV- study period, the country had widespread and infected people in resource-limited settings. violent civil war and high political instability. About one third of HIV-infected people are co-infected with TB, most of them living in Study design and data collection low- and middle-income countries.2 In Nepal, We conducted a cross-sectional study to 80 000 people have active TB infection at any measure the proportion of active pulmonary one time, with about 40 000 new cases and tuberculosis and analysed the data comparing 5000–7000 deaths per year.3 As many as 11% patients with PTB and without PTB, to identify of newly detected TB cases are infected with the associated factors/co-morbidities. A total HIV and up to 23% of newly detected HIV cases of 1807 participants were included in the are co-infected with TB.4,5 A high prevalence of study from March 2005 to December 2008. HIV/TB co-infection is described elsewhere.6-9 The median age of participants was 30 years Estimating the magnitude of pulmonary (interquartile range, 25–35 years), 66% of tuberculosis (PTB), and the factors associated whom were male, and 90% were married. with it in HIV-infected people, may be useful The study was explained to participants, who at policy level as well as for management were assured about the confidentiality and of individual patients for better outcomes. anonymity of the collected information. An In resource-limited settings, HIV is often informed verbal consent was obtained from diagnosed in late stage of the disease and in all the volunteers. association with other co-morbidities related to deterioration of the immune system. In this Definition of study variables study, we aimed to estimate the prevalence of Pulmonary tuberculosis was defined as a PTB and identify the factors associated with participant whose sputum specimen was it in HIV-infected people visiting the national positive for acid fast bacilli (AFB) by microscopic public health laboratory (NPHL) for assessing examination at NPHL; or who had a prior eligibility to highly active antiretroviral therapy medical record of being diagnosed with PTB (HAART). on the basis of other laboratory investigations (sputum smear examination or microbiological Methods culture) from other health-care facilities. Setting and study population Participants with tuberculosis other than PTB were excluded from the study. Participants This study was conducted in people infected with pulmonary tuberculosis already receiving with HIV, attending NPHL for CD4 T-cell count HAART were also excluded from the analysis. in Kathmandu, Nepal, primarily for the purpose The status of diarrhoea was ascertained by of assessing their eligibility for HAART. NPHL self-reporting of participants having loose

WHO South-East Asia Journal of Public Health 2012;1(4):404-411 405 Factors associated with high prevalence of pulmonary Bishnu R Tiwari et al. tuberculosis in HIV-infected people in Kathmandu

stools 3 or more times a day. Information interaction by using a test of homogeneity. on parasitic infection was collected from the The chi-square test was used to test statistical examination of stool specimens from a random significance in categorical variables and subset of the study sample. Information about the Wilcoxon ranksum test was used for age, sex, possible mode of transmission, and continuous variables. A p-value of < 0.05 was date of HIV diagnosis was collected from a considered statistically significant. patient register maintained at NPHL. Ethics Laboratory investigation The study was approved by the Ethics The volunteers were instructed on appropriate Committee of Tribhuvan University, sputum specimen collection. Briefly, all sputum Kathmandu. Participation was fully voluntary smears were stained with gram stain and based oninformed verbal consent. Participants observed under low power (10×) objective. diagnosed with TB or being infected with Samples having less than 10 epithelial cells intestinal parasites were referred appropriately and more than 25 pus cells per low-power for treatment. field were accepted and those not meeting the criteria were requested to submit a repeat sample. Sputum specimens collected Results and stained following AFB staining methods The median time since the first diagnosis of were examined microscopically following the HIV was 23 weeks (inter-quartile range, 7–36 standard protocol. At least 200 fields were weeks). The major modes of transmission evaluated before reporting negative.10 reported were injecting drug use (41.2%) and commercial sex (36.6%). The median CD4 and Microscopic examination of stool samples CD8 T-cell count was 257/µL (inter-quartile was performed by wet mount, formal-ether range, 132–419) and 1010/µL (inter-quartile sedimentation technique and modified acid range, 657–1432), respectively. About 86% 10 fast staining. Three millilitres of blood from of the participants had a CD4/CD8 cell ratio each participant was collected for CD4 T-cell of ≤0.5. count, performed using flow cytometry.11 Of the 1807 participants, 32.4% (95% CI, 30.25–34.56) were ascertained to Statistical analysis have pulmonary tuberculosis infection. The Data were entered in a Microsoft Excel proportion of TB was not significantly different spreadsheet and statistical analysis was among males (33.8%) and females (29.5%). performed by STATA (version 10; StataCorp, Among patients with PTB, 82.7% had a CD4 College Station, United States of America). T-cell count of less than < 200/µL and 73.9% The measure of association between different had a CD4/CD8 T-cell ratio of < 0.25. In the explanatory variables and TB infection univariate analysis, patients with PTB were (outcome variable) was expressed in terms more likely to be older, married, and have a of odds ratio (OR) with corresponding 95% longer duration since the diagnosis of HIV, confidence interval. Factors significantly diarrhoea, parasitic co-infection, lower CD4T- associated with TB infection at p<0.05 cell count, and lower CD4/CD8 ratio. The in univariate analysis were adjusted for distribution of factors across patients with and confounders in a backward stepwise multiple without PTB, and the measure of association logistic regression model. The independently in terms of corresponding unadjusted odds associated variables were tested for statistical ratios, are shown in Table 1.

406 WHO South-East Asia Journal of Public Health 2012;1(4):404-411 Bishnu R Tiwari et al. Factors associated with high prevalence of pulmonary tuberculosis in HIV-infected people in Kathmandu

Table 1: Factors associated with tuberculosis in univariate analysis

Variable Patients Patients Unadjusted odds P-value with PTB without PTB ratio (95% CI) (%) (%) Number of participants 585 1222 - - Median age (Q1, Q3) 32(28, 36) 29(24, 35) - <0.001 Marital status 561(96.1) 1071(87.6) 3.43(2.17–5.65) <0.001 Male 403(68.89) 788(64.48) 1.21(0.98–1.51) 0.06 Median duration in weeks since 24(12, 36) 21(6–36)* - 0.01 HIV diagnosis (Q1, Q3) Diarrhoea 271(59.2) 117(12.8) 9.92(7.51–13.11) <0.001 Parasitic infection 112(64.7) 38(8.7) 9.21(5.95–14.36) <0.001 Possible mode of transmission Mother to child 17(2.9) 98(8.1) 1(Reference) Sex with partner 75(12.8) 209(17.1) 2.07(1.13–3.94) 0.011 Injecting drug use 255(43.6) 490(40.1) 3.00(1.78–5.27) <0.001 Commercial sex 237(40.5) 425(34.8) 3.21(1.90–5.65) <0.001 Age group (in years) Less than 11 21(3.6) 106(8.7) 1(Reference) 11–20 6(1.1) 37(3.1) 0.82(0.28–2.13) 0.72 21–30 227(38.8) 584(47.8) 1.96(1.21–3.28) 0.006 31–40 256(43.8) 413(33.8) 3.13(1.93–5.22) <0.001 More than 40 75(12.8) 82(6.7) 4.62(2.64–8.20) <0.001 Median CD4 T-cell count 107(52, 171) 347(232– - <0.001 (Q1, Q3) 487) Median CD8 T-cell count 710 (443, 1124(778– - <0.001 (Q1, Q3) 1085) 1526) CD4 T-cell <200 484(82.7) 211(17.3) 51.5(34.20–79.24) <0.001 200–350 74(12.7) 405(33.1) 4.10(2.61–6.57) <0.001 >350 27(4.6) 606(49.6) 1(Reference) CD4/CD8 T-cell count ratio 407 860 <0.25 301(73.9) 290(33.7) 10.7(6.36–18.84) <0.001 0.25–0.5 90(22.1) 405(47.1) 2.29(1.33–4.13) 0.002 >0.5 16(3.9) 165(19.2) 1(Reference)

* CI, confidence interval,;Q1=25th quartile, Q3=75th quartile.

WHO South-East Asia Journal of Public Health 2012;1(4):404-411 407 Factors associated with high prevalence of pulmonary Bishnu R Tiwari et al. tuberculosis in HIV-infected people in Kathmandu

All the factors significantly associated with is very high compared with similar studies PTB in univariate analysis were included in done in Cambodia and the United Republic of the multiple logistic regression model. The Tanzania,12–14 although the Cambodian study backward stepwise logistic regression revealed involved active case-finding of TB in a home- that only the CD4 T-cell counts of < 200/µL care programme of HIV-infected people. It and 200–350/µL, diarrhoea, parasitic infection should be noted that in our study, cases of PTB and sex with partner as probable mode of included newly diagnosed as well as existing transmission were independently associated. TB cases that had not commenced HAART. Participants with PTB were about 12 times This may partly explain the higher prevalence more likely to have a CD4 T-cell count of observed in our study. Also, since our study < 200/µL and two-and-a-half times more likely participants were being assessed for HAART to have a CD4 T-cell count of 200–350/µL eligibility on the basis of CD4 T-cell count, it compared with participants with a CD4 count is clear that at least one third of HIV patients of > 350/µL. Similarly, participants with had already developed PTB before the start of TB were more than twice as likely to have HAART. Such a high TB burden adds complexity diarrhoea and parasitic infection/infestation. to HIV treatment as well as to treatment of However, participants with probable mode TB itself.15,16 In addition, the finding has of transmission by “sex with partner” were implications for the possible impact of HIV on significantly less likely to have active TB a TB epidemic:since patients were diagnosed infection. The corresponding adjusted OR from by positive sputum smear microscopy, it is multivariate analysis is shown in Table 2. probable that they are highly infectious and a powerful source of transmission to others Discussion in their close environment, particularly their families.17 Although it has been shown that This is the largest study in the country to patients with CD4 T-cell counts of < 200/µL date involving participants from different risk may be less infectious, representing a lower groups to estimate the prevalence of PTB risk of transmission,11HIV-infected TB patients and associated factors and co-morbidities can be highly infectious in case of multi-drug in HIV-infected people before the start of resistant tuberculosis.18 HAART. Our study revealed that about one third of the participants presenting at NPHL This study showed that the HIV-infected had a pulmonary tuberculosis infection, which people attending for a CD4 T-cell count already suggests an alarming PTB epidemic among had a very poor immune status. The median HIV-infected people in Nepal. Our finding CD4 count was 257/µL and about 86% of all

Table 2: Factors associated with tuberculosis infection (multivariate analysis)

Variable Adjusted odds ratio P-value (95% CI) CD4 T-cell count <200/µL 11.69(6.23–21.94) <0.001 CD4 T-cell count 200–350/µL 2.52(1.30–4.89) <0.001 Diarrhoea 2.77(1.78–4.31) <0.001 Parasitic infection 3.34(2.02–5.50) <0.001 “Sex with partner” as possible mode of transmission 0.44(0.20–0.93) 0.03

408 WHO South-East Asia Journal of Public Health 2012;1(4):404-411 Bishnu R Tiwari et al. Factors associated with high prevalence of pulmonary tuberculosis in HIV-infected people in Kathmandu

participants had a CD4/CD8 ratio of ≤ 0.5. The prevalence of malnutrition in HIV/TB co- fact that the median time since the diagnosis infected patients.13 of HIV was less than six months indicates that When ‘sex with partner’ was a possible the participants were diagnosed in late stages mode of transmission, the odds of TB of the disease, which is not uncommon in status were almost halved. This is plausible resource-limited settings.19 because, compared with participants whose The lower CD4 count was strongly associated potential mode of transmission was injection with the higher odds of PTB. This finding drug use or commercial sex, those with ‘sex suggests the extent of missed opportunity of with partner’ as mode of transmission were preventing TB, probably due to a delayed start married females (87%), probably housewives, of HAART; earlier initiation, at a CD4 count who may be considered less vulnerable to TB. of >350/µL, may have a significant effect on However, supplementary data are lacking to reducing the incidence of TB. Lower CD4 count support this proposition. is a well-established risk factor for development One limitation of the study was that the of active TB among HIV-infected people.20–22 PTB status in most of the participants was The guideline of the World Health Organization ascertained on the basis of sputum smear to initiate HAART at CD4 T-cell count ≤ 350/µL microscopy. Microscopy is a less sensitive should therefore be followed strictly to prevent method than culture when the bacterial load TB or other opportunistic infections.23 in the sputum is lower.26 The prevalence of PTB The study showed diarrhoea and intestinal may therefore be marginally underestimated. parasitic infections to be important co- In addition, we do not have data on the status morbidities among participants with pulmonary of anti-tuberculosis treatment or its duration. tuberculosis. Participants with TB had more In the scenario of treatment of TB for more than twice the odds of having diarrhoea after than 3–4 weeks, the patient may convert to adjustment for other possible confounders. It smear negative, thus leading to a further is not entirely clear how diarrhoea is related underestimation of true prevalence. Since the to a TB outcome; however a randomized study was of cross-sectional nature, we have controlled trial from Zambia has demonstrated no information on the temporal relationship that preventive TB treatment can also prevent between diarrhoea and parasitic infection with diarrhoea among HIV-infected patients.24 It active PTB; these variables should therefore is very important to investigate further the be considered as co-morbidities rather than relationship between diarrhoea and active TB any indication of risk of developing TB. Data in HIV-infected patients. Similarly, participants on possible modes of transmission are based with TB had more than three times higher on patient information. However, one of the odds of having an intestinal parasitic infection. strengths of our study is that it is based on Opportunistic intestinal parasitic infections are real world observed settings. common in late stage HIV disease, and have 25 been described elsewhere. Given the high Conclusions proportion of participants in our study with diarrhoea and/or intestinal parasitic infection, We conclude that the prevalence of pulmonary it is likely that the prevalence of under-nutrition tuberculosis is alarmingly high in our study may also be very high. Although we did not population. The majority of participants had test for under-nutrition, a study from the a very poor immune status as assessed on United Republic of Tanzania has shown 70% the basis of the CD4/CD8 ratio before the

WHO South-East Asia Journal of Public Health 2012;1(4):404-411 409 Factors associated with high prevalence of pulmonary Bishnu R Tiwari et al. tuberculosis in HIV-infected people in Kathmandu

initiation of HAART. The odds of PTBs were 7. Munseri PJ, Bakari M, Pallangyo K, Sandstrom E. higher in participants with lower CD4 T-cell Tuberculosis in HIV voluntary counselling and testing counts, diarrhoea, and intestinal parasitic centres in Dar es Salaam, Tanzania. Scand J Infect Dis. 2010 Oct; 42(10): 767-74. infection. These findings may be useful for formulating better strategies for HIV and TB 8. Gao L, Zhou F, Li X, Jin Q. HIV/TB co-infection in mainland China: a meta-analysis. PLoS One. 2010; treatment programmes in the country. For 5(5): e10736. example, early diagnosis of HIV infection; motivation of patients to visit CD4 T-cell 9. Iliyasu Z, Babashani M. Prevalence and predictors of tuberculosis coinfection among HIV-seropositive count facilities as early as possible; initiation patients attending the Aminu Kano Teaching Hospital, of HAART on higher threshold values of CD4 northern Nigeria. J Epidemiol. 2009; 19(2): 81-7. T-cell counts; and treatment of intestinal 10. Cheesbrough M. District laboratory practice in tropical parasitic infection as an integral part of HIV countries, Part 1. 2nd edn. Cambridge: Cambridge programmes, may prove beneficial for better University Press, 2000. treatment outcomes. 11. Kenyon TA, Creek T, Laserson K, Makhoa M, Chimidza N, Mwasekaga M, et al. Risk factors for transmission Acknowledgements of mycobacterium tuberculosis from HIV-infected tuberculosis patients, Botswana. Int J Tuberc Lung We are thankful to all the volunteers who Dis. 2002 Oct; 6(10): 843-50.

participated in the study, and the laboratory 12. Mtei L, Matee M, Herfort O, Bakari M, Horsburgh CR, staff of the National Public Health Laboratory, Waddell R, et al. High rates of clinical and subclinical Kathmandu, Nepal. tuberculosis among HIV-infected ambulatory subjects in Tanzania. Clin Infect Dis. 2005 May 15; 40(10): 1500-7. References 13. Ngowi BJ, Mfinanga SG, Bruun JN, Morkve O. Pulmonary tuberculosis among people living with 1. Blaser MJ, Cohn DL. Opportunistic infections in HIV/AIDS attending care and treatment in rural patients with AIDS: clues to the epidemiology of AIDS northern Tanzania. Bmc Public Health. 2008 Sep and the relative virulence of pathogens. Rev Infect 30; 8: 341. Dis. 1986 Jan-Feb; 8(1): 21-30. 14. Kimerling ME, Schuchter J, Chanthol E, Kunthy T, 2. World Health Organization. Tuberculosis. Fact sheet Stuer F, Glaziou P, et al. Prevalence of pulmonary No. 104, 2012. Geneva: WHO, 2012. tuberculosis among HIV-infected persons in a home 3. Nepal, Ministry of Health and Population. Annual care program in Phnom Penh, Cambodia. Int J Tuberc report of tuberculosis control programme, 2009/2010. Lung Dis. 2002 Nov; 6(11): 988-94. Kathmandu: DoHS, National Tuberculosis Centre, 15. Kwara A, Flanigan TP, Carter EJ. Highly active 2010. antiretroviral therapy (HAART) in adults with 4. Ghimire P, Dhungana JR, Bam DS, Rijal B. Tuberculosis tuberculosis: current status. Int J Tuberc Lung Dis. and HIV co-infection status in United Mission Hospital, 2005 Mar; 9(3): 248-57. Tansen-Western Nepal. Journal of Tuberculosis, Lung 16. Nahid P, Gonzalez LC, Rudoy I, de Jong BC, Unger A, Diseases and HIV/AIDS. 2004; 1: 32-8. Kawamura LM, et al. Treatment outcomes of patients 5. Dhungana GP, Ghimire P, Sharma S, Rijal BP. with HIV and tuberculosis. Am J Respir Crit Care Med. Tuberculosis co-infection in HIV infected persons 2007 Jun 1; 175(11): 1199-206. of Kathmandu. Nepal Med Coll J. 2008 Jun; 10(2): 17. Singh M, Mynak ML, Kumar L, Mathew JL, Jindal 96-9. SK. Prevalence and risk factors for transmission of 6. Bhattacharya MK, Naik TN, Ghosh M, Jana S, Dutta infection among children in household contact with P. Pulmonary tuberculosis among HIV seropositives adults having pulmonary tuberculosis. Arch Dis Child. attending a counseling center in Kolkata. Indian J 2005 Jun; 90(6): 624-8. Public Health. 2011 Oct-Dec; 55(4): 329-31.

410 WHO South-East Asia Journal of Public Health 2012;1(4):404-411 Bishnu R Tiwari et al. Factors associated with high prevalence of pulmonary tuberculosis in HIV-infected people in Kathmandu

18. Escombe AR, Moore DA, Gilman RH, Pan W, Navincopa 22. Taha M, Deribew A, Tessema F, Assegid S, Duchateau M, Ticona E, et al. The infectiousness of tuberculosis L, Colebunders R. Risk factors of active tuberculosis patients coinfected with HIV. PLoS Med. 2008 Sep in people living with HIV/AIDS in Southwest Ethiopia: 30; 5(9): e188. a case control study. Ethiop J Health Sci. 2011 Jul; 21(2): 131-9. 19. Thanawuth N, Chongsuvivatwong V. Late HIV diagnosis and delay in CD4 count measurement 23. World Health Organization. Antiretroviral therapy among HIV-infected patients in Southern Thailand. for HIV infection in adults and adolescents: AIDS Care. 2008 Jan; 20(1): 43-50. recommendations for a public health approach: 2010 revision. Geneva: WHO, 2010. 20. Antonucci G, Girardi E, Raviglione MC, Ippolito G. Risk factors for tuberculosis in HIV-infected persons. 24. Mwinga A, Hosp M, Zulu I, Farthing MJ, Mulambo A prospective cohort study. The Gruppo Italiano di S, Kelly P. Tuberculosis preventative treatment also Studio Tubercolosis AIDS (GISTA). JAMA. 1995 Jul prevented diarrhoea in HIV-infected patients in 12; 274(2): 143-8. Zambia. AIDS. 2002 Mar 29;16(5): 806-8.

21. Silveira JM, Sassi RA, de Oliveira Netto IC, Hetzel 25. Gomez Morales MA, Atzori C, Ludovisi A, Rossi JL. Prevalence of and factors related to tuberculosis P, Scaglia M, Pozio E. Opportunistic and non- in seropositive human immunodeficiency virus opportunistic parasites in HIV-positive and negative patients at a reference center for treatment of human patients with diarrhoea in Tanzania. Trop Med immunodeficiency virus in the southern region of the Parasitol. 1995 Jun; 46(2):109-14. state of Rio Grande do Sul, Brazil. J Bras Pneumol. 26. Steingart KR, Ng V, Henry M, Hopewell PC, Ramsay 2006 Jan-Feb; 32(1): 48-55. A, Cunningham J, et al. Sputum processing methods to improve the sensitivity of smear microscopy for tuberculosis: a systematic review. Lancet Infect Dis. 2006 Oct; 6(10): 664-74.

WHO South-East Asia Journal of Public Health 2012;1(4):404-411 411 Original research

Compliance of off-premise alcohol retailers with the minimum purchase age law Areekul Puangsuwana, Kannapon Phakdeesettakuna, Thaksaphon Thamarangsia, Surasak Chaiyasonga,b

Background: In Thailand, the 2008 Alcoholic Beverages Control Act set the minimum purchase age (MPA) at 20 years old in order to limit new drinkers as part of the overall alcohol control effort. This study aims to assess the compliance of off-premise alcohol retailers with MPA restrictions and to identify factors affecting sales to adolescents.

Methods: A decoy protocol was used to quantify compliance of 417 alcohol retailers from three categories, namely grocers, modern minimarts and department stores. Multi-stage sampling was applied to obtain the samples in four provinces: Bangkok, Nakorn Sawan, Songkhla and Surin. Each alcohol retailer was visited twice by 17–19 year-old male and female adolescents who tried to buy alcohol. Information collected from focus groups and in-depth interviews with vendors and management officers were analysed for the qualitative methodology.

Results: Of all 834 buying attempts undertaken by the underage adolescent, 98.7% were successful in buying alcohol. Only 0.9% were asked for age and 0.1% were requested to show an ID card. Age and ID verifications were statistically significant to buying success as well as province, while number of vendors, gender and age of vendors and buyers, type of outlet, law cautions and advertisement signs in the outlet demonstrated no significant association.

Conclusions: The results showed that vendors fail to comply with the law despite the fact that they know the law. Enforcement needs to be strengthened to effectively limit new drinkers.

Key words: sting or decoy operation, alcohol purchase survey, law compliance, minimum purchase age (MPA).

Introduction Alcohol consumption has become more that Thailand has approximately 260 000 common and accepted in Thailand. The most new drinkers each year.1 Among the 12.8% worrisome trend is the rise in the number of adolescents aged 15–19 years who have of new drinkers. A study in 2008 revealed become regular drinkers, 21.6% are binge

aCenter for Alcohol Studies, International Health Policy Program, Ministry of Public Health, Nonthaburi, Thailand. bFaculty of Pharmacy, Mahasarakam University, Maha Sarakham, Thailand. Correspondence to Ms Areekul Puangsuwan, Center for Alcohol Studies, International Health Policy Program, Ministry of Public Health, Soi Satharanasook 6, Tiwanon Road, Nonthaburi 11000, Thailand, (Tel: +66 (0) 590 2316; Fax: +66 (0) 590 2380; e-mail: [email protected])

412 WHO South-East Asia Journal of Public Health 2012;1(4):412-422 Areekul Puangsuwan et al. Compliance of off-premise alcohol retailers with minimum purchase age law

drinkers and 40.4% are intoxicated youth To represent the population of 579 413 drinkers.2 These figures have drawn the outlet licensees in Thailand in 2009,14 a attention of concerned parties as drinking sample size of 385 was calculated based on among adolescents leads to many related the assumption that 50% of retailers do sell problems other than health, such as drug use, alcohol to adolescents (95% confidence and premature sexual intercourse, low educational 5% precision levels). The sample size was performance, crime and suicide.1 then increased by approximately 10% for any missing or incomplete cases, resulting in The Alcoholic Beverages Control Act, in an adjusted sample size of 428. Multi-stage effect since 2008, attempts to deter youths sampling was used to obtain the samples in from drinking by restricting the minimum four provinces. In the first-stage sampling, purchase age (MPA) to 20 years.3 The law four provinces were purposively selected with does not impose penalties on purchasers but different geographical locations and density of instead on vendors. In fact, the responsible alcohol licensees per capita: Nakorn Sawan government authorities have been trying to (northern) and Surin (north-east), with a high promote the law and its enforcement, and density of licensees per capita; and Bangkok civil society has increasingly launched special (central) and Songkhla (south), with a low to campaigns to emphasize its enforcement middle density. The samples were distributed in parallel to their usual educational and proportionally according to the number of awareness-raising campaigns. However, in a licences registered in each province. In the number of recent surveys, vendors in many second stage of sampling, main streets or provinces reported that 17–31% of their business areas of these four provinces were customers in the three months prior to being purposively selected to increase the likelihood surveyed seemed to be underage1,4,5 and of achieving the specified number of samples more than 50% of the general public had and for feasibility of the survey. witnessed sales made to underage customers.6 This led to research questions about whether Following several surveys conducted vendors consciously sell alcoholic beverages to enumerate and develop a map of the to underage customers and whether there physical locations of the retailers, a total of are any factors that influence such illegal 417 samples were used in this study. These sales. Previous studies that evaluated law were broken down as follows: Bangkok, 163 enforcement in Thailand through self-report (39.2%); Nakorn Sawan, 63 (15%); Songkhla, methodologies4,5 are not able to answer these 74 (17.8%) and Surin, 117 (28%). Alcohol questions due to methodological limitations. retailers were classified into three categories, An unobtrusive test purchase methodology, namely grocers, modern mini-marts and which has been applied in studies in other department stores. Temporary or canvas countries,7-13 is an alternative research outlets, mobile sales units, restaurants and methodology available. This study used test pharmacies that sold alcohol were excluded. purchases to assess the compliance of off- Independent variables in this study premise alcohol vendors on MPA restrictions included province, gender of buyer and in naturalistic settings. vendor, age of buyer and vendor, number of vendors, age check, ID card check, type of Methods outlet, law caution and advertisement signs. This study applied mixed methods of both The dependent variable was the result of a quantitative and qualitative methodologies. purchase attempt.

WHO South-East Asia Journal of Public Health 2012;1(4):412-422 413 Compliance of off-premise alcohol retailers Areekul Puangsuwan et al. with minimum purchase age law

For the qualitative methodology, 29 key Each adolescent had a different code informants who were vendors or owners, and was paired with an adult volunteer or including management staff from all outlet assistant researcher. The adolescent volunteer categories, were selected purposively from all was assigned to buy a particular beer brand provinces for focus groups (n=26, 13 vendors from a particular number of alcohol outlets or owners and 13 management staff) and in- appearing on the map. The adult volunteer depth interviews (n=3, 2 vendors or owners was instructed to enter the outlet before and 1 management staff). the adolescent to help observe the protocol and necessary information required for the Purchase protocol data collection form, record the conversation between vendors and the adolescent and A purchase protocol was designed based on for security purposes. The adolescents were guidance from the Pacific Institute for Research trained to tell the truth when challenged 15 and Evaluation and a data collection form about their age, to carry their ID cards and was also developed. Both the protocol and to show them if requested. When the vendor the form were tested and adjusted prior to sold alcohol to the adolescent, even with the actual purchase survey. age or ID card verification, that purchase Recruitment of the adolescent volunteers, attempt was considered successful. If the purchase procedure, particular brand and type vendor did not sell alcohol, even if they had of alcoholic beverage, timing of purchase and asked for age or ID, that purchase attempt conversation dialogue were strategically and was considered unsuccessful. The adolescents purposely planned. A total of 40 adolescents – were instructed to leave without complaint. 20 male and 20 female – who met the selection Each adult/adolescent pair had to complete criteria were recruited. The adolescents were the data collection form immediately after 17–19 years old with an appearance that each purchase attempt, in a place far from the matched their age. Their racial and/or ethnic vendor’s sight and from that of the next outlet characteristics were also harmonious with on the list. Each outlet was visited twice by those of the local residents in the purchase different gender adolescent volunteers. survey sites. The adolescents received permission from their parents or guardians Ethical considerations and submitted written consent for participation The covert test purchase of alcohol and use of in the survey. The minimum number of underage adolescents raised some significant adolescents participating in the survey was ethical issues. The rights of the vendors were four males and four females and the total well respected although advance survey number varied according to the sample size notice was not provided. The results of the in each province. A local partner organization study would have been invalid or distorted facilitated the recruitment of the adolescents, without the use of underage adolescents in a except in Bangkok where the adolescents were natural setting. This principle is in line with the recruited by the researchers. The adolescents International Ethical Guidelines for Biomedical and adult volunteers or assistant researchers Research Involving Human Subjects.16 were trained to note the information required Prosecution was not brought against vendors for the data collection form and undertook who sold alcohol to underage customers. To role-play rehearsals to become acquainted allay the concerns of researchers that the with purchase procedures and dialogue before survey might encourage the test purchasing they undertook the fieldwork.

414 WHO South-East Asia Journal of Public Health 2012;1(4):412-422 Areekul Puangsuwan et al. Compliance of off-premise alcohol retailers with minimum purchase age law

adolescents to initiate future alcohol use, 0.1% (n=1) were asked to show an ID card. orientation and training were provided. In 27.1% of the alcohol retailers placed alcoholic addition to permission from their parents beverages in front of the outlet, 33.1% at the or guardians and their written consent, the side, 16.5% at the back and 23.3% in other adolescents could withdraw their participation areas. The majority of the alcohol outlets at any time. The names of all volunteers, did not display relevant law cautions and alcohol retailers, vendors, and key informants advertisement signs (Table 1). remain confidential. This study was approved by the Ethical Committee of the Institute Overall results of the purchase for the Development of Human Research survey Protections, Ministry of Public Health. Overall, 98.7% (n=808) of purchase attempts were successful. While three provinces had Data analysis a very small proportion of sales refusals, Categorical variables were descriptively 100% of purchase attempts in Songkhla analysed using simple frequencies and province were successful. The proportion of percentages for general information obtained successful attempts made in different outlet from the survey such as purchase success, age categories, outlets with a different number of and ID card verifications, category of alcohol vendors and outlets displaying law cautions outlet, gender, age and number of vendors. and advertisement signs presented similar To examine associations among categorical results (Figure 1). variables, chi-square and Fisher’s exact tests were performed. Statistical significance was Association between variables assessed at p<0.05. SPSS version 15 and and purchase success STATA version 10 for Windows were used The study results demonstrated that the for the analysis. Content analysis of the proportion of successful and unsuccessful qualitative findings of the focus groups and purchases in the four provinces were statistically in-depth interviews were conducted. different. When vendors checked the age of the adolescent, the proportion of successful Results purchases dropped significantly from 99.3% to 28.6%.The purchase attempt was unsuccessful Characteristics of alcohol (0%) when the volunteer’s ID card was verified. retailers As a consequence, the verification of age and ID card had a statistically strong association When classified into the three different with purchase success (Table 2). outlet categories, there were 253 (60.7%) conventional grocers, 17 (4.0%) department Factors that had no significant association stores and 147 (35.3%) mini-marts. At the with purchase success were the gender and time of purchase, 68% of the alcohol outlets age of the adolescent and the gender, age had only one vendor, 32% had two or more and number of vendors. The proportion of vendors, and 67% of all vendors were female. sales to adolescent girls and boys was almost The estimated age of a third of vendors (35%) equal. A similar result was found across the from observation was under 30 years old. In different ages of test-purchasing adolescents. 0.8% (n=7) of the purchase attempts the When considering the characteristics of volunteers were asked for their age and only the vendors, the proportion of successful

WHO South-East Asia Journal of Public Health 2012;1(4):412-422 415 Compliance of off-premise alcohol retailers Areekul Puangsuwan et al. with minimum purchase age law

Table 1: Characteristics of alcohol retailers Characteristics No. % 1. Alcohol retailers by province Bangkok 163 39.2 Nakorn Sawan 63 15.0 Songkhla 74 17.8 Surin 117 28.0 Total 417 100.0 2. Alcohol retailers by category Department store 17 4.0 Grocer 253 60.7 Mini-mart 147 35.3 Total 417 100.0 3. Number of vendors in the outlet 1 567 68.0 2 203 24.3 3 64 7.7 Total 834 100.0 4. Gender of vendor Female 561 67.3 Male 273 32.7 Total 834 100.0 5. Estimated age of vendor < 30 years old 292 35.0 30–45 years old 272 32.6 46–60 years old 189 22.7 > 60 years old 81 9.7 Total 834 100.0 6. Placement of alcohol in the outlet Front 226 27.1 Side (left or right) 276 33.1 Back 138 16.5 Other 194 23.3 Total 834 100.0 7. Law caution in the outlet Yes 277 33.2 No 557 66.8 Total 834 100.0 8. Content of the law caution Restriction on time of sales 106 38.3 Restriction on minimum purchase age 170 61.3 Other 1 0.4 Total 277 100.0 9. Alcohol advertisements in the outlet Yes 102 12.2 No 732 87.8 Total 834 100.0

416 WHO South-East Asia Journal of Public Health 2012;1(4):412-422 Areekul Puangsuwan et al. Compliance of off-premise alcohol retailers with minimum purchase age law

Figure 1: Proportion of purchase successes by variable

100.0 99.6 90.0 100.0 98.9 96.6 100.0 98.8 80.0 98.7 99.5 100.0 98.1 70.0 98.1 98.4 99.0 98.2 60.0

50.0 99.3

Percentage 40.0

30.0

20.0 28.6

10.0

0.0

n

1

7

811

Total

Surin

endor

Sawa

Grocer

endors

endors

Bangkok

1v

Songkhla

Mini-mart

2v

3v

Law caution

check (N= )

Nakorn

No law caution

ID

Age check (N=7)

Department store

ID

ge check (N= ) Variable

No check (N=81 )

No a Unsuccessful Successful

purchases from male and female vendors was penalties. On the other hand, management virtually the same (98.9% vs 98.5%) as was staff of department stores and modern mini- the percentage of sales made by vendors who mart chains had a more precise knowledge of appeared below 60 years old. Sales occurred in the law and reported paying serious attention every instance when the underage adolescents to compliance with the law pertaining bought an alcoholic beverage from vendors who to the sale of alcoholic beverages. Their were older than 60. The number of vendors vendor employees were strictly instructed in an outlet was not associated with purchase not to sell alcoholic beverages to underage outcome; neither were the presence of law customers. cautions or advertisement signs at the outlet. Reasons for illegal sales and sale Highlighted findings from the refusals qualitative method Most vendors reported that they sold alcoholic beverages to underage customers because Knowledge of the law there was no legal comeback for doing so. The information collected from focus groups One vendor said that he might as well sell and in-depth interviews with vendors or it to adolescents because if he didn’t, they owners, including management staff of alcohol could easily buy it from another outlet with outlets, revealed that most vendors knew of total impunity. Sometimes, vendors asked restrictions on the sale of alcohol, but not of the for age if they suspected that the customer latest age restriction. Most vendors thought the was underage, but still sold them alcoholic MPA was still 18 years and were uncertain about beverages.

WHO South-East Asia Journal of Public Health 2012;1(4):412-422 417 Compliance of off-premise alcohol retailers Areekul Puangsuwan et al. with minimum purchase age law

Table 2: Association between variables and purchase success

Result Variable Total P Successful Unsuccessful Purchase attempt (%) 819 808 (98.7) 11 (1.3) Province 0.042* Bangkok 320 314 (98.1) 6 (1.9) Nakorn Sawan 118 114 (96.6) 4 (3.4) Surin 233 232 (99.6) 1 (0.4) Songkhla 148 148 (100.0) 0 (0.0) Gender of buyer 0.752 Female 408 402 (98.5) 6 (1.5) Male 411 406 (98.8) 5 (1.2) Age of buyer 0.544 17 years old 83 81 (97.6) 2 (2.4) 18 years old 282 279 (98.9) 3 (1.1) 19 years old 454 448 (98.7) 6 (1.3) Gender of vendor 1.000 Female 554 546 (98.6) 8 (1.4) Male 265 262 (98.9) 3 (1.1) Estimated age of vendor 0.768 <30 years old 280 277 (98.9) 3 (1.1) 30–45 years old 271 266 (98.2) 5 (1.8) 46–60 years old 188 185 (98.4) 3 (1.6) >60 years old 80 80 (100.0) 0 (0.0) Number of vendors present 0.386 1 561 551 (98.2) 10 (1.8) 2 194 193 (99.5) 1 (0.5) 3 64 64 (100.0) 0 (0.0) Age check <0.001* Not asked 811 805 (99.3) 6 (0.7) Asked 7 2 (28.6) 5 (71.4) ID card verification 0.013* Not asked 817 807 (98.8) 10 (1.2) Asked 1 0 (0.00) 1 (100.0) Type of outlet 0.084 Department store 32 32 (100.0) 0 (0.0) Grocer 501 493 (98.4) 8 (1.6) Mini-mart 286 283 (99.0) 3 (1.0) Law caution 0.354 None 550 544 (98.9) 6 (1.1) Yes 269 264 (98.1) 5 (1.9) Advertisement sign 0.633 None 719 710 (98.7) 9 (1.3) Yes 100 98 (98.0) 2 (2.0)

*Statistically significantP <0.05 at 95% confidence level.

418 WHO South-East Asia Journal of Public Health 2012;1(4):412-422 Areekul Puangsuwan et al. Compliance of off-premise alcohol retailers with minimum purchase age law

It was also found that illegal sales were particularly those in remote areas – are partially related to culture. Parents or senior aware of this law. This would also facilitate staff often instruct their children or subordinates the imposition of penalties. to perform tasks for them, including buying alcoholic beverages. Vendors of local grocers Discussion reported that they sold alcoholic beverages to children because they knew the adult drinker, The primary aim of this study was to assess were familiar with their family members, and the compliance of alcohol retailers with the believed that the alcohol was not intended for law in different outlet categories and in the child or adolescent in question. naturalistic situations. As a consequence, the methodology and findings of the study are The main reason cited for refusing a sale unique among the limited pool of research into was apprehension regarding the law. However, law enforcement and illegal sales to underage inconsistencies in practice were found for customers in Thailand. Our study also provides vendors who asked for age and ID. During the results of test purchases in cases when law peak time, age and ID verification was not cautions and advertisements were displayed in performed. Almost all vendors believed that the outlet, whereas other studies, both locally the decision to sell was primarily based on the and internationally, have not paid attention morals of the vendor. to these two variables. Thus, there were no comparable data. Perspectives on enforcement of the law and suggested measures to prevent The findings of this study have similarities illegal sales and disparities with other international studies. Generally, the proportion of purchase Most vendors had a relatively positive attitude success is consistent with the results of toward MPA. They strongly supported law studies in other countries10,17,18 although the enforcement but felt that awareness of figures in Thailand are much higher. The its importance, equity, wider coverage of underage adolescents had very little difficulty enforcement and more stringent penalties in obtaining alcoholic beverages, with a very were necessary. Random checks by authorized low risk of being challenged for proof of age. entities would encourage compliance. This is similar to the results of studies in São Advertisements and public campaigns run on Paulo, Brazil, and Washington, DC, United a regular basis could help stimulate vendors to States of America where the proportions be cautious when selling alcoholic beverages of purchase success were as high as 80% to young customers. and 97% respectively.8,10 The findings of Vendors did not consider themselves a the study in Brazil were also consistent with cause, but rather a consequence of youth our study in that the age and gender of the drinking. Therefore, they suggested a law adolescent and the type of alcohol outlet did against alcohol purchase and possession by not influence purchase success, while a study adolescents, and reflected that the law should in the Netherlands found that underage female impose penalties on the adolescents, as well adolescents were more successful with alcohol as their parents, to prevent illegal purchase. purchases than males.9 It would be interesting It was also noted that while basic knowledge to study further the effectiveness of MPA of the relevant law should be provided, the in different countries and characteristics of authorities concerned should work more interventions that could promote sustainable proactively to ensure that all vendors – enforcement of this particular restriction.

WHO South-East Asia Journal of Public Health 2012;1(4):412-422 419 Compliance of off-premise alcohol retailers Areekul Puangsuwan et al. with minimum purchase age law

With regards to the age of the adolescent locations of alcohol outlets was created, based and the vendors, this study did not show any on an enumeration survey undertaken by significant relationship of these two variables the researchers. Lastly, it is quite common with purchase success, while the study in for grocers to provide a few tables for Brazil found that vendors older than 30 years their customers, but this is for occasional were seven times more likely to refuse sales to use and not mainly for on-premise alcohol underage customers. Concerning the number consumption. These grocers were therefore of vendors in the outlet, one study observed included in the survey, but were very small that sales to underage customers were more in number. Thus, the study limitations were likely when the vendor was alone.7 In contrast, dealt with effectively to reduce any possible this study found that sales were less likely influence on the results. when there was only one vendor in the outlet, although this relationship is not statistically Conclusions and significant. recommendations We also reflected on the unexpected result The results of our study reflect that law in Surin. Despite major campaigns and strong enforcement, in respect of MPA restrictions, networks of civil society organizations working still needs to be strengthened despite the fact actively for alcohol control at the provincial that the Alcoholic Beverages Control Act has level, the purchase success in Surin was very been in effect since 2008. Age and ID card high at 99.6%. A preliminary recommendation verifications are effective interventions to limit would be for such campaigns and networks to access to alcoholic beverages of underage try to focus on promoting law enforcement, adolescents. When the vendors asked for rather than raise public awareness of the age, sales of alcohol to the adolescents harms of alcohol and to change people’s dropped to approximately one third, whereas attitudes towards drinking alcohol in general. ID card verification led to a complete sales This proposal is in line with a report from refusal. Alcohol retailers should therefore an evaluation of the status and direction of be encouraged to perform age and ID card alcohol control campaigns in Bangkok during verifications when faced with young customers 2010–2011.19 to help prevent youth drinking.20

Limitations of the study This study also suggests that many parties concerned with law enforcement should A major difficulty was being able to identify cooperate to reduce sales to adolescents. whether or not the samples (alcohol outlets) The potential benefits of enforcement of MPA studied actually held a sales licence. Inclusion restrictions are substantial and the costs are and exclusion criteria were clearly established low.21 The Excise Department could contribute to increase the chance of having a proper by emphasizing the restriction on sales to sample. Moreover, it was considered that underage persons when a licence is granted. alcohol outlets in certain physical locations A range of practical interventions ranging would normally hold sales licences. In addition, from simple to complex should be considered, we could not use the physical addresses of the such as enclosure of law caution stickers alcohol outlets officially registered because it and leaflets with the licence, brief verbal could not be proved that they were actually introductions to the law, guidance on what to operating at the registered address and in do when encountering young customers and what category. Thus, a map of the physical

420 WHO South-East Asia Journal of Public Health 2012;1(4):412-422 Areekul Puangsuwan et al. Compliance of off-premise alcohol retailers with minimum purchase age law

licensee meetings. A short training course advertisemnet and promotion of alcohol after the as a prerequisite for alcohol sales licensees 1st anniversary of the enactment of the Alcohol would also help promote and raise awareness Beverages Control Act 2008 Bangkok: Assumption University, 2010. of the law. In addition to the existing penalties under the Alcoholic Beverages Control Act, 7. Willner P, Hart K, Binmore J, Cavendish M, Dunphy E. Alcohol sales to underage adolescents: an unobtrusive the licence should be terminated when observational field study and evaluation of a police alcohol retailers infringe the law. Increased intervention. Addiction. 2000; 95(9): 1373-88. monitoring and enforcement activities, such 8. Romano M, Duailibi S, Pinsky I, Laranjeira R. Alcohol as random visits and warning letters, by purchase survey by adolescent in two cities of state enforcement or regulatory-related authorities of Sao Paulo, Southeastern Brazil. Rev Saude Publica. could also lead to positive changes in age and 2007; 41(4): 1-6. 18 ID card verification practices. Campaigners, 9. Gosselt J, Hoof JJV, Jong MDT, Prinsen S. Mystery communities, mass media and civil society shopping and alcohol sales: Do supermarkets and could also contribute actively to reducing liquor stores sell alcohol to underage customers? illegal sales to minors. Campaigns specifically Adolescent Health. 2007; 41(2007): 302-08. addressing the MPA, community-based 10. Preusser DF, Williams AF, Weinstein HB. Policing interventions with multisectoral partnerships, underage alcohol sales. Journal of Safety Research. and media advocacy can all be part of the 1994; 25(3): 127-33. overall effort to prevent underage alcohol sales 11. Wolfson M, Toomey TL, Murray DM, Forter JL, Short and related problems.22 BJ, Wagenaar AC. Alcohol outlet policies and practices concerning sales to undreage people. Addiction. 1996; 91(4): 589-602.

12. Forster LJ, Murray MD, Wolfson M, Wagenaar AC. References Commercial availability of alcohol to young people: 1. Assanangkornchai S, Muekthong A, Indhanon T. A result of alcohol purchase attempts. Preventive surveillance of drinking behaviors and other health- medicine. 1995; 24: 342-47. risk behaviours among high school students in 13. Rehnman C, Larsson J, Adreasson S. The beer Thailand. Bangkok, 2008. campaign in Stockholm - attempting to restrict the 2. Sornphaisarn B, Kaewmongkul J, Imsarnphang S, availability of alcohol to young people. Alcohol. 2005; Wathanaporn K, Nasueb S, Unlamlert W. Thai Alcohol 37: 65-71. Situation. Bangkok: Center for Alcohol Studies, 14. Tubwong J, et. al. Comparison of licensing measure 2007. for controling physical availability of alcohol in 3. Rex BA. Alcohol Consumption Control Act B.E.2551 Thailand and other countries. Nonthaburi: Center for (2008), 2008 - http://www.thaiantialcohol.com/ Alcohol Studies, 2009. eng/images/law/alcohol_beverage_control_act.pdf 15. Grube JW, Stewart K. Guide to conducting alcohol - accessed 16 April 2013. purchase surveys. Cambridge: Office of Juvenile 4. Ratchadapanthikul C, Rungpatchim C. Evaluation of Justice and Delinquency Prevention, 1999. law enforcement of the Alcohol Beverages Control 16. The Council for International Organizations of Medical Act 2008 in Bangkok. Nothaburi: Center for Alcohol Sciences (CIOMS) in Collaboration with the World Studies, 2010. Health Organization. International for biomedical 5. Ratchadapanthikul C, Rungpatchim C. Evaluation of research involving human subjects. Nonthburi: Ethical law enforcement of the Alcohol Beverages Control Act Committee of the Institute for the Development of 2008 in Nonthaburi, Nakornprathom and Singburi. Human Research Protections, 2009. Nothaburi: Center for Alcohol Studies, 2010. 17. Preusser DF, Williams AF. Sales of alcohol to 6. Academic Network for Community Happiness underage purchasers in three New York Countries Observation and Research (ANCHOR). A study and Washington, DC Palgrave Macmillan. 1992; 13(3 on perception, law compliance, accessibility, Autumm): 306-17.

WHO South-East Asia Journal of Public Health 2012;1(4):412-422 421 Compliance of off-premise alcohol retailers Areekul Puangsuwan et al. with minimum purchase age law

18. Huckle T, Greenaway S, Broughton D, Conway K. Zealand to improve age checks for young people The use of an evidence-based community action pruchasing alcohol. Health Promotion International. intervention to improve age verification practices 2005; 2(2): 147-55. for alcohol purchase. Informa Healthcare. 2007; 21. Wagenaar AC, Wolfson M. Deterring sales and 42: 1899-914. provision of alcohol to minors: a study of enforcement 19. SthapitanondaP et al. Results of an evaluation in 295 countries in four stages. Public Health Report. on status and direction of campaigns for alcohol 1995; 110: 419-27. consumption control in 2010-2011 Bangkok: Center 22. Grube JW, Stewart K. Preventing sales of alcohol to for Alcohol Studies, 2009. minors: Results from a community trial. Addiction. 20. Huckle T, Conway K, Casswell S, Pledger M. Evaluation 1997; 92 (Supplement 2): S251-60. of a regional community action intervention in New

422 WHO South-East Asia Journal of Public Health 2012;1(4):412-422 Original research

External quality assessment in blood group serology in the World Health Organization South-East Asia Region

Patravee Soisangwana

Background: The quality of blood transfusion services (BTS) is essential for the treatment of patients who need blood or blood products. BTS involve several steps, including the acquisition of the donor’s blood, blood grouping, unexpected antibody screening, blood storage, transfusion, etc. There is a need to check the effectiveness of all elements in the BTS can be assessed and monitored by an external quality assessment.

Aim: To assess and evaluate the performance of ABO and Rh(D) blood grouping and unexpected antibody screening of the selected World Health Organization (WHO) South-East Asia Region Member country laboratories.

Methods: WHO Collaborating Centre on Strengthening Quality of Health Laboratories (Thailand) organized a regional external quality assessment scheme for blood group serology (REQAS-BGS) between 2002 and 2008 for laboratories in countries of the WHO South-East Asia Region. Test items for ABO and Rh(D) blood groupings and unexpected antibody screening and identification were distributed three cycles per year to BTS laboratories in Bangladesh, Bhutan, India, Indonesia, Maldives, Myanmar, Nepal, Sri Lanka and Thailand. By the end of the project, a total of 20 BTS laboratories had participated for differing lengths of time.

Results: It was found that 87.5%, 93.3%, 81.3%, 92.3%, 100% and 87.5% of laboratories returned the test results in 2002, 2003, 2004, 2006, 2007 and 2008, respectively. Laboratories with excellent quality or a trend of quality improvement for ABO and Rh(D) blood grouping, unexpected antibody screening and identification during the six years were 60% (12/20), 50% (10/20), 52.9% (9/17) and 81.8% (9/11), respectively. At the initiation of the scheme, most laboratories were using substandard methods for ABO and Rh blood groupings, i.e. performing only direct blood grouping alone but subsequently adopted the standard methods, i.e. performing both direct and reverse blood groupings.

Conclusion: REQAS-BGS in South-East Asia countries has been useful for assessing, monitoring and improving the quality of testing. Challenges such as high costs and regulatory requirements for international shipment of blood samples could be solved by amending the regulation(s) for shipment, or establishing a national EQAS.

Key words: Regional external quality assessment scheme (REQAS), blood transfusion services, laboratories, quality, ABO, Rh(D) blood grouping, unexpected antibody screening and identification.

a Bureau of Laboratory Quality Standards, Department of Medical Sciences, Ministry of Public Health, Thailand. Correspondence to Patravee Soisangwan (e-mail: [email protected])

WHO South-East Asia Journal of Public Health 2012;1(4):423-431 423 External quality assessment in blood group serology in the Patravee Soisangwan WHO South-East Asia Region

Introduction WHO EQA schemes5 have been established in the main areas of testing in transfusion The quality of laboratory investigation results laboratory practice by the United Kingdom is essential for the treatment of patients, National External Quality Assessment Service monitoring the progress of disease, and disease (UK-NEQAS). Because of the absence of prevention and surveillance. The quality of the national EQAS in SEAR countries, a regional results in the blood transfusion services (BTS) EQAS was initiated by WHO’s Regional office is critical since the human blood product is for South-East Asia to create awareness directly transfused to patients. Any error in the among the countries and to provide them with test result (false negative or false positive) can technical support to set up and improve the have serious consequences including death of quality of their performance. Accordingly, a the recipients. To achieve a continuous quality Regional External Quality Assessment Service system in BTS, the effectiveness of all critical in Blood Group Serology (REQAS-BGS) was associated elements should be assessed, organized in 2002 by the Bureau of Laboratory using both internal and external mechanisms. Quality Standards (BLQS), Department of Internal assessment can be done by the use of Medical Sciences, Thailand (WHO Collaborating batch controls, staff competency testing and Centre on Strengthening Quality of Health audits. But evaluation of the whole process Laboratories) to strengthen BTS in SEAR in a laboratory, requires an external quality countries. All countries participated, except assessment (EQA).1 EQA can be conducted by the Democratic People’s Republic of Korea delivering test items of known compositions and Timor-Leste since these two countries that closely simulate the clinical materials lacked BTS and experienced difficulties with to participating health laboratories and the shipment route of the test items. The aim subsequently comparing their results. EQA of the REQAS-BGS project was to evaluate the has several objectives: (i) assess laboratory performance of the laboratories for ABO and performance and the ability to determine the Rh blood grouping and unexpected antibody correct results; (ii) identify any errors in the screening and identification of the participating laboratory; (iii) stimulate laboratory personnel laboratories during 2002–2008. to improve their performance and the use of standard methods; and (iv) encourage the establishment and implementation of quality Materials and methods 2,3 systems with the relevant standards. In this cross-sectional, retrospective study in To minimize errors in the blood banking a 20 laboratories from nine SEAR countries procedures, the WHO’s Regional office for participated. Each laboratory was given a South-East Asia has been implementing confidential code number, which was used effective quality assurance for BTS in Member for performance evaluation reporting of countries since 1998.4 This external quality data results. Data on evaluation were made assessment scheme (EQAS) is recognized available only to the respective participating as an important tool for laboratory quality laboratory n and to the coordinator at the WHO improvement. EQAS data are also recognized South-East Asia Regional Office (SEARO). by accreditation bodies as important elements A breakdown of the laboratories participating for meeting quality standards for laboratory in the EQA scheme is shown in Table 1. quality as per ISO 15189 and ISO/IEC 17025.

424 WHO South-East Asia Journal of Public Health 2012;1(4):423-431 Patravee Soisangwan External quality assessment in blood group serology in the WHO South-East Asia Region

Table 1:Breakdown of reports of the total 20 participating laboratories per distribution cycle

Distribution Number of reporting laboratories cycle Bangladesh Bhutan India Indonesia Maldives Myanmar Nepal Sri Thailand (#Cycle/Year) Lanka 1 (1/2002) 2 1 3 1 2 2 2 1 1 2 (2/2002) 2 1 4* 1 2 2 2 1 1 3 (1/2003) 2 1 4 1 2 2 1 1 1 4 (2/2003) 2 1 4 1 2 2 1 1 1 5 (3/2003) 2 1 4 1 2 2 1 1 1 6 (1/2004) 2 1 4 1 2 2 1 1 1 7 (2/2004) 2 1 4 1 2 2 2 1 1 8 (3/2004) 0** 1 4 1 2 2 2 1 1 9 (1/2006) 0 1 4 2* 0** 2 2 1 1 10 (1/2007) 0 1 4 2 0 2 2 1 1 11 (2/2007) 2*** 1 4 2 1*** 2 2 1 1 12 (3/2007) 2 1 4 2 1 2 2 1 1 13 (1/2008) 2 1 4 2 1 2 2 1 1 14 (2/2008) 2 1 4 2 1 2 2 1 1 15 (3/2008) 2 1 4 2 1 2 2 1 1 Total 4 1 4 2 3 2 2 1 1

*One additional laboratory in India and Indonesia from distribution cycles 2 and 9, respectively. **Withdrawn from EQA participation from distribution cycles 8 and 9, respectively. ***New laboratories (two from Bangladesh and one from Maldives participated from distribution cycles 11 to 15. The total number of the participating laboratories was 20.

The packed red cells and converted sera for a single or mixed unexpected antibody) used in the preparation of the test items were for screening and identification were also sent provided by the National Blood Center, Thai in the same shipment. The test sets were Red Cross Society (the WHO Collaborating dispatched 15 times, three distribution cycles Centre for Training in Blood Transfusion per year during 2002–2008, to each laboratory Medicine) They were tested negative for HBV, by courier in non-cold package condition HCV and HIV infections and syphilis. The (Table 1). Participating laboratories in India test items were prepared for samples of ABO received the test items through SEARO during and Rh(D) blood groupings and unexpected distribution cycles 3–15. Details of the test antibody screening test (screening) and items distributed are shown in Table 2. identification. The participating laboratories were For each distribution cycle, three sets of requested to conduct only the tests currently the test items (red blood cell suspensions and available in their routine services. Laboratory corresponding ABO blood group sera) were performance evaluation was based on the distributed to the designated laboratories. status of report return within the closing date, Three serum samples (negative or positive completeness of the test data, the standard

WHO South-East Asia Journal of Public Health 2012;1(4):423-431 425 External quality assessment in blood group serology in the Patravee Soisangwan WHO South-East Asia Region

Table 2:Details of the test items distributed to participating laboratories

Blood grouping Distribution ABO1 ABO 2 ABO3 Rh 1 Rh2 Rh 3 UnAb 1 UnAb 2 UnAb 3 cycle 1 AB A O Pos Neg Pos Neg Anti-D Neg 2 O AB A3 Pos Neg Pos Anti-Mia Neg Neg 3 AB O A Pos Neg Weak-D Neg Anti-D Neg 4 Subgroup B AB O Pos Neg Pos Neg Neg Anti-E 5 O B A Neg Pos Pos Anti-D Neg Neg 6 B AB O Pos Neg Pos Neg Anti-D Neg 7 B AB O Pos Weak-D Pos Anti-Mia Anti-D Neg 8 A AO Pos Neg Neg Anti-E Neg Neg 9 A BO Pos Neg Weak-D Neg Neg Anti-E+ Anti-C 10 A B O Neg Pos Pos Anti-Mia Neg Neg 11 OOA Pos Neg Pos Anti-Mia Anti-D Anti-E 12 O B AB Weak-D Neg Pos Neg Neg Anti-D 13 A B O Neg Pos Neg Neg Anti-D Anti-Mia + anti- Leb 14 B O AB Neg Pos Pos Neg Anti-D Neg 15 A AO Pos Neg Weak-D Anti-D Anti-D Neg

Neg, negative; Pos, positive; UnAb, unexpected antibody screening test and identification.

method for the test, patterns and degrees of for each testing was 4.0. Quality ranking was cell agglutination, consistency of agglutination based on the full score, i.e. excellent (4.0), pattern, blood group interpretation and very good (3.5–3.99), good (3.0–3.49?), detection, and type of unexpected antibody.6 borderline (2.5–2.99), and unacceptable The results were compared with the consensus (< 2.49). modes derived from the Thai-NEQAS results. A performance evaluation report for Professional judgment by the Expert Committee each distribution cycle with suggestions and (comprising the experts from the Thai National comments for improvement was submitted to Red Cross Society, Deans of the Faculty of the individual participants. A summary report, Medical Technology and Faculty of Allied Health which included the data of all participants’ Sciences of the Thai universities, etc.) was performance and types of reagents, was also used for the final decision when consensus dispatched to all participants for review and mode could not be calculated. The REQAS test comparison of their performance with the items were the same sample sets organized others. To ensure confidentiality, participating for the Thai-NEQAS. The overall scores for laboratories received a summary of the performance evaluation of each testing were evaluation report without mentioning names calculated from the results of the three test of the laboratories.) items in each distribution cycle. A full score

426 WHO South-East Asia Journal of Public Health 2012;1(4):423-431 Patravee Soisangwan External quality assessment in blood group serology in the WHO South-East Asia Region

Results Performance evaluation was calculated based on the system used for the Thai national At the beginning of the project, 15 laboratories EQAS in blood group serology currently from nine countries participated. Two additional conducted by BLQS for individual testing. laboratories from Bangladesh and one from The evaluation method was developed by the Maldives joined the project to replace those Bureau of Laboaratory Standards accredited that withdrew at the 8th, 9th and 10th according to ISO/IEC 17043 by NATA, cycle. The total number of participating Australia. Participating laboratories followed laboratories in the REQAS-BGS project was 20 testing manual. For ABO blood grouping, three (Table 1). Because certain laboratories did not laboratories showed consistently excellent always receive the test items, mainly due to quality for all cycles while four had unacceptable customs clearance problems, the percentage performance in at least one distribution cycle of laboratories that returned reports was caused by either the use of a substandard calculated based on those that received the method (cell grouping only), incorrect ABO test items, not on the total number of delivery blood grouping, inconsistent results between cycles. It was found that 14 of 20 (70%) cell and serum groupings, or incomplete laboratories returned results for every set records of the results in the report form. of tests received. Two laboratories reported For Rh(D) blood grouping, two laboratories on two tests (ABO and Rh blood grouping), showed consistently excellent quality for 7 laboratories reported on three tests (ABO, all cycles, whereas 11 had unacceptable Rh blood grouping and unexpected antibody performance in at least one distribution cycle. screening) and 11 reported on all four tests The main causes of the errors were the use (ABO, Rh blood grouping, unexpected antibody of a substandard method to identify weak-D, screening and identification) (Table 3). incorrect Rh(D) blood group, or incomplete Test items showing mixed-field agglutina- records of the results (Table 4). tion, i.e. A3 and B subgroups, were sent to Performance evaluation for unexpected participating laboratories. Only 1 of 14 labora- antibody screening showed that 3, 6 and tories reported the correct A3 subgroup result, 11 laboratories reported on no, some and the remainder reporting A blood group without all distribution cycles, respectively. Of the mixed-field agglutination, and while all the 11 complete reports, the results of laboratories identified B blood group, only 2 3 laboratories were excellent and 8 were reported the correct subgroup.

Table 3: Details of tests reported by participating laboratories for the performance evaluation

No. of Type of test No. of countries tests 2 ABO and Rh(D) blood grouping 2 Maldives (1) Nepal (1) 3 ABO, Rh(D) blood grouping and 7 Bangladesh (3), Maldives (2), unexpected antigen screening Myanmar (1), Nepal (1)

ABO, Rh(D) blood grouping, 11 Bangladesh (1), Bhutan (1), India (4), 4 unexpected antigen screening and Indonesia (2), Myanmar (1), identification Sri Lanka (1), Thailand (1)

WHO South-East Asia Journal of Public Health 2012;1(4):423-431 427 External quality assessment in blood group serology in the Patravee Soisangwan WHO South-East Asia Region

of unacceptable quality. The major cause of of potentially infectious materials, and errors was false-negative results. The ratio to infrequent flights from Bangkok to the of negative to false positive in all errors was destination. In addition, multiple steps to 21:4. Only three laboratories successfully clear the test samples from customs, despite identified the unexpected antibody for all prior arrangement of the delivery process distribution cycles, 8 laboratories reported for in accordance with the International Air some distribution cycles and 9 did not report. Transport Association (IATA) regulations, Four of the 6 laboratories that only returned resulted in delayed deliveries, non-deliveries reports in certain distribution cycles had and deterioration of the test samples. These excellent quality for all returned results. The delays were reduced in later cycles, which unacceptable quality was due to false-positive, may have been due to better planning and false-negative, unidentified and reported mixed management of the delivery schedules and types of antibody reports. more awareness and experience of the individuals involved. Moreover, the delivery of The status of the quality system in each batch materials to a designated collaborator laboratory was determined by analysing the for forwarding to participating laboratories in quality ranking of each testing. The status the same country would significantly reduce was classified into three groups, i.e. excellent transportation costs as well as the delivery quality for all cycles; showed a trend of quality time. improvement; and showed no significant quality improvement. The number of laboratories Although the test items were shipped classified into each group is shown in Table 5. outside the cold chain, we carried out random checks on the homogeneity and stability of Discussion the aliquots stored or transported at room temperature in all 15 distribution cycles, to The REQAS-BGS project was first initiated in ensure that the samples were still acceptable SEAR in 2002 under the auspices of SEARO. if they reached their destination five days A major challenge encountered was the after dispatch. shipment of the test items to certain countries due to regulations relating to the importation

Table 4: Number of laboratories with unacceptable quality and the main causes of the errors

ABO Rh Unacceptable quality at least in 4 11 1 of 15 distribution cycles Cause of error • use of substandard method • laboratory did not perform (cell grouping only) indirect antiglobulin test to identify weak-D • reported incorrect ABO blood grouping • laboratory performed indirect antiglobulin test • inconsistent results between but reportedRh negative for cell and serum groupings weak-D or weak-D for Rh • incomplete record of results positive in report form • incomplete record of results in report form

428 WHO South-East Asia Journal of Public Health 2012;1(4):423-431 Patravee Soisangwan External quality assessment in blood group serology in the WHO South-East Asia Region

Table 5: Summary of the quality status of laboratories from the performance evaluation results (Answer: right)

ABO blood Rh blood Unexpected Unexpected grouping grouping antibody antibody screening identification Number of laboratories Excellent quality for 4 (20.0%) 2 (10.0%) 4 (23.5%) 5 (45.4%) all cycles Trend of quality 8 (40.0%) 8 (40%) 5 (29.4%) 4 (36.4%) improvement No significant 8 (40.0%) 10 (50%) 8 (47.1%) 2 (18.2%) trend of qualityimprovement Total number of 20 20 17 11 laboratories

All the 20 BTS laboratories nominated to Since the five laboratories that could not participate in the project were well-established correctly identify the weak-D blood group with responsibility for the supply of blood probably performed only the immediate spin products within their respective countries. method, it is recommended that the indirect However, they are not representative of all antiglobulin test be used for correct blood laboratories in the country in the majority of group identification. This test is needed for cases. Their performance evaluation in the cross-matching in all BTS laboratories. REQAS-BGS project reflects their current and With regard to the test items with mixed- potential competency in BTS, and should be field agglutination, initially only 1 reported used to promote a fully functioning quality the correct A3 and 2 laboratories reported the system for the acquisition and transfusion of correct B subgroups. However, after receiving the blood products needed by the patients. the evaluation results, most laboratories The REQAS-BGS project involved the reported the correct results in subsequent performance evaluation of four routine tests delivery cycles. This showed that the BTS in BTS, i.e. ABO and Rh(D) blood grouping, laboratories could improve their competency and unexpected antibody screening and for blood groups with irregular reactions. identification. Only 11 of 20 laboratories BTS laboratories must be able to screen from seven Member countries – Bangladesh, and identify unexpected antibodiesin patients Bhutan, India, Indonesia, Myanmar, Sri Lanka in order to select blood products without the and Thailand – provided results on all the corresponding antigen. This is essential to four tests. Furthermore, not all laboratories avoid any transfusion reaction or antibody performed the standard ABO blood grouping responses that may result in further difficulty procedures (both cell and serum grouping), in selecting blood products, especially for either because they lacked the standard cells patients with thalassemia.8,9 In this study, or because they were unable to differentiate unexpected antibody screening test could be the weak-D grouping. The latter is essential performed by 18 participating laboratories, to prevent the transfusion of weak-D blood to among which 11 could perform the antibody a Rh-negative patient which would result in identification. Nine laboratories prepared production of the corresponding antibodies.7

WHO South-East Asia Journal of Public Health 2012;1(4):423-431 429 External quality assessment in blood group serology in the Patravee Soisangwan WHO South-East Asia Region

their own screening cells but only 2 of them Since 2002, BTS staff from 9 of the 11 SEAR reported the correct results. False-negative Member countries have received WHO grants results, especially anti-Mia, were reported by for training on the establishment of EQAS in laboratories that did not have the screening BGS and laboratory quality systems. These cells for the antibody. Since anti-Mia is countries should thus have sufficient knowledge common in Asian populations, screening and experience to establish their own EQAS cells that can detect the antibody should be rather than relying on regional or international available in BTS laboratories in SEAR.10–12 The quality assessment networks. This would not REQAS-BGS project stimulated laboratory only provide all local laboratories with quality quality in the participating BTS since many of performance assessment and an overall BTS them have started to procure the appropriate quality management record at the national screening and panel cells for their routine level, but also future planning of quality use, and have improved the quality of their improvement by policy-makers. services by evaluating the root causes of Presently, Bhutan has begun to provide error.13 However, it is expected that the cost national EQAS in BGS. India has EQAS for would be significantly reduced if laboratories blood banking, although this service is available could prepare their own cells since commercial only for transfusion-transmitted infectious cells are expensive and normally short-lived. markers.15 Some Member countries, such as The preparation of standardized cells should India, Indonesia and Thailand, have already be obtainable from the selected local donor established strong laboratory quality systems panel and this would facilitate the use of as well as an accreditation body, which could the standard method of blood grouping and facilitate and promote the improvement of provide an easy and low-cost supply of the quality systems in sister countries. standard cells. WHO has been promoting quality assurance Even though some laboratories reported in countries of the Region since the 1980s. the correct BGS results, the number of test Several consultations and workshops have items was small compared with the number stimulated countries to practise quality of the samples routinely tested in which an assurance with emphasis on internal quality error could occur. Samples that give inclusive control and EQAS.16 results by cell grouping alone need to be retested with the serum grouping technique. Therefore, an appropriate quality system must Conclusions be established for most accurate and efficient The results of our project point to the importance laboratories. of quality systems and continuous quality Continuous training in both quality and assessment, both internal and external, in BTS academic aspects of BTS has been provided to laboratories. The limitations and obstacles in laboratories since the initiation of the quality organizing REQAS should encourage policy- management training project by SEARO in makers and responsible institutions to initiate 2001. However, a number of BTS still require EQAS at the national instead of the regional quality improvement. Indeed, the WHO global level. The retrospective data can also be used database on blood safety summary report 2011 to follow up improvements in the BTS quality indicates that 47% of donations are tested in systems to sustain the success of participating laboratories without quality assurance.14 laboratories in the REQAS-BSG project.

430 WHO South-East Asia Journal of Public Health 2012;1(4):423-431 Patravee Soisangwan External quality assessment in blood group serology in the WHO South-East Asia Region

References 8. Pahuja S, Pujani M, Gupta SK, Chandra J, Jain M. Alloimmunization and red cell autoimmunization 1. World Health Organization, Regional Office for South- in multitransfused thalassemics of Indian origin. East Asia. Quality assurance and accreditation: report Hematology. 2010 Jun; 15(3): 174-7. of intercountry consultation, Yangon, Myanmar, 16- 19 November 1999. New Delhi: WHO-SEARO, 2000. 9. Cheng CK, Lee CK, Lin CK. Clinically significant Document No. SEA-HLT-323 - http://203.90.70.117/ red blood cell antibodies in chronically transfused PDS_DOCS/B3393.pdf - accessed 17 April 2013. patients: a survey of Chinese thalassemia major patients and literature review. Transfusion. 2012 Feb 2. World Health Organization. External quality 17. doi: 10.1111/j.1537-2995.2012.03570.x. assessment of transfusion laboratory practice: guidelines on establishing an EQA scheme in blood 10. Prathiba R, Lopez CG, Usin FM. The prevalence of GP group serology. Geneva: WHO, 2004. Document No. Mur and anti-”Mia” in a tertiary hospital in Peninsula WHO/EHT/04.09 - http://www.who.int/bloodsafety/ Malaysia. The Malaysian Journal of Pathology. 2002 EQA_in_Blood_Group_Serology.pdf - accessed 17 Dec; 24(2): 95-8. April 2013. 11. Broadberry RE, Lin M.The distribution of the MiIII 3. Popovich BW. Advances for proficiency testing (Gp.Mur) phenotype among the population of Taiwan. genetic laboratory science. Accreditation and Quality Transfusion Medicine. 1996 Jun; 6(2): 145-8. Assurance. 2002; 7(8-9): 351-536. 12. Mak K H, Banks JA, Lubenko A, Chua KM, Torres 4. World Health Organization, Regional Office for South- de Jardine AL, Yan KF. A survey of the incidence of East Asia. Quality assurance in blood transfusion Miltenberger antibodies among Hong Kong Chinese services in SEAR Countries: report of an inter- blood donors. Transfusion. 1994 Mar; 34(3): 238- country workshop, Bangkok, Thailand, 24-28 August 41. 1998. New Delhi: WHO-SEARO, 1999. Document 13. Chaudhary R, Das SS, Ojha S, Khetan D, Sonker No. SEA/HLM/317 - http://www.searo.who.int/EN/ A.The external quality assessment scheme: Five Section10/Section17/Section58/Section219_753. years experience as a participating laboratory. Asian htm - accessed 17 April 2013. Journal of Transfusion Science. 2010 Jan; 4(1): 5. World Health Organization. WHO EQA scheme for 28-30. blood group serology. Geneva: WHO - http://www. 14. World Health Organization. Global database on blood who.int/bloodsafety/quality/external_assessment/ safety: summary report 2011. Geneva: WHO, 2011 immunohaematology/en/index.html accessed 17 – http://www.who.int/bloodsafety/global_database/ April 2013. GDBS_Summary_Report_2011.pdf - accessed 17 6. Harmening Denise M. Modern blood banking and April 2013. transfusion practices. 5th edn. Philadelphia, F.A.: 15. BEQAS. External quality assessment scheme Davis Company, 2005. pp. 108-111. for blood banks in India. http://www.nabh.co/ 7. The Rh System. In: Technical Manual: 50th main/bloodbank/EQAS_General_Information.pdf - anniversary AABB edition 1953-2003. Bethesda, accessed 17 April 2013. Md.: American Association of Blood Banks, 2002. 16. World Health Organization, Regional Office for South- pp. 295-313. East Asia. Quality assurance in health laboratory services: a status report. New Delhi: WHO-SEARO, 2003. Document No. SEA-HLM

WHO South-East Asia Journal of Public Health 2012;1(4):423-431 431 Original research

Prevalence of group A genotype human rotavirus among children with diarrhoea in Nepal, 2009–2011 Jeevan B Sherchanda, W William Schluterb, Jatan B Sherchanc, Sarmila Tandukara, Jyoti R Dhakwad, Ganga R Choudharyb, Chandeshwar Mahasethd

Background: Rotavirus as a causative agent of childhood diarrhea is known to cause serious illness among children less than 5 years of age. This study examined the epidemiology of rotavirus disease burden and diversity of G and P genotypes of rotavirus in Nepal.

Methods: Stool samples were tested for rotavirus by Enzyme Immuno Assay and Group A rotaviruses were detected by using both ELISA and RT-PCR in 2718 samples between 2009 and 2011.

Results: Rotavirus was more frequently detected among inpatients (28.5%) than outpatients (15.2%). Over the three-year study period, 653 (24.4%) cases were positive for rotavirus by ELISA. Genotyping by RT-PCR was done on 638 samples. The most prevalent genotype was G12P [6] (60.4%). Mixed infections were not uncommon (14% in 2009, 29% in 2010 an 7% in 2011). However, 41 were partially typed and 23 were completely untyped over the study period.

Conclusions: This study highlights the rotavirus disease burden and diversity of rotavirus strains circulating in Nepal. Continued sentinel surveillance will provide useful information to policy makers with regard to rotavirus vaccine introduction.

Keywords: Human rotavirus, prevalence, genotype, RT-PCR

Introduction Worldwide, rotavirus is the most common cause rotavirus causes approximately 111 million of severe gastroenteritis among infants and episodes of acute gastroenteritis requiring young children.1,2 While diarrhea is the second home care, 25 million clinic visits, and 2 million most common cause of fatal childhood illness, hospitalizations5 in the U.S. and Europe.3 about 1.34 million deaths occur worldwide Rotavirus belongs to the Reoviridae virus among children aged less than 5 years due to family and the virion comprises of three rotavirus.3,4 Though the incidence of rotavirus concentric protein layers. The outer capsid infection among children in developed and consists of two proteins, VP7 and VP4 that developing countries is similar; outcomes are used to classify rotavirus strains into vary widely with 82% of fatalities estimated G (glycoprotein) and P (protease sensitive) to occur in developing countries.5 Each year,

aTribhuvan University Institute of Medicine, Public Health Research Laboratory and Microbiology. bProgram for Immunization Preventable Diseases (IPD), Kathmandu, Nepal. cDepartment of Microbiology, Kathmandu University School of Medical Sciences, Dhulikhel, Nepal. dKanti Children’s Hospital, Maharajgunj, Kathmandu, Nepal. Correspondence to Jeevan B. Sherchand (e-mail: [email protected])

432 WHO South-East Asia Journal of Public Health 2012;1(4):432-440 Jeevan B Sherchand et al. Prevalence of group A genotype human rotavirus among children with diarrhea in Nepal, 2009-2011

genotypes, respectively.7,8 Of the 12 G and was explained to the caregivers and written 15 P genotypes known to infect humans, informed consent obtained. genotypes G1P[8], G2P[4], G3P[8], G4P[8] Study setting, sample and data and G9P[8] cause over 90% of rotavirus collection: This study was conducted between disease worldwide.9 January 2009 to December 2011 in Tribhuvan Two effective rotavirus vaccines, a single- University, Institute of Medicine, Public strain attenuated human rotavirus vaccine Health Research Laboratory, Maharajgunj, (Rotarix, GlaxoSmithKline Biological) and Kathmandu, Nepal. a multi-strain bovine-human reassortant Case enrollment was done on the basis vaccine (RotaTeq, Merck and Company) of inclusion and exclusion criteria among are now available. WHO recommends the children visiting Kanti Children’s hospital with inclusion of rotavirus vaccines in all national symptoms of acute watery diarrhea, fever, routine immunization schedules. In countries vomiting, and abdominal pain. All children less where diarrheal deaths account for ≥10% of than five years of age who were admitted for mortality among children <5 years of age, the treatment of acute gastroenteritis diarrhoea introduction of rotavirus vaccine is strongly and/or vomiting with or without intake of recommended.10 Efficacy of these vaccines medications were included. Diarrhoea was has ranged from 80% to 98% in industrialized defined by the occurrence of three or more countries, including Latin America, and 39% liquid stools in a 24-hour period and presence to 77% in developing countries, such as Africa of diarrhea at the time of clinical presentation and Asia.3,4 with admission to the inpatient diarrheal The present study summarizes the treatment unit. The exclusion criteria were findings from three years of hospital-based as follows: Hospital-acquired diarrhea, which rotavirus sentinel surveillance in Nepal and was defined as onset of diarrhea more than 48 the epidemiology of rotavirus gastroenteritis hours after hospitalization; bloody diarrhea; and characterizes circulating rotavirus strains and chronic and/or persistent diarrhea, which among children aged less than five years in was defined as diarrhea that lasted for more Nepal. than two weeks.

Sampling: A total of 2718 stool samples Material and methods were collected. Specimens were stored at −70°C until analysed using ELISA and Clinical methods molecular methods. All case-based data Children presenting to the pediatric emergency and clinical information was recorded on a unit of Kanti Children’s hospital were screened standardized questionnaire and was analyzed for diarrhea by the treating medical officer using EPI-Info software. or nurse. Children aged less than 60 months Measurement variables: Data were with acute diarrhea and from whom stool collected using a pre-coded data collection sample could be recovered within 24 hours of tool. Variables included child’s age, sex, hospitalization were recruited. Ethical approval symptoms, and duration of symptoms, water was obtained from the Institutional Review source, and home hand washing practices, Board (IRB), Institute of Medicine, Tribhuvan diarrhea contacts, housing facilities, education University, Kathmandu, Nepal. The study

WHO South-East Asia Journal of Public Health 2012;1(4):432-440 433 Prevalence of group A genotype human rotavirus among children Jeevan B Sherchand et al. with diarrhea in Nepal, 2009-2011

and occupation of the caregiver, temperature Results of the child, hydration status, and rotavirus results. Epidemiology of Rotavirus From January 2009 to December 2011, 2718 Rotavirus detection: The specimens cases meeting the enrollment criteria were were tested by a solid-phase sandwich-type evaluated for rotavirus. Of these 653 (24.0%) enzyme immunoassay method (Rotavirus Ag cases were positive for rotavirus by EIA, ELISA, Pro Spect, USA). among the positive cases, 638 samples were processed for genotyping by RT-PCR. Rotavirus Genotyping Detection of rotaviruses in clinical specimens Clinical Presentation among and determination of the G-type and P-type Hospitalized Children was accomplished by extraction of the viral Abdominal pain was the most common RNA from fecal specimens and analysis symptom among the 2718 children with by reverse-transcription polymerase chain diarrhea. (See Table 1). Approximately 25% reaction (RT-PCR) with primers specific for the of cases with each of the evaluated symptoms VP7 genes of G serotypes 1,2,3,4, 8, 9, 10, were found to be positive for rotavirus. The and 12 and VP4 genes of P serotypes 4, 6, 8, degree of dehydration among rotavirus cases 9, 10 and 11. Rotavirus genotyping was done was more likely to be moderate to severe in collaboration with Christian Medical College, rather than mild. Vellore – Rotavirus Regional Reference Laboratory and WHO Collaborating Centre, Distribution of rotavirus by EIA India. A total of 638 samples were rotavirus positive [2009 (n=221), 2010 (n=151), and Rotavirus was detected by EIA among 25.4% 2011 (n= 266)] and were analyzed by RT- (230/906) in 2009, among 23.9% (195/815) PCR method described by Gouvea et al12 and in 2010 and among 22.7% (228/997) in 2011. Gunasena et al.13 The predominant strain was G12P [6] during all three years of the study period, (45.7% in

Table 1: Clinical presentation in hospitalized patients

Symptoms No. cases Rotavirus positive cases Abdominal pain 2439 (98.7%) 583 (23.9%) Fever 1392 (51.2%) 327 (23.4%) Nausea 1310 (48.1%) 352 (26.8%) Vomiting 1315 (48.3%) 367 (27.9%) Degree of dehydration Mild 261 (9.6%) 34 (13.02%) Moderate 1881 (66.7%) 487 (25.8%) Severe 535 (19.7%) 123 (22.9%)

434 WHO South-East Asia Journal of Public Health 2012;1(4):432-440 Jeevan B Sherchand et al. Prevalence of group A genotype human rotavirus among children with diarrhea in Nepal, 2009-2011

2009; 28.5% in 2010, and 30.1% in 2011). Age and year-wise distribution Previously unidentified strains were also of rotavirus positive cases in detected, such as G9P [6] was identified in hospitalized children 2010 (6%) and G12P [4] was first identified Hospitalization and enrollment of children 0 to in 2011 (1.1%). There were a number of 11 months was more common than among diarrheal cases with mixed infections (14.0% older aged children. In 2009, rotavirus was in 2009 and 29.8% in 2010 and 6.7% in most commonly identified (28.1%) among 2011). Forty-one samples were partially typed children 0 to 11 months, whereas the rates and 23 were completely untyped over the of rotavirus as the cause of diarrhea among three-year period. hospitalized children was more common among children 12–23 months of age in 2009 Month wise distribution of (35.5%) and in 2011 (26.7%). rotavirus positive in hospitalized children Distribution of major Rotavirus among hospitalized children showed combination of G and P types of a seasonal distribution with the highest percent rotavirus in hospitalized children of diarrhea cases caused by rotavirus during The most common circulating genotype in the winter months from December to March. the population under surveillance was G12P The month with the highest percentage of [6] (45.7% in 2009, 28.5% in 2010 and 30% rotavirus positive cases was January (38.5%) in 2011) over the three-year investigation, in 2009, February (53.4%) in 2010, and April though the mixed type was slightly higher in (34.9%) in 2011. 2010 (29.8%).

Figure 1: Month-wise distribution of rotavirus from 2009 to 2011

WHO South-East Asia Journal of Public Health 2012;1(4):432-440 435 Prevalence of group A genotype human rotavirus among children Jeevan B Sherchand et al. with diarrhea in Nepal, 2009-2011

Figure 2: Age and year-wise distribution of rotavirus

Discussion with rotavirus infection compared to other causes of diarrhea11. Rotavirus was a major Diarrhea continues to be a major public cause of pediatric gastroenteritis and was the health problem and leading cause of death in underlying etiology of diarrhea for one-fourth developing countries.11-16 The present study is of children 6 to 24 months of age hospitalized a continuation of a prospective study among with an acute diarrheal illness.23 children less than five years of age with diarrhea evaluated at Kanti Children’s Hospital, which The present study found a clear seasonal is the only government-supported children’s pattern of acute rotavirus gastroenteritis that hospital in Nepal. The majority of the children peaked in winter, i.e., December to February who tested positive for rotavirus antigen had (31.3%, 35.4%, and 30.6% in year 2009, clinical features of diarrhea, vomiting, and 2010, and 2011, respectively) followed by moderate to severe dehydration. Overall, the early spring i.e. March to May, (31.3%, prevalence of rotavirus during the three-year 32.9%, and 26.1% in 2009, 2010, and 2011, study period was 24.0% which is similar to respectively) with only 15.3% in both summer previously published studies from the same (June to August) and autumn (September to institution,17-22 but comparatively lower than November), which is in contrast to similar evaluations at other study sites.2,5,7,8 Fever study conducted in Nepal 2006.19 and vomiting were more often associated

436 WHO South-East Asia Journal of Public Health 2012;1(4):432-440 Jeevan B Sherchand et al. Prevalence of group A genotype human rotavirus among children with diarrhea in Nepal, 2009-2011

Table 2: Distribution Rotavirus genotyping and combination of G and P types of rotavirus in hospitalized children ( 2009, 2010 and 2011)

Rotavirus genotyping

Year Genotypes Total 2009 2010 2011 G1P[8] 6 13 34 53 (8.3%) G12P[4] 0 0 3 3 (0.5%) G12P[6] 101 43 80 224 (35.1%) G12P[8] 7 4 2 13 (2.0%) G2P[4] 32 7 74 113 (17.7%) G1P[4] 0 0 0 0 G1P[6] 9 3 27 39 (6.1%) G2P[8] 0 0 0 0 G2P[6] 0 0 0 0 G3P[8] 0 0 0 0 G9P[8] 3 3 4 10 (1.6%) G9P[4] 2 3 10 15 (2.4%) G9P[6] 0 10 0 10 (1.6%) Mixed 31 45 18 94 (14.7%) Partially typed 16 19 6 41 (6.4%) Untyped 14 1 8 23 (3.6%) Total 221 151 266 638

Global epidemiologic surveys have of all specimens. This strain was uncommon identified G1P [8], G2P [4], G3P [8], G4P [8], in 2009 (2.7%). G1P [6] was the fourth most and G9P [8] as the most common G/P common type identified but was uncommon genotype combinations associated with in 2010 (2.0%), but relatively more common diarrhea in humans. However, recent studies in 2011 (10.2%). Thus, variations were found in developing countries have shown increased to occur in genotype distribution from year to identification of a high proportion of rotavirus year over the three-year study period. Results strains with unusual G/P combinations, which from other countries differ from ours.2,3,6,8,26 In may have implications for vaccine efficacy.24,25 this study, the strain G9P6 that occurred only The genotypic distribution pattern from this in year 2010 (6.6%) and was the predominant study identified G12P[6] as the predominant strain in a study from India in 1996.27 The strain (35.1%), with G2P[4] strains as the incidence of G9 with either P[8] or P[6] was second most common genotype identified also reported by Jain et al in 2001 to be 17% (17.7%) over the three years study. G1P[8], in stool samples collected from hospitals in which is the single strain present in the Bhopal, New Delhi, Davengere, Lucknow, licensed Rotrix oral vaccine9 was the third Nagpur and Shimla from 1996 to 1998.28 most common strain identified among 8.3%

WHO South-East Asia Journal of Public Health 2012;1(4):432-440 437 Prevalence of group A genotype human rotavirus among children Jeevan B Sherchand et al. with diarrhea in Nepal, 2009-2011

G12 genotype was first detected in the admitted less often to referral centers. Thirdly, Philippines,6 and appears to be more prevalent this study was based on a single sentinel in recent years in developing countries.29 This site representing a catchment population of study demonstrates that a variety of different about three million, which does not represent rotavirus types circulate simultaneously in complete scenario of the nation to ward Nepal. rotavirus disease burden in Nepal.

The observation of G and P strains identified This year-round surveillance of rotavirus in this study suggests the circulation of various genotypes will contribute to identifying rotavirus strains in Nepal and may help to important and emerging strains of rotavirus characterize the antigenic variation of the among Nepalese children. Continuation and rotavirus. In the present study, the third most expansion of laboratory-based rotavirus frequent circulating genotypes in the population sentinel surveillance in other areas of Nepal under surveillance were mixed genotypes. High is likely to yield important information for mixed infection with different rotavirus strains improved public health control of rotavirus may reflect frequent contamination of water associated outbreaks, as well as aid in the sources with rotavirus strains that facilitate tracking of changes in strains that cause generation of novel rotavirus strains through serious diarrhea among infants and young a re-assortment process.29 Non-typeable children. rotavirus strains were present in 3.6% over the three-year study period. Partially typed Conclusions virus strains were identified in 6.4% of children with acute diarrhea. The presence There is high disease burden of rotavirus of mixed genotypes may indicate infection gastroenteritis among children less than five because of poor sanitation or transmission of years of age. Genotyping analysis highlights recombinant strains transferred from animals that the existence of significant diversity to humans. 30, 31 of rotavirus strains with unusual G and P combinations. Therefore, rotavirus sentinel Our study has some limitations. Firstly, surveillance and genotyping will provide in the laboratory, we did not examine for important background information for decision- all pathogens but only for rotavirus and making about whether to introduce rotavirus enteric adenoviruses in addition to pathogens vaccine into Nepal National Immunization for which routine screening is performed Program. according to hospital protocols. Results of these investigations have provided evidence of the burden of disease associated with Recommendation rotavirus, but did not determine an etiology In Nepal, we have no information about the for nearly half of all diarrheal cases. Additional medical and social costs either to parents, examinations for noroviruses, sapoviruses, for treatment of their children when they astroviruses, and a variety of enteric bacteria are ill with rotavirus diarrhea, or to the would be needed to complete this picture. health sector, for the direct medical care Secondly, the sentinel-hospitals the only provided by the government. Such an analysis government children’s hospital in the country might help advance the case for a rotavirus and may not be representative of admissions vaccine and establish the economic value to smaller or private hospitals. It is possible of routine immunization. In addition, we that children with uncomplicated diarrhea are have no exact information about the cost of

438 WHO South-East Asia Journal of Public Health 2012;1(4):432-440 Jeevan B Sherchand et al. Prevalence of group A genotype human rotavirus among children with diarrhea in Nepal, 2009-2011

vaccines currently licensing (Rota Teq; Merck 5. Bonkoungou IJO, Damanka S, Sanou I, Tiendre and RotaRix; GlaxoSmithKline) and their F, Coulibaly SO, Bon F , Haukka K, Traore AS, effectiveness in developing Asian countries Barro N, Armah GE. Genotype diversity of group a rotavirus strains in children with acute diarrhea in like Nepal. Hence, we recommend continuing urban Burkina Faso, 2008–2010. Journal of Medical the rotavirus surveillance in other sentinel Virology. 2011; 83:1485–90. site of Nepal to obtain more information and 6. Kobayashi N, Ishino M, Wang YH, Chawla-Sarkar to represent national scenario of rotavirus M, Krishnan T, and Naik TN. Diversity of G-type disease burden estimation, which can strongly and P-type of human and animal rotaviruses and support-evidence based decision making its genetic background. Communicating Current on vaccine introduction, monitor serotype/ Research and Educational Topics and Trends in genotype distribution, and allow evaluation of Applied Microbiology. 2007. pp. 847-58. vaccine impact after vaccine introduction. 7. Nakawesi JS, Wobudeya E, Ndeezi G, Mworozi EA and Tumwine JK. Prevalence and factors associated with rotavirus infection among children admitted with Acknowledgements acute diarrhea in Uganda. Nakawesi et al. BioMed We express our profound gratitude to WHO Central Pediatrics. 2010; 10(62): 2-5. for supporting this study. We are also grateful 8. Bourdett-Stanziola L, Jiménez CDR, and Ortega-Barria to WHO Collaborating Center, CMC Vellore for E. Diversity of human rotavirus G and P genotypes in Panama, Costa Rica, and the Dominican Republic confirmation of genotyping samples. American Journal Tropical Medicine. Hygiene. 2008; 79(6): 921–4. Conflict of interest 9. Donato C, Molecular epidemiology of rotavirus in the None to declare era of vaccination. Microbiology Australia. 2012. pp. 64-66 – http://journals.cambridgepublishing.com. au/UserDir/CambridgeJournal/Articles/06donato411. pdf - accessed 17 April 2013.

References 10. Rotavirus vaccine: an update. Weekly Epidemiological 1. Payne DC, Wikswo M, Parashar UD. VPD surveillance Record. 2009; 84: No. 51-52 - http://www.who. manual. 5th edn. 2011 Rotavirus: Chapter 13-1 - int/wer/2009/wer8451_52.pdf - accessed 17 April http://www.cdc.gov/vaccines/pubs/surv-manual/ 2013. chpt13-rotavirus.pdf - accessed 17 April 2013. 11. Mangayarkarasi V, Prema A, Gunasekaran P, Babu 2. Kargar M, Najafi A, Zandi K, Hashemizadeh Z. BVS and Kaveri K. A Unique human rotavirus (non Genotypic distribution of rotavirus strains causing vaccine) G9P4 genotype infection in a child with severe gastroenteritis in children under 5 years old gastroenteritis. Indian pediatrics. 2012; 49: 569- in Borazjan, Iran. African Journal of Microbiology 71. Research. 2011; 5(19): 2936-41. 12. Tate JE, Burton AH, Boschi-Pinto C, Steele AD, Duque 3. Delogu RR, Petouchoff T, Tcheremenskaia O, Petris J, Parashar UD, and the WHO-coordinated Global SD, Fiore L and the RotaNet-Italy. Study group Rotavirus Surveillance Network. 2008 estimate of molecular characterization of rotavirus strains worldwide rotavirus-associated mortality in children from children with diarrhea in Italy, 2007–2009. younger than 5 years before the introduction of Franco Maria. Journal of Medical Virology. 2011; 83: universal rotavirus vaccination programmes: a 1657–68. systematic review and meta-analysis. Lancet 4. Patel MM, Steele D, Gentsch JR, Wecker J, Glass RI Infect Dis. 2012 Feb; 136-41. DOI:10.1016/S1473- and Parashar UD. Real-world Impact of Rotavirus 3099(11)70253-5. Vaccination. The Pediatric Infectious Disease Journal. 13. Kheyami AM, Nakagomi T, Nakagomi O, Dove W, 2011; 30(1): 1-5. Hart CA, and Cunliffe NA. Molecular epidemiology of rotavirus diarrhea among children in Saudi Arabia:

WHO South-East Asia Journal of Public Health 2012;1(4):432-440 439 Prevalence of group A genotype human rotavirus among children Jeevan B Sherchand et al. with diarrhea in Nepal, 2009-2011

first detection of G9 and G12 strains. J Clin Microbiol. EM, Pazvakavambwa I,Armah GE, Seheri LM, Kiulia 2008; 46(4): 1185–91. NM, Widdowson MA and Steele AD. Burden and Epidemiology of Rotavirus Diarrhea in Selected 14. Gilany AH and Hammad S. Epidemiology of diarrheal African Countries: Preliminary Results from the diseases among children under age 5 years in African Rotavirus Surveillance Network. J Infect Dakahlia, Egypt. East Mediterr Health J. 2005; 11(4): Dis.2010; 202 (1): Suppl. SS-S11. 762-75. 23. Kirkwood CD, Roczo S, Boniface K, Bishop RF, 15. Bresee J, Fang ZY, Wang B, Nelson EAS, Tam J, Barnes GL, Australian Rotavirus Surveillance Group. Soenarto Y, Wilopo SA, Kilgore P, Kim JS, Kang JO, Australian Rotavirus Surveillance Program, 2010/2011 Lan WS, Gaik CL, Moe K, Chen KT, Jiraphongsa C, Annual reports. CDI. 2011; 35(4): 281-7 Pongsuwanna Y, Man NV, Tu PV, Luan LT, Hummelman E, Gentsch JR, Glass R and the members of the Asian 24. Urbina D, Rodríguez JG, Arzuza O, Parra E, Young G, Rotavirus Surveillance Network. First report from the Castro R, Portillo P. G and P genotypes of rotavirus Asian Rotavirus Surveillance Network. Emerg Infect circulating among children with diarrhea in the Dis. 2004; 10(6): 988-995. Colombian northern coast. Int. Microbiol. 2004 Jun; 7(2): 113-20. 16. Ward KA, McIntyre PB, Kirkwood CD, Roche PW, Ferson MJ, Buynder PGV, Witteveen ARR, Kesson AM, 25. Linhares JA, Stupka A, Ciapponi AE, Bardach D, Krause VL, McAnulty JM. Rotavirus surveillance in Glujovsky PK Aruj A, Mazzoni JA, Colindres R. Burden Australia. Commun Dis. Intell. 2008; 32(1): 82-7. and typing of rotavirus group A in Latin America and the Caribbean: systematic review and meta-analysis 17. Sherchand JB, Nakagomi O, Dove W, Nakagomi Alexandre C. 2011; John Wiley & Sons, Ltd. Rev. Med. T, Yokoo M, Pandey BD, Cuevas LE, Hart A and Virol. 2011 Mar; 21(2): 89-109. Cunliffe NA. Molecular epidemiology of rotavirus diarrhea among children aged <5 years in Nepal: 26. Lodha R and Shah D. Prevention of rotavirus diarrhea predominance of emergent G12 strains during 2 in India: is vaccination the only strategy? Indian years. J Infect Dis. 2009; 200: 4222-6. pediatrics. 2012; 49: 441-3.

18. Pun SB, Nakagomi T, Sherchan JB, Pandey BD, 27. Ramachandran M, Das BK, Vij A, Kumar R, Bhambal Cuevas LE, Cunliffe NA, et al. Detection of G 12 SS, Kesari N, Rawat H, Bahl L, Thakur S, Woods PA, human rotavirus in Nepal. Emerg Infect Dis. 2007; Glass RI, Bhan M and Gentsch J. Unusual Diversity 13: 482-83. of Human Rotavirus G and P Genotypes in India. J of Clinical Microbiology. 1996; 34(2): 436–439. 19. Uchida R., Pandey BD, Sherchand JB, Ahmed K, Yokoo M, Nakagomi T, et al. Molecular epidemiology 28. Jain V, Dass BK, Bhan MK, Glass RI, Gentsch JR. of rotavirus diarrhea among children and adults in Great diversity of group A rotavirus strains and high Nepal: detection of G12 strains with P[6] or P[8] prevalence of mixed rotavirus infections in India. J of and a G11P[25] strain. J Clin Microbiol. 2006; 44: Clinical Microbiology. 2001; 39: 3524-3529. 3499–505. 29. Ogilvie I, Khoury H, Khoury AC and Goetghebeur MM. 20. Sherchand JB, Cunliffe NA, Tandukar S, Yokoo Burden of rotavirus gastroenteritis in the pediatric M, Pandey BD, Niraula P, Panta AR, Nakagomi O. population in Central and Eastern Europe. Human Rotavirus disease burden and molecular epidemiology Vaccines. 2011: 7: (5): 523-533. in children with acute diarrhea age less than 5 Years 30. Gentsch JR, Laird AR, Bielfelt B, Griffin DD, in Nepal. J. Nepal Paediatr. Soc. 2011; 31(3): 215- Ramachandran M, Cunliffe NA, Nakagomi O, Kirkwood 18. CD, Parashar UD, Bresee JS, Jiang B and Glass 21. Parashar UD, Holman RC, Clarke MJ, Bresee JS, and RI. Serotype dversity and reassortment between Glass RI. Hospitalizations associated with rotavirus human and animal rotavirus strains: implications diarrhea in the United States, 1993 through 1995: for Rotavirus vaccine programs. J Infect Dis. 2005; surveillance based on the new ICD-9-CM rotavirus- 192: 146–159. specific. J Infect Dis. 1998; 177: 13–7. 31. Sherchand JB, and Haruki K. Rotavirus diarrhea in 22. Mwenda JM, Ntoto KM, Abebe A, Enweronu-Laryea children and animals of urban and rural Nepal. J. C, Amina I, Mchomvu J, Kisakye A, Mpabalwani Nepal Health Res. Council. 2004; 2: 5-8.

440 WHO South-East Asia Journal of Public Health 2012;1(4):432-440 Research brief

Evidence of HPV subtypes linked with cervical cancer in Nepal

Chop L Bhusala, Sulochana Manandharb, Meeta Singhc, Aarati Shahd, Sushharma Neupanea, Dibesh Karmacharyab, Kate Cuschierie, Heather Cubiee, Duncan C Gilbertf, Sameer M Dixitb

Objectives: Cervical cancer is the commonest malignancy among women in Nepal but data are limited on which subtypes of human papillomavirus (HPV) are associated with cancer in this population. Now that vaccines against HPV types 16 and 18 are available, this evidence is of vital importance in obtaining further support for a vaccination programme.

Methods: Cervical swabs from 44 histologically confirmed invasive cervical cancer cases were obtained from two tertiary referral hospitals in Nepal. Evidence of HPV subtypes was identified using an HPV multiplex polymerase chain reaction (PCR), and confirmed at the Scottish HPV Virus Reference Laboratory.

Results: HPV types 16 and 18 were present in 70% of samples, along with other high-risk subtypes. HPV 6 and 11 were not observed. Epidemiological data assessment appeared to indicate that patient age, age of marriage and age of first pregnancy were associated with increased HPV infection in patients.

Conclusions: This study provides further evidence of the importance of HPV types 16 and 18 in cervical cancer in Nepal and adds support to a nationwide vaccination programme and the use of HPV detection in screening programmes.

Key words: Cervical cancer, HPV, Nepal, multiplex polymerase chain reaction

Introduction Cancer of the cervix is responsible for the mortality,2 cancers in Nepal tend to present greatest number of cancer deaths among at advanced stage with significant associated women in Nepal,1 with up to 2000 deaths per morbidity and mortality. annum. Unlike in more developed countries, It is appreciated that worldwide the where comprehensive screening programmes majority of cervical cancers are associated with have contributed to reduced incidence and infection with high-risk, oncogenic subtypes

a Nepal Health Research Council, Nepal. b Centre for Molecular Dynamics Nepal, Nepal. c Institute of Medicine, Nepal. d National Academy of Medical Sciences, Nepal. e Scottish HPV Viral Reference Laboratory, Edinburgh, United Kingdom. f Brighton and Sussex Medical School, Brighton, United Kingdom. Correspondence to Sameer M Dixit (e-mail: [email protected])

WHO South-East Asia Journal of Public Health 2012;1(4):441-445 441 Evidence of HPV subtypes linked with cervical cancer in Nepal Chop L Bhusal et al.

of human papillomavirus (HPV),3,4 acquired China). Briefly cells were lysed by lysis buffer through sexual contact. Viral oncogenes VL followed by vortexing. DNA was precipitated E6 and E7 downregulate p53 and pRb, and captured on spin columns, then washed respectively, resulting in cellular proliferation with buffers RBW and RNW, each followed by and a malignant phenotype. centrifugation. Finally the purified DNA was eluted in DNAse/RNAse free water. The DNA The development of vaccines for HPV5,6 has was stored at -20 °C until used for polymerase brought into focus the precise nature of this chain reaction (PCR). association in terms of which subtypes are prevalent in which populations. Preliminary HPV-specific multiplex PCR was carried data from Nepal from 54 women with invasive out using Seeplex HPV4A ACE screening kit. cervical cancer demonstrated HPV-16 in 68.5% Following the manufacturer’s instructions,3 μl and HPV-18 in 22.2%,the next most frequent of undiluted extracted DNA was used as subtype being HPV-45 (in 5.6% patients).7 template DNA in sample tubes to make the This is broadly in keeping with similar final volume of 20 μl. Negative and positive population studies from India8 and Pakistan9 controls provided were included in each assay. and in a wider context the United States of PCR thermocycling was performed: 5 μl of America and the United Kingdom. Provisional each PCR product was electrophoresed in experiences of instigating an HPV vaccination 2% agarose gel along with the DNA marker programme in Nepal have been encouraging10 provided with the kit. The result was visualized with 99.3% of 1096 girls completing the under ultraviolet illumination and the bands course of 3 doses of Gardasil®. were analysed in the reference of marker DNA (Figure 1). The result was considered valid To further support the introduction of HPV only in the presence of the distinct band of the vaccination in Nepal, we present here data internal control indicating PCR success. on the presence of oncogenic high-risk HPV subtypes from 44 invasive squamous cell Results for HPV types 16 and 18 were carcinomas of the uterine cervix from Nepal. validated and samples further genotyped at the Scottish HPV Virus Reference Laboratory, Methods Edinburgh.Samples were assessed using a multimetrix HPV genotyping kit. This work has approval from the Ethical Review Board of the Nepal Health Research Council, Ministry of Health and Population, Government Results of Nepal. Of the 44 cervical samples analysed, 82% (36) showed HPV infection by PCR, of which Cervical swab samples were collected from 80% were HPV-16 or 18. Eleven HPV types, the cervix of women fulfilling inclusion criteria including nine high-risk groups, were observed using sterile cytobrushes and immediately in this study. Types 42 and 43 are not inserted into a container containing phosphate considered high-risk but were identified in the buffer saline (PBS) as transport medium samples. HPV 16 and 18 DNA was identified and stored at 4–8 °C. Sample vials were in 68% of the total sample and 83% of HPV brought to room temperature and vortexed positive samples. The prevalence of other vigorously to detach the exfoliated cells. types in the population were observed at a Deoxyribonucleic acid (DNA) was extracted much lower percentage (Table 1). using the GeneALLRibo-Spin VRD Kit (GeneAll,

442 WHO South-East Asia Journal of Public Health 2012;1(4):441-445 Chop L Bhusal et al. Evidence of HPV subtypes linked with cervical cancer in Nepal

Figure 1: Representative agarose gel electrophoresis image showing PCR products for HPV types 16 and 18 and a combined high-risk group

Table 1: HPV types observed among In terms of histological subtypes, 35 of all cervical samples tested 44 were squamous cell carcinomas. Of these, in this study 78% were positive for HPV DNA while for adenocarcinoma (n=9), HPV infection rates Type Frequency of % were 22%. The majority of cases in this cohort detection (n=44) (n=39) were stage I or II with only five cases HPV 16 22 50 in advanced stage (III/IV) carcinoma. HPV 18 8 18 The majority of patients in this study had HPV 31 1 2 married at an early age with 29 of the 41 HPV 33 2 5 patients, where data were available, having HPV 42 2 5 been married prior to the age of 18 (Nepal’s HPV 43 3 7 legal age of marriage until 2002 was 16). Out of these 29 patients, 25 had evidence of HPV HPV 45 2 5 infection (86%). Eight of the 12 respondents HPV 52 2 5 who reported marrying later than aged 18 HPV 53 3 7 showed evidence of HPV (67%), though this HPV 59 3 7 does not reach statistical significance (Fisher’s HPV 66 1 2 exact test p=0.2). More respondents reported having a child under (n=22), as opposed to No result 8 18

WHO South-East Asia Journal of Public Health 2012;1(4):441-445 443 Evidence of HPV subtypes linked with cervical cancer in Nepal Chop L Bhusal et al.

older (n=19) than 18 years of age. Within previous study reported low rates of compliance the two groups, those that became pregnant with colposcopy following abnormal screening at an earlier age had a higher HPV infection results.7 As a result, mortality and morbidity prevalence (61%) than those in the second from cervical cancer remain unacceptably group (39%). high. The development and successful clinical introduction of HPV vaccines, however, present Discussion a major opportunity to make inroads into cancer prevention in Nepal.11 The data presented here confirm earlier findings that the majority of cancers of Furthermore, HPV DNA typing can benefit the cervix in Nepal are associated with screening programmes.DNA typing in cervical HPV infection, and furthermore that HPV- cancer screening is used in many countries; 16 and -18 represent the major subtypes results from large randomized controlled trials responsible. Specifically, 36 of 44 (82%) cases testing for HPV DNA suggest that most women demonstrated the presence of HPV infection. with cervical cytologic abnormalities should be HPV-16 was present in 22 cases (61% of assessed using this form of screening.12 The HPV-positive tumours) and HPV-18 in 8 cases role of HPV screening in neighbouring India (22%);the remaining cases demonstrated the has been widely discussed.13-15 involvement of multiple subtypes. In summary, these results are consistent Combined with the limited previous data,7 with those from elsewhere in South Asia this comprises 98 cases of invasive cervical and the wider world and should provide cancer from Nepal. Of these, 83 (85%) were further support for the introduction of an associated with HPV infection, namely HPV-16 HPV vaccination programme across Nepal. in 59 cases (71% HPV positive, 60% all cervical Work to assess HPV infection rates in the cancers) and HPV-18 in 20 cases (24% and healthy female population of Nepal should be 20% respectively). Therefore, 80% of cervical extended to identify high-risk HPV types in cancers in Nepal are theoretically preventable the population and facilitate early screening to with currently available vaccines. achieve comprehensive prevention of cancer of the cervix. Marriage and pregnancy at an early age both appear to be linked to a higher likelihood of HPV infection in cervical cancer cases in this Acknowledgements study. This concurs with data from a previous The authors wish to thank the Nepal Health publication, where early age at first sexual Research Council (NHRC) for funding this intercourse (AFSI), early age at first marriage research. We acknowledge the inputs of Dr SP (AFM) and early age at first pregnancy Singh, Member Secretary, and Mr Purushottam (AFP) were linked to higher incidence of HPV Dhakal, Biostatistician, NHRC. We also thank infection and cervical carcinoma in women of the National Academy of Medical Sciences and developing countries. the Institute of Medicine for helping acquire samples for the study. Our appreciation While effective screening programmes is extended to the entire team at Intrepid have reduced morbidity and mortality from Nepal Pvt Ltd for facilitating the laboratory cancer of the cervix in the developed world,2 a requirements of this project. multitude of challenges beset the introduction of such a programme across Nepal. Indeed,a

444 WHO South-East Asia Journal of Public Health 2012;1(4):441-445 Chop L Bhusal et al. Evidence of HPV subtypes linked with cervical cancer in Nepal

References 8. Pillai RM, Babu JM, Jissa VT, Lakshmi S, Chiplunkar SV, Patkar M, et al. Region-wise distribution of 1. Pradhananga KK, Baral M, Shrestha BM. Multi high-risk human papillomavirus types in squamous institution hospital-based cancer incidence data for cell carcinomas of the cervix in India. Int J Gynecol Nepal: an initial report. Asian Pac J Cancer Prev. Cancer. 2010; 20(6): 1046-51. 2009; 10: 259-62. 9. Raza SA, Franceschi S, Pallardy S, Malik FR, Avan 2. Peto J, Gilham C, Fletcher O, Matthews FE. The cervical BI, Zafar A, et al. Human Papillomavirus infection in cancer epidemic that screening has prevented in the women with and without cervical cancer in Karachi, UK. Lancet. 2004; 364(9430): 249-56. Pakistan. 2010. British Journal of Cancer. 2010; 102, 3. Walboomers JM, Jacobs MV, Manos MM, Bosch FX, 1657–1660. Kummer JA, Shah KV, et al. Human papillomavirus 10. Singh Y, Shah A, Singh M, Verma S, Shrestha BM, is a necessary cause of invasive cervical cancer Vaidya P, et al. Human papilloma virus vaccination worldwide. J Path. 1999; 189: 12-19. in Nepal: an initial experience in Nepal. Asian Pac J 4. zur Hausen H. Papillomaviruses and cancer: from Cancer Prev. 2010; 11(3): 615-7. basic studies to clinical application. Nat Rev Cancer. 11. Gilbert DC. Human papilloma virus and cancer of the 2002; 2: 342–350. cervix: A challenge facing Nepal. Kath Univ Med J. 5. Joura EA, Leodolter S, Hernandez-Avila M, Wheeler 2006; 15, 396-397. CM, Perez G, Koutsky LA, et al. Efficacy of a 12. Goldie SJ, Kuhn L, Denny L, Pollack A, Wright TC. quadrivalent prophylactic human papillomavirus Policy analysis of cervical cancer screening strategies (types 6, 11, 16, and 18) L1 virus-like-particle in low-resource settings: clinical benefits and cost- vaccine against high-grade vulval and vaginal lesions: effectiveness. JAMA. 2001; 285(24): 3107-15. a combined analysis of three randomised clinical trials. Lancet. 2007; 369: 1693-702. 13. Sankaranaryanan R, Budukh AM, and Rajkumar R. Effective screening programmes for cervical cancer 6. Ault KA. Effect of prophylactic human papillomavirus in low- and middle-income developing countries. Bull L1 virus-like-particle vaccine on risk of cervical World Health Organ. 2001; 79, 954-962. intraepithelial neoplasia grade 2, grade 3, and adenocarcinoma in situ: a combined analysis of 14. Sankaranarayanan R, Bhatla N, Gravitt PE, Basu four randomised clinical trials. Lancet. 2007; 369: P, Esmy PO, Ashrafunnessa KS, et al. Human 1861-8. papillomavirus infection and cervical cancer prevention in India, Bangladesh, Sri Lanka and Nepal. Vaccine. 7. Sherpa AT, Clifford GM, Vaccarella S, Shrestha S, 2008; 26S: M43-52. Nygård M, Karki BS, et al. Human papillomavirus infection in women with and without cervical cancer 15. Sankaranarayanan R, Nene BM, Shastri SS, Jayant in Nepal. Cancer Causes Control. 2010; 21(3): 323- K, Muwonge R, Budukh AM, et al. HPV screening for 30. cervical cancer in rural India. N Engl J Med. 2009; 360, 1385-94.

WHO South-East Asia Journal of Public Health 2012;1(4):441-445 445 Policy and practice

Institutionalizing district level infant death review: an experience from southern India Sanjeev Upadhyayaa, Sudeep Shettyb, Selva S Kumarc, Amol Dongred, Pradeep Deshmukhe

Background: An Infant Death Review (IDR) programme was developed and implemented in two districts of Karnataka.

Objective: We explored the processes that led to the development of the IDR programme with a view to improving the existing pilot programme and to ensuring its sustainability.

Methods: A sequential mixed-methods design was followed in which quantitative data collection (secondary data) was followed by qualitative data collection (in-depth interviews). Quantitative data were entered using EpiInfo (version 3.5.1) software and qualitative data were analysed manually.

Results: Apart from ascertaining the cause of infant deaths, the IDR Committee discusses social, economic, behavioural and health system issues that potentially contribute to the deaths. As a result of the IDR programme, key actors perceived an improvement in infant death reporting at district level, the development of a rapport with the local community, and elaboration of a feedback system for corrective actions. This has led to improved health care during pregnancy.

Conclusions: We found that involvement of the different stakeholders in planning and implementing the IDR programme offered a platform for collective learning and action. Impediments to the success of the programme need to be addressed by corrective actions at all levels for its future sustainability.

Key words: Verbal autopsy, infant death, rural, district, India.

Introduction The infant mortality rate (IMR) is universally Population Policy (2000) envisions reducing regarded as an important indicator of the IMR to less than 30 per 1000 live births,2 health and economic status of communities, echoing Millennium Development Goal 4 to and the effectiveness of maternal and child reduce under-five mortality by two thirds health services. According to the Sample between 1990 and 2015.3 Cause-of-death data Registration System (SRS) of India, IMR in are critical to formulating good public health Karnataka, a southern state of India, was 38 programmes; developing national, regional, per 1000 live births in 2010.1 India’s National and global policies; and implementing

aHealth Specialist, United Nations Children’s Fund, Field Office, Hyderabad, India. bSenior Consultant, Child Health, United Nations Children’s Fund, Bangalore, India. cMission Director, National Rural Health Mission, Karnataka, India. dProfessor, Department of Community Medicine, Sri Manakuala Vinayagar Medical College and Hospital, Pondicherry, India. eProfessor, Department of Community Medicine, Mahatma Gandhi Institute of Medical Sciences, Sewagram, India. Correspondence to Sanjeev Upadhyaya (e-mail: [email protected])

446 WHO South-East Asia Journal of Public Health 2012;1(4):446-456 Sanjeev Upadhyaya et al. Institutionalizing district level infant death review: an experience from southern India

and evaluating public health action.4,5 In programme actors at various levels. We hope developing countries, most deaths are neither that our findings may be useful to refine the attended by physicians nor certified medically, pilot programme, and for health-care systems so two thirds to three quarters of the world’s that wish to implement a similar programme population remain outside any systematic in a developing country context. mortality surveillance.6–10 Studies in Asia and Africa in the 1950s and 1960s used systematic Material and methods interviews with physicians to assess causes of death,11 a technique that the Narangwal project in India later labelled verbal autopsy (VA); the Setting method subsequently evolved, particularly The present study on the Infant Death Review during the 1970s when the World Health programme was carried out in Raichur and Organization (WHO) discussed lay reporting of Dakshina Kannada districts of Karnataka. health information by people with no medical These two districts were chosen in consultation background.11–14 Today, VA remains the best with the Director of the National Rural Health available approach to assess the cause of Mission, Karnataka, based on the performance death in communities in which most deaths of its health system. Raichur district, located occur at home. Separate questionnaires for 413 km north of the state capital Bangalore neonatal, child, maternal, and other adult with a population of 1 924 773, has poor deaths have been suggested.15 The need for an health indicators and a female literacy rate of infant death audit system was therefore crucial 49.6%. On the other hand, Dakshina Kannada, to map the rise or decline in infant deaths, to 347 km west of Bangalore, has a population identify common causes of death, as well as of 2 083 625 with better health indicators and for prioritization of actions, fund allocation and 84% female literacy. The crude birth rate of monitoring and evaluation. Karnataka was 19.2 per 1000 population with 34.1% home deliveries.1,18 This review was The United Nations Children’s Fund done during February and March 2012. (UNICEF), Hyderabad, India in partnership with the Department of Health and Family Welfare Services, Karnataka, has been implementing Method a pilot Infant Death Review (IDR) programme A sequential mixed-methods design was to strengthen the district health system review followed where quantitative data collection of infant deaths. So far, the research on verbal (secondary data) was followed by qualitative autopsy has focused on technical aspects data collection (in-depth interviews). such as the development of standardized To begin with, verbal autopsy forms and the tools, validated algorithms and computerized minutes of meetings and reports related to the algorithms.14,16 However, for VA methods to programme were reviewed. Reports from 566 extend their reach successfully from research infant deaths from both districts from January to routine application, feasibility demonstration to June 2011 were obtained from the district at local level and implementation research health system. Following quantitative analysis have been suggested to resolve emerging of the verbal autopsy records, 38 in-depth challenges as the programme develops.17 interviews (IDI) were conducted at district and Hence, we explored the processes leading to sub-district level with health-care providers the development of the IDR programme within who were actively involved in programme the health system, and the perspectives of key activities and willing to participate freely

WHO South-East Asia Journal of Public Health 2012;1(4):446-456 447 Institutionalizing district level infant death review: Sanjeev Upadhyaya et al. an experience from southern India

Table 1: Category of respondents and number of in-depth interviews conducted

Number of IDI

Respondents Dakshina Raichur Total Kannada

District level – District Health Officer, Reproductive and Child Health (RCH) Officer, District RCH Programme 3 5 8 Assistant, members of faculties of Community Medicine and Paediatrics of affiliated medical colleges

Peripheral health institutions – Medical officers, lady 10 9 19 health visitors, auxiliary nurse midwives

Village level – Accredited social health activists, 4 7 11 anganwadi workers

Total 17 21 38

(Table 1). Semi-structured guidelines were its administration, mortality classification, used to conduct the interviews, the purpose ascertainment of cause of death and forces of which was to explore the respondents’ ‘for’ and ‘against’ the IDR programme. experiences related to the IDR programme. Transcripts were reviewed and a list of codes Interviews were conducted at a date, time worked out. Manual content analysis was done and place convenient for the respondents. and text data were presented under each With their permission, interviews were audio coding category. Data collection and analysis recorded and transcribed verbatim. The was supervised by a public health expert interviewer (second author) was sensitized to who was trained in handling qualitative data. qualitative data collection. Findings of the analysis were reviewed by other authors to ensure their completeness Data analysis and trustworthiness. Statements ‘for’ and ‘against’ the programme were calculated for Quantitative data was entered using Epi-Info different levels of health-care providers. (version 3.4.3) software. The proportional mortality was calculated for neonatal and Informed consent was obtained from the post-neonatal periods. The inter-observer respondents of the IDI. Permission for this reliability between the Medical Officer and review was obtained from the Director of the the district Infant Death Review Committee National Rural Health Mission, Karnataka. for causes of death was assessed by kappa 16 statistics. This measures the agreement Results between two or more observers taking into account that they may agree or disagree Analysis of verbal autopsy purely by chance. For qualitative data records analysis, structural coding was done using Based on the prevalent infant mortality rates codes such as process, questionnaire design, applied by the state to the rural district of

448 WHO South-East Asia Journal of Public Health 2012;1(4):446-456 Sanjeev Upadhyaya et al. Institutionalizing district level infant death review: an experience from southern India

Raichur and the urban district of Dakshina place in government health facilities followed Kannada, we expected 1208 infant deaths by private health care facilities (32.8%), to be reported during January to June 2011. at home (18.9%) and in transit (12.1%) However, only 566 infant deaths were reported (Table 2). – 447 (79%) from Raichur and 119 (21%) Regarding the category of infant deaths from Dakshina Kannada – for which we in each district from January to June 2011, obtained verbal autopsy information. Dakshina Kannada reported 83 (69.7%) Of the 556 infant deaths, 316 (55.8%) neonatal and 36 (30.3%) post-neonatal were male and 250 (44.2%) were female; deaths, while Raichur reported 294 (65.8%) 451 (79.7%) belonged to parents who had neonatal and 153 (34.2%) post-neonatal below-poverty-line cards provided by the state deaths (Tables 2 and 3). government. Scheduled castes\scheduled As ascertained by the district IDR tribes (SC\ST) accounted for 239 (42.2%) Committee, the most common causes of of the deceased infants, more than half of infant deaths in Dakshina Kannada were whom occurred in Raichur. Noteworthy is the pneumonia (19.3%), birth asphyxia\injury breakdown between the two districts: 42% (13.4%), congenital malformations (12.6%), infant deaths in Raichur took place at home, low birth weight (9.2%) and prematurity the next highest proportions being in private (8.5%). Overall, the inter-code reliability health facilities (28.8%), government health between cause of death as ascertained by facilities (19.9%) and in transit (9.6%). In medical officers and the District IDR sub- Dakshina Kannada, 36.2% infant deaths took

Table 2: Sociodemographic characteristics of deceased infants in Raichur and Dakshina Kannada districts, January to June 2011

Socio demographic Raichur Dakshina Kannada Total characteristics (n=447)(%) (n=119)(%) (n=566)(%) Sex Male 241 (53.9) 75 (63.0) 316 (55.8) Female 206 (46.1) 44 (37.0) 250 (44.2) Socioeconomic status (colour of ration card) Above poverty line 72 (16.1) 43 (36.1) 115 (20.3) Below poverty line 375 (83.9) 76 (63.9) 451 (79.7) Caste General 109 (24.4) 31 (26.1) 140 (24.7) Scheduled caste\Scheduled tribe 207 (46.3) 32 (26.9) 239 (42.2) Other backward classes 70 (15.7) 21 (17.6) 91 (16.1) Others 61 (13.6) 35 (29.4) 96 (17.0)

Place of death (n=543; n1=427 and n2=116) Home 178 (41.7) 22 (18.9) 200 (36.8) Government health facility 85 (19.9) 42 (36.2) 127 (23.4) Private health facility 123 (28.8) 38 (32.8) 161 (29.7) Transit 41 (9.6) 14 (12.1) 55 (10.1)

WHO South-East Asia Journal of Public Health 2012;1(4):446-456 449 Institutionalizing district level infant death review: Sanjeev Upadhyaya et al. an experience from southern India

Table 3: Level of agreement on cause of death ascertained by Medical Officer and by Infant Death Review Committee, Dakshina Kannada District, Karnataka

Cause of death ascertained by Cause of death ascertained by Infant Overall medical officer (a) Death Review Committee (b) Kappa agree- Post- Post- Causes of death Neonatal Neonatal ment neonatal Total neonatal Total period period between period No.(%) period No. (%) No. (%) No. (%) a and b, No. (%) No. (%) (n=117) Birth asphyxia/ 11 (13.3) 0 (0) 11 (9.2) 16 (19.3) 0 (0) 16 (13.4) injury Congenital 12 (14.5) 4 (11.1) 16 (13.5) 11 (13.3) 4 (11.1) 15 (12.6) malformations Diarrhoea 0 (0) 1(2.8) 1 (0.8) 0 (0) 1 (2.8) 1 (0.8) Feeding problem 1 (1.2) 0 (0) 1 (0.8) 0 0 0 Infection of central nervous 0 (0) 5 (13.9) 5 (4.2) 0 (0) 4 (11.1) 4 (3.4) system Low birth weight 12 (14.5) 0 (0) 12 (10.1) 11 (13.3) 0 (0) 11 (9.2) Malnutrition Kappa 0 (0) 1 (2.8) 1 (0.8) 0 (0) 1 (2.8) 1 (0.8) related death agree- ment Neonatal sepsis 6 (7.2) 0 (0) 6 (5.0) 6 (7.3) 0 (0) 6 (5.1) – 0.809 Others* 7 (8.4) 5 (13.9) 12 (10.1) 5 (6.0) 5 (13.9) 10 (8.5) (95% Pneumonia 8 (9.6) 11 (30.5) 19 (16.1) 10 (12.0) 13 (36.0) 23 (19.3) CI; 0.73 – 0.88) Postnatal 7 (8.4) 5 (13.9) 12 (10.1) 7 (8.4) 4 (11.1) 11 (9.2) aspiration Prematurity 8 (9.7) 0 (0) 8 (6.7) 10 (12.0) 0 (0) 10 (8.4)

Respiratory 6 (7.2) 1 (2.8) 7 (5.9) 4 (4.8) 2 (5.6) 6 (5.1) distress syndrome

Sepsis 2 (2.4) 0 (0) 2 (1.6) 0 0 0 Sudden death 2 (2.4) 3 (8.3) 5 (4.2) 2 (2.4) 2 (5.6) 4 (3.4) Unknown 1 (1.2) 0 (0) 1 (0.8) 1 (1.2) 0 (0) 1 (0.8) Total 83 (69.7) 36 (30.3) 119 (100) 83 (69.7) 36 (30.3) 119 (100)

*Others include accidents, homicide, and severe form of kernicterus

committee was good (kappa agreement 0.81; Processes leading to the 95% confidence interval 0.73–0.88) (Table 2). development of the IDR In Raichur, the most common causes of programme death as ascertained by medical officers were (i) Preparatory process: In March 2010, birth asphyxia\injury (26.2%), pneumonia UNICEF Hyderabad organized a national (17.4%), and congenital malformations (11%) consultative workshop on Infant Death (Table 4).

450 WHO South-East Asia Journal of Public Health 2012;1(4):446-456 Sanjeev Upadhyaya et al. Institutionalizing district level infant death review: an experience from southern India

Table 4: Causes of death ascertained by Medical Officer in Raichur, Karnataka

Post-neonatal Neonatal period Total Cause of death period No. (%) No. (%) No.(%) Birth asphyxia/injury 116 (39.5) 1 (0.7) 117 (26.2) Congenital malformations 28 (9.5) 21 (13.7) 49 (11.0) Diarrhoea 1 (0.3) 6 (3.9) 7 (1.5) Feeding problem 4 (1.4) 3 (2.0) 7 (1.5) Hypothermia 5 (1.7) 0 (0) 5 (1.1) Infection of central nervous 0 (0) 12 (7.8) 12 (2.7) system Low birth weight 18 (6.1) 1 (0.7) 19 (4.3) Malnutrition related death 0 (0) 7 (4.6) 7 (1.6) Neonatal sepsis 39 (13.3) 2 (1.3) 41(9.2) Others 4 (1.4) 13 (8.5) 17 (3.8) Pneumonia 22 (7.5) 56(36.6) 78 (17.4) Postnatal aspiration 3 (1.0) 3 (2.0) 6 (1.3) Prematurity 14 (4.8) 1 (0.6) 15 (3.4) Respiratory distress syndrome 26 (8.8) 0 (0) 26 (5.8) Sepsis 0 0 0 Sudden death 3 (1.0) 8 (5.2) 11 (2.5) Unknown 11 (3.7) 19 (12.4) 30 (6.7) Total 294 (65.7) 153 (34.3) 447 (100)

*Others include accidents, homicide, and severe form of kernicterus

Review in Bangalore in collaboration with the (ii) Questionnaire design: a comprehensive Government of Karnataka. Experts at the verbal autopsy questionnaire for neonatal workshop discussed the VA questionnaire, and post-neonatal deaths was developed in mortality classification and developed guidelines consultation with experts as a checklist with for IDR implementation at district level, and in filters based on existing tools and literature. July 2010 the programme was initiated. This It was a simple colour-coded questionnaire involved government health-care providers; (pink for newborn deaths and yellow for programme managers from RCH and the post-neonatal deaths) in the local language department of Women and Child Development; Kannada. The questionnaire was pre-tested paediatrics and community medicine specialists on a sample of 12 deaths to incorporate from medical colleges; and representatives local words and terms. The VA tool consisted from private hospitals involved in the planning of 17 pages including 3 pages for writing process of the IDR programme. The workshop descriptive notes such as case summaries. It participants finalized the verbal autopsy tool covered comprehensive information on socio- and guidelines for the IDR programme. demographic characteristics, antenatal, natal and postnatal care and health-care seeking

WHO South-East Asia Journal of Public Health 2012;1(4):446-456 451 Institutionalizing district level infant death review: Sanjeev Upadhyaya et al. an experience from southern India

behaviour to capture the context in which the Officer of the Primary Health Centre (PHC) death occurred. on a short form called the First Information Report (FIR).In addition, they should send (iii) Capacity-building at various levels: a Short Message Service (SMS) to a unique Training started with a consultation of officials mobile number at district headquarters. In from the State Directorate and a few districts, one Raichur district block, the district health followed by rapid training courses at district system with support from UNICEF provided and sub district levels. At district level, an self-addressed and printed postal envelopes initial launch meeting was conducted to for reporting FIR. sensitize all stakeholders, including medical colleges, professional bodies, and other public (v) VA administration and cause of death service departments (e.g. Women and Child ascertainment: The Medical Officer was Development) in both districts. A separate asked to pay a home visit and interview the sensitization programme on modalities of close carer of the deceased infant within 15 reporting infant deaths at the community and days of its reporting to ascertain the cause facility level was conducted for the Indian of death. The completed VA forms were then Academy of Paediatrics chapter in Dakshina forwarded to the district Infant Death Review Kannada and for the Indian Medical Association subcommittee (IDRSC). This committee in Raichur. A one-day training programme was consisted of the District Health Officer, district organized for all medical officers separately RCH officer, paediatricians from regional in each district on administration of verbal medical colleges, public health experts, district autopsy and ascertainment of cause of programme managers from the department of death. Women and Child Development, and medical officers from peripheral health centres. Skill-building was carried out by reviewing all verbal autopsy questions in an interview In Dakshina Kannada, during monthly IDR in the local language, with illustrations on Committee meetings, all the submitted VA how to ascertain the cause of death based forms and clinical case records of the deceased on diagnostic criteria. Following this, a few infants available from parents were discussed. trained, skilled and motivated medical officers In addition to ascertaining the cause of death, were identified by the district RCH officer to IDRSC discussed socioeconomic factors conduct further training for auxiliary nurse and events related to pregnancy, delivery, midwives, lady health visitors and other postnatal period, health-care seeking, gaps in paramedical staff at taluk level. Medical the health system and health-care services. In officers of primary health centres sensitized Raichur, IDRSC reviewed only selected cases anganwadi workers (AWW) and accredited of infant deaths from community and hospital social health activists (ASHA) under their settings. The cases were selected on the basis service area. In order to create awareness of completeness of records, geographical of the IDR programme, Gram-panchayat location and cause of death. IDRSC provided (local self-government) members, leaders of feedback to peripheral health institutions women’s self-help groups, and representatives regarding common causes of infant deaths from local nongovernmental organizations and implementation of strategies tailored to were sensitized in specific training sessions. local needs. The proceedings of IDRSC were submitted to the district IDR Committee to (iv) Information on infant deaths: Village refine needs-based strategies to prevent infant workers such as AWW and ASHA were asked deaths. to report details of infant deaths to the Medical

452 WHO South-East Asia Journal of Public Health 2012;1(4):446-456 Sanjeev Upadhyaya et al. Institutionalizing district level infant death review: an experience from southern India

(vi) Positive and negative effects of the of a medical officer in the village to conduct Infant Death Review system and options verbal autopsy confers a sense of importance for improvement to issues related to infant deaths (Table 5). Positive elements of the IDR system for the district health authorities and medical college Discussion faculties were their new partnership and the Decision-makers and programme managers involvement of regional medical colleges in need simple data that can be linked to planning, implementation and training. This improving the provision of care. These created a rapport with the local community potential users of verbal autopsy methods and improved infant death reporting. Forces have different perspectives from researchers, ‘against’ the programme were an initial lack tending to prefer simple VA instruments for of motivation at district and peripheral level their feasibility and programme relevance over to the new initiative, incomplete information technical performance. Our study explored in verbal autopsy reports, late reporting, the different perspectives of district health lengthy form, and overburdened staff. Options authorities, staff of the Primary Health Centre suggested for future sustainability were to and village-level staff as key actors in the IDR make the VA reporting form more concise, to programme. This contextual knowledge is train district level staff and medical officers useful for health-care managers and planners and to improve health-care facilities. for ensuring programme sustainability. In Health workers of peripheral health particular, we noted that health system institutions reported that the identification of managers sought concrete and programme- gaps in the health-care system and financial relevant data, implying an interest in the incentives for AWW and ASHA to report deaths socioeconomic determinants of mortality. had contributed to infant death reporting. They The Survey of Causes of Death by the also saw an opportunity to educate family Registrar General of India has been merged members on the prevention of infant deaths. with the Sample Registration System, resulting On the negative side, these health workers in a modified verbal autopsy method with felt overburdened with their existing workload, physician review. The current SRS sample size and suffered from poor cooperation from does not allow sub district level estimation caregivers and a lack of transport facilities. of IMR, which varies across districts. Thus, a Options to redress this situation included reduction in the average IMR in a State does financial incentives for medical officers to not provide a complete picture of mortality complete verbal autopsies, improved logistics, decline, necessitating identification of high and forming sub district committees where mortality-prone areas and planning innovative infant deaths are high. strategies for IMR reduction in these areas.1 Anganwadi and ASHA workers noted Hence the need for sub-district ‘infant death that, as a result of the IDR programme, audit systems’ to gain accurate and reliable they had more information on the causes estimates for planning at the district and sub- of infant deaths, and had feedback. This in district level. turn had improved antenatal care. However, Our study proposes a feasible approach they have to face poor cooperation from to strengthen the existing health-care caregivers, including occasional anger. These system at district level. We found that village workers considered that the arrival the involvement of stakeholders – the

WHO South-East Asia Journal of Public Health 2012;1(4):446-456 453 Institutionalizing district level infant death review: Sanjeev Upadhyaya et al. an experience from southern India

Table 5: Forces ‘for’ and ‘against’ the IDR programme at various levels

Options for Stakeholder Perceived advantage Perceived impediment sustainability

District level Public–private Incomplete information in Make reporting form authorities partnership(private verbal autopsy forms concise and medical medical colleges) college Involvement of Late reporting and poor Provide repeated faculties pediatricians feedback training and feedback to medical officers Involvement of Lack of motivation partners in planning, Improve health-care implementation and Lack of interest in facilities training peripheral staff Train district level staff Closer rapport with the Lengthy VA; staff do not community always understand the technical content Improved reporting and ensuing actions Overburdened staff

Form in local language

Primary health Gaps in health service Overburdened with existing Financial incentives care staff identified and rectified national health programme for medical officers to complete VA Improved reporting of Coordination of many infant deaths meetings Regular supply of forms

Financial incentives for Non-availability of, and Taluk (sub-district) village level anganwadi or incomplete information level committee may ASHA workers to report from caregivers, requiring be formed if number of deaths more than two visits deaths reported is high

Opportunity to educate Transport problems to family members for access remote villages prevention of infant deaths

Village-level Improved quality of Poor cooperation from Medical officers visiting workers antenatal care caregivers (as they believe the family of the village workers are paid for deceased would confer Move from just reporting VA) the seriousness of the to finding out cause of issue; including for infant death and events Anger faced from family neighbours leading to it members for deficiency in health care (if death happens at health-care delivery centre)

Forms, or some terms therein, in English

Cost of using mobile phone to send SMS to announce infant death

454 WHO South-East Asia Journal of Public Health 2012;1(4):446-456 Sanjeev Upadhyaya et al. Institutionalizing district level infant death review: an experience from southern India

Department of Health, Department of Women Family Welfare Services, Karnataka, for the and Child Development, local medical colleges, opportunity to pilot this study in two districts district hospital authority, peripheral health in the state. We also thank the District Health institutions, local representatives from WHO and Family Welfare Officers and District and UNICEF – in planning and implementing Reproductive and Child Health Officers in the IDR programme provided a platform for Raichur and Dakshina Kannada for allowing collective learning and action. The causes of this study to be conducted in their districts. deaths derived from verbal autopsies may not Finally, we express our appreciation to all staff be as reliable as hospital-based certification, and expert members of the IDR Committee but information on risk factors and the health who volunteered information, as well as to care sought prior to death made health-care all individuals whose contributions made the providers more aware of the causes of infant work possible. deaths. The potential impediments identified to the success of the IDR programme need to be addressed by corrective actions at the relevant References level by policy-makers and programmers for 1. Sample registration system of India, Sample its sustainability. Registration Bulletin. 2011 Dec; 46(1) - http://pib. nic.in/archieve/others/2012/feb/d2012020102.pdf - Conclusion accessed 16 April 2013. 2. Gvernment of India, National Commission on Involving the different stakeholders such as Population. National population policy 2000 - government health departments and private URL:http://populationcommission.nic.in/npp_obj. health-care providers in the IDR programme htm - accessed 16 April 2013. creates synergy and is essential for its 3. Tracking the Millennium development goals in Asia sustainability. In addition to ascertaining the and the Pacific. 2009 - www.mdgasiapacific.org/.../ cause of infant deaths, the IDR Committee Regional_MDGs_report_2_chapter1.pdf - accessed discusses social, economic, behavioural 16 April 2013. and health system issues that may have 4. Losos JZ. Routine and sentinel surveillance methods. contributed to the death. As a result of the East Mediterr Health J. 1996; 2(1): 45-60

IDR programme, respondents perceived an 5. Byass P. Who needs cause-of-death data. PLoS Med. improvement in infant death reporting at the 2007; 4(11): e333.

district level, a closer rapport with the local 6. Beaglehole R, Bonita R. Challenges for public health community, and the welcome development in the global context—prevention and surveillance. of a feedback system for corrective actions. Scand J Public Health. 2001; 29(2): 81-83.

This has led to improved quality of care during 7. Byass P. Person, place and time—but who, where, and pregnancy as perceived by health service when? Scand J Public Health. 2001; 29(2): 84-86.

providers. 8. Cleland J. Demographic data collection in less developed countries 1946–1996. Popul Stud (Camb). Acknowledgements 1996; 50(3): 433-450. 9. Begg S, Rao C, Lopez AD. Design options for sample- The authors thank Dr Arun Aggarwal, Professor based mortality surveillance. Int J Epidemiol. 2005; of Community Medicine, Postgraduate Institute 34(5): 1080-1087. of Medical Education and Research, Chandigarh, 10. Hill K. Making deaths count. Bull World Health Organ. for his critical review and comments. We are 2006; 84(3): 162. grateful to the Directorate of Health and

WHO South-East Asia Journal of Public Health 2012;1(4):446-456 455 Institutionalizing district level infant death review: Sanjeev Upadhyaya et al. an experience from southern India

11. Garenne M, Fauveau V. Potential and limits of verbal 15. AbouZahr C. Verbal autopsy: Who needs it? autopsies. Bull World Health Organ. 2006; 84(3): Population Health Metrics. 2011; 9: 19 - http:// 164. www.pophealthmetrics.com/content/9/1/19 - 16 April 2013. 12. World Health Organisation. Lay reporting of health information. Geneva: WHO, 1978. 16. Bang AT, Bang RA and SEARCH team. Diagnosis of causes of childhood deaths in developing countries by 13. Garenne M, Fontaine O. Assessing probable causes verbal autopsy: suggested criteria. Bulletin of World of deaths using a standardized questionnaire: a Health organization. 1992; 70(4): 499-507. study in rural Senegal. In: Vallin J, D’Souza S, Palloni A, editors. Measurement and analysis of 17. Government of India, Ministry of Health and Family mortality. Proceedings of the International Union for Welfare. District level household and facility survey. the Scientific Study of Population seminar, Sienna, Fact sheet Karnataka. Mumbai: International Institute Italy, July 7–10, 1986. Oxford, Clarendon Press, for Population Sciences - http://www.rchiips.org/pdf/ 1990. pp. 123-142. rch3/state/Karnataka.pdf - 16 April 2013.

14. Soleman N, Chandramohan D, Shibuya K. Verbal 18. Fleiss JL. Statistical methods for rates and proportions. autopsy: Current practices and challenges. Bull World New York: John Wiley, 1981. pp. 38-46. Health Organ. 2006; 84 (3): 239-245.

456 WHO South-East Asia Journal of Public Health 2012;1(4):446-456 Policy and practice

Health systems responsiveness and its correlates: evidence from family planning service provision in Sri Lanka W L S P Pereraa, Lillian Mwanrib, Rohini de A Seneviratnec, Thushara Fernandod

Background: The World Health Organization has described health systems responsiveness (HSR) as a multi-domain concept encompassing eight non-medical expectations of health-care service seekers. HSR is a valuable measure of health systems performance, providing policy-makers and service providers much information to improve services. This paper presents findings of a cross- sectional survey conducted to assess HSR and its correlates through family planning (FP) services in Colombo district, Sri Lanka.

Methods: A Health Systems Responsiveness Assessment Questionnaire, developed and validated in Sri Lanka, was used. Trained interviewers administered the questionnaire in 38 FP clinics randomly selected to sample 1520 clients.

Results: The rating of responsiveness as ‘good’ for six domains ranged from 88% (n=1338) to 72% (n=1094). The overall HSR was rated to be ‘good’ by 83.4% (1268). Ethnicity being majority Sinhalese, persons who were currently employed and those using oral contraceptive pills (OCP) or condoms were negatively associated with rating of HSR. Positive associations with the HSR assessment were a family income of less than Rs 40 000 (US$ 303) per month, satisfaction with current FP method, use of only one method within the past year, use of only one FP clinic within the past year, health service provider being a medical officer, intention to use the FP clinic services in future, and satisfaction with overall services of the clinic.

Conclusions: Though overall HSR was rated by the majority as ‘good’, some aspects need more attention in delivering FP services.

Key words: health systems responsiveness, family planning, Sri Lanka.

Introduction

Health systems responsiveness (WHO), identified three main goals to improve The 2000 World health report (WHR), the performance of health-care systems.1 published by the World Health Organization The degree to which the legitimate non-

aActing Consultant Community Physician, Ministry of Health, Sri Lanka. bPublic Health Physician (FAFPHM), Course Coordinator, Master of Health and International Development, Discipline of Public Health, School of Medicine, Flinders University, Adelaide, Australia. cHead/ Department of Community Medicine, Faculty of Medicine, University of Colombo, Sri Lanka. dCountry Representative, World Health Organization, Bangladesh. Correspondence to W L S P Perara (e-mail: [email protected])

WHO South-East Asia Journal of Public Health 2012;1(4):457-466 457 Health systems responsiveness and its correlates: evidence W L S P Perera et al. from family planning service provision in Sri Lanka

medical expectations of people were met was WHO assessed HSR in 71 Member States one of these goals, named health systems’ in 2002–2004 during the World Health Survey responsiveness (HSR).1,2 (WHS).4 Inpatient and outpatient ‘modules’ were used for the assessment, which covered Assessment of HSR the entire health system of the country and included government, private and traditional The need to assess and improve national medicine systems. Results were presented HSR has been documented. The assessment, as a percentage of respondents, rating each essentially, has to be made by the service item of the modules as moderate, bad or very seeker who is in the best position to state bad. Table 1 describes the published data on how non-medical expectations were met.1,2 patient services for Sri Lanka. Assessments based on individual perspectives are influenced by sociodemographic factors of While assessing the responsiveness of the the respondents as well as the sociocultural entire health system is important, assessing milieu. Nevertheless, this approach can provide a specific service or institution would have valuable information on the performance much practical value.2 In line with this of health-care systems, especially if the recommendation, a tool has been developed assessment tool is tailored to the specific and validated in the Sri Lankan setting to sociocultural context of the country.2,3 assess HSR of family planning services.5,6

Table 1: Health system responsiveness in Sri Lanka

Outpatient care Inpatient care assessment assessment Domain/item of responsiveness % users who rated % users who rated moderate, bad or moderate, bad or very bad very bad Travel time to health centre 29.8 37.7 Prompt attention Waiting time 38.0 38.4 Talked respectfully 13.1 20.0 Dignity Privacy 23.2 17.5 Clear explanation 17.4 26.0 Communication Time for questions 29.2 38.0 Treatment information 33.1 38.7 Autonomy Involvement 29.6 38.7 Talked privately 24.3 36.7 Confidentiality Confidentiality of records 23.3 34.0 Choice 30.9 42.2 Cleanliness 30.6 39.8 Basic amenities Space 40.7 48.4 Family visit NA 36.6 Social support External contact NA 44.0 Source: WHO, 2004/2005.

458 WHO South-East Asia Journal of Public Health 2012;1(4):457-466 W L S P Perera et al. Health systems responsiveness and its correlates: evidence from family planning service provision in Sri Lanka

Correlates of HSR districts.9 The Colombo district houses Many factors – related to both the health approximately 20% of this population and sector as well as the sociocultural context – has a network of public as well as private determine people’s assessment of HSR. health-care institutions providing FP services. These correlates play an important role in The public health sector runs 64 registered FP the assessment and could provide insight clinics in 12 Medical Officer of Health (MOH) 7,8 for policy-makers and service providers to divisions. improve service provision. To our knowledge, A cross-sectional study was conducted on these factors have not been scientifically clients seeking services from FP clinics of MOH appraised and published. divisions in Colombo district. Sample size was calculated using the formula for estimating Family planning services prevalence for descriptive studies.10 Prevalence in Sri Lanka of ‘good’ HSR was taken as 60% as it was the Family planning (FP) services were introduced lowest estimation made in the country to 4,10 to Sri Lanka in 1953 and were well accepted.7 ensure maximum sampling size. Estimated Today, they are a key component of the sample size was 753 for a precision of 3.5%. maternal and child health (MCH) package Cluster sampling was used: clients attending provided by the Government with the objective one clinic session comprised a cluster. Cluster 13 of improving the health and quality of life of size was 40 patients and the sample size families.7–8 The Family Health Bureau of the was multiplied by two to account for design 11–12 Ministry of Health coordinates provision of effect. Thirty-eight clinics were randomly services within the government sector and selected from the list of registered FP clinics registers FP clinics. Five methods of FP are and within the clinic, systematic sampling 13 provided through these registered clinics. was used to identify study participants. The calculated sample size was 1506 and a total Though applicable to every facet of health- of 1520 was taken as final. care delivery, HSR is especially relevant for services that are sought repeatedly over many We developed and validated the Health years, e.g. long-term use of FP. Family planning Systems Responsiveness Assessment is a sensitive topic in Sri Lanka where non- Questionnaire (HESRAQ) in Sri Lanka to assess 5,6 medical expectations of clients often determine the HSR of FP services. Since correlates their willingness to use services, even if they of HSR had not been identified earlier, focus are widely available. Also, the satisfaction group discussions and in-depth interviews of clients plays a critical role in ensuring the were conducted with clients and non-users proper use of FP methods, and minimizing of FP, service providers and policy-makers to method failure. This paper describes the identify possible correlates. Twenty-five factors findings of a cross-sectional survey conducted were identified and assessed in the cross- to assess the responsiveness of FP services of sectional survey. The tool was administered the public health sector through examination by trained research assistants who were pre- of HSR correlates. intern medical officers to persons as they left the FP clinics. Methodology The results of the rating of HSR were categorized into three groups. Group 1 Sri Lanka, with an estimated population of combined ‘very good’ and ‘good’ and indicated 20.86 million, is divided into 25 administrative

WHO South-East Asia Journal of Public Health 2012;1(4):457-466 459 Health systems responsiveness and its correlates: evidence W L S P Perera et al. from family planning service provision in Sri Lanka

good HSR; Group 2 contained the ‘moderate’ rating. Logistic regression analysis of a rating responses, indicating neutrality; and Group 3 being ‘good’ is summarized in Table 4. consolidated ‘bad’ and ‘very bad’, signifying Domains that had negative HSR outcomes that the HSR was bad. Correlates were were Sinhalese ethnicity (OR=0.189, CI identified using logistic regression analysis. 0.062–0.579), being employed (OR=0.247, First, bivariate analysis was performed using CI 0.104–0.587) and using OCP/condoms chi square test followed by logistic regression as contraception (OR=0.09, CI 0.02-0.31). with significant variables. SPSS version 16 On the other hand, factors associated with was used to determine the odds ratios (OR) positive HSR were a family income of less than and 95% confidence intervals (95% CI) of Rs 40 000 (US$ 303) per month (OR=19.31, the correlates for rating the overall HSR as CI 4.70–79.27), satisfaction with current FP ‘good’ or ‘bad’. method (OR=10.68, CI 4.80–23.74), use of only one method in the past year (OR=6.69, Results CI 2.74–16.33), use of only one FP clinic in Response rate was 100%. All respondents the past year (OR=9.91, CI 3.87–25.36), were married females and came from health service provider being a medical officer different socioeconomic backgrounds. (OR=19.77, CI 3.59–108.88), intention to use The sociodemographic characteristics of the clinic for FP services in future (OR=14.24, respondents and their FP use are presented CI 4.13–49.08) and satisfaction with overall in Table 2. services of the clinic (OR=69.07, CI 20.31– 234.87). The findings of the assessment carried out using HESRAQ are given in Table 3. Discussion Although the vast majority (83.4%) of Our study showed that the majority of participants stated that the overall HSR of respondents were happy with the HSR in FP services was good, 3.8% were unhappy relation to FP services provided by the and a further 12.8% remained neutral in Government. This is consistent with findings their assessment. The highest rated domain of the only other local assessment on HSR in was ‘being treated with dignity’ (88%). the WHS.4 Since the WHS only published the Regarding the confidentiality maintained on percentage of respondents rating each item medical information, and the ease of access as ‘moderate/bad/very bad’, and only on a to FP services, 84% and 83% of respondents five-point Likert-like scale, the percentage were happy, respectively. On the other hand, rating of each item as ‘good/very good’ would almost one fifth expressed dissatisfaction or be reasonably interpreted as the balance remained neutral about the state of the clinic percentage of respondents. This percentage environment, with another 23.9% stating would reasonably compare with the percentage the same regarding adequate communication rating items as ‘very good/good’ in the present during medical encounters. ‘Client choice’ was study. Other studies conducted in Sri Lanka on rated lowest with 28% of respondents stating patient satisfaction and quality of health care they were either unhappy or neutral on the – two concepts with similarities to HSR –also adequacy of choice given to them on the clinic, reveal a high rating of non-medical aspects of provider and FP method. health-care delivery by the Government.14–18 Though 25 correlates were identified, only Though the overall rating was high, there was 10 were found to be associated with an HSR much variation in different domains of HSR.

460 WHO South-East Asia Journal of Public Health 2012;1(4):457-466 W L S P Perera et al. Health systems responsiveness and its correlates: evidence from family planning service provision in Sri Lanka

Table 2: Sociodemographic characteristics and use of modern family planning methods among respondents

Sociodemographic characteristic No. % Age < 25 334 22.0 25–34 984 64.7 ≥ 35 202 13.3 Marital status Currently married 1520 100.0 Ethnicity Sinhala 1240 81.6 Sri Lankan Tamil 128 8.4 Indian Tamil 14 0.9 Muslim 138 9.1 Religion Buddhist 1152 75.8 Hindu 120 7.9 Roman Catholic/Christian 110 7.2 Islam 138 9.1 Level of education No schooling 0 0.0 Grade 1–5 186 12.2 Grade 6–10 806 53.0 Passed GCE O/L 278 18.3 Grade 12 – 13 66 4.4 Passed GCE A/L* 148 9.7 Above GCE A/L qualifications 36 2.4 Current employment status Currently employed 378 24.9 Currently unemployed 1142 75.1 Method of family planning Oral contraceptive pill 74 4.9 Condoms 10 0.7 Depot-medroxyprogesterone acetate (DMPA) 1168 76.8 Implant 56 3.7 Intrauterine contraceptive device (IUD) 212 13.9

*GCE O/L – General Certificate of Education Ordinary Level. **GCE A/L – General Certificate of Education Advanced Level.

WHO South-East Asia Journal of Public Health 2012;1(4):457-466 461 Health systems responsiveness and its correlates: evidence W L S P Perera et al. from family planning service provision in Sri Lanka

Table 3: Health systems responsiveness assessment of family planning services

Rating Total

Domain Good Moderate Bad No. (%) No. (%) No. (%) No. (%)

Being treated with dignity 1338 (88.0) 125 (8.2) 57 (3.8) 1520 (100)

Confidentiality 1277 (84.0) 185 (12.2) 58 (3.8) 1520 (100)

Ease of access to services 1262 (83.0) 200 (13.2) 58 (3.8) 1520 (100)

Clinic environment 1218 (80.1) 244 (16.1) 58 (3.8) 1520 (100)

Communication 1157 (76.1) 301 (19.8) 62 (4.1) 1520 (100)

Client choice 1094 (72.0) 356 (23.4) 70 (4.6) 1520 (100)

Overall HSR 1268 (83.4) 194 (12.8) 58 (3.8) 1520 (100)

Table 4: Correlates of ‘good’ health systems responsiveness

95% CI for Exp Correlate (β)

Exp very good/good Sig. Lower Upper (β)

1 Sinhalese in ethnicity 0.004 0.189 0.062 0.579

2 Currently employed 0.002 0.247 0.104 0.587

Family income less than Rs 40 000 (US$ 303)/ 3 0.000 19.31 4.70 79.27 month

4 Current method of FP: OCP/condoms 0.000 0.09 0.02 0.31

5 Satisfied with current method of FP 0.000 10.68 4.80 23.74

6 Used only one FP method in past year 0.000 6.69 2.74 16.33

7 Used FP services from only one clinic in past year 0.000 9.91 3.87 25.36

8 Service provider is a medical officer 0.001 19.77 3.59 108.88

9 Intent to use the clinic for FP services in the future 0.000 14.24 4.13 49.08

10 Satisfied with overall services of the clinic 0.000 69.07 20.31 234.87

Constant 0.000

Exp (β) – Odds ratio

462 WHO South-East Asia Journal of Public Health 2012;1(4):457-466 W L S P Perera et al. Health systems responsiveness and its correlates: evidence from family planning service provision in Sri Lanka

Respondents expressed satisfaction The majority of respondents expressed on being treated with dignity in clinics. satisfaction with the environment of FP clinics. The domain ‘dignity’ of the WHS module The WHS assessed OPD facilities under the assessing outpatient department (OPD) domain ‘basic amenities’, where the items settings is comparable with the domain in were rated very good and good by 69.4% the present study.4 The two items comprising and 59.3%. The higher levels of rating in our the domain in WHS were rated as very good/ study could be explained because the setting good by 86.9% and 76.8% of respondents was an FP clinic at which a limited number of respectively, while the rating in the present mothers attend, unlike OPD settings which study was 88%.4 There is much respect for are generally overcrowded, especially in dignity of mothers attending MCH clinics in the larger hospitals where most people go for local culture compared with general medical outpatient services. Therefore, it is evident services.14 Also, the closeness of clients to that the FP clinic setup – the general MCH service providers, who are mainly public health setup in the public health sector – offers a midwives, leads to much respect between the better environment than the general OPD client and service provider in the field.5 facilities.5

Confidentiality was the second highest Communication in the WHS was assessed rated domain. The ratings in the WHS were on two items: clear explanations and time lower, i.e. 75.7% against 76.7% in the present for questions, rating very good or good at study.4 Since the WHS was a general OPD 82.6% and 70.8% respectively. Only 76.1% assessment, the higher rating in the present of respondents in our study said that their study could be explained by the cultural expectations had been met with regard to respect for mothers seeking FP services in communication. The reason for this may be relation to their confidentiality, in addition a lack of explanation on the more technical to there being very little medical information aspects of FP methods and even the side collected at the FP clinic as opposed to OPD effects of FP methods.6 On the contrary, the settings.5,19 time given for questions was rated lower in the WHS, given the general overcrowding and Ease of access to services was rated by thus waiting time in OPD settings for patients 83.0% as being good. The domain ‘prompt to receive clarifications. attention’ was comparable with this domain in our study.4–5 Both elements were rated as The client choice was rated lowest, i.e. very good and good in the WHS (70.2% and either bad or average, at 28%. The WHS 62%, respectively). The WHS was an overall assessed the aspect of choice in two domains, assessment on OPD services where the time namely autonomy and choice of provider.4–5 The travelling to larger but far-away hospitals choice of provider in the WHS study met the and the longer waiting time in overcrowded expectations of 69.1% of people, which largely hospital settings is reflected in the results. matched the finding of our study. Autonomy, or FP services, which are provided at smaller the involvement of patients in making decisions institutions and outreach clinics, offer services in OPD settings was rated lower in the WHS, closer to home, and reduce the waiting time, assessed on providing treatment information given that mothers are generally given a and involvement in deciding treatment. Only specific time to attend the clinic.5 66.9% and 70.4% rated these aspects very

WHO South-East Asia Journal of Public Health 2012;1(4):457-466 463 Health systems responsiveness and its correlates: evidence W L S P Perera et al. from family planning service provision in Sri Lanka

good/good respectively. Involving the patient Monthly family income was analysed on in decision-making was similar in the present two levels, with a cut-off based on the median study, although the provision of treatment income. A monthly family income of less than information was lower in the WHS. This could Rs 40 000 (US$ 303) had a positive effect on reflect a lack of attention to informing patients the HSR rating. Respondents with a higher adequately on their medical condition in the family income may have higher expectations general outpatient settings.4 and hence tend to be ‘harsher raters’.22 Though respondents with high income have Correlates the option of obtaining services from the private sector, there was more confidence in Regarding the analysis of ethnicity, the the government public health services despite Sinhalese (the majority in Sri Lanka) rated the inconvenience.5 the HSR lower. However, qualitative research on HSR by the Principal Investigator in Sri Contraceptive methods were categorized Lanka showed that non-Sinhalese respondents into two groups; methods that can be obtained had lower levels of expectations.5 Literature during domiciliary services (OCP/condoms) presumes that minority ethnic groups may and methods that can only be obtained from have lower expectations of government an FP clinic (DMPA/implant/IUD). Use of OCP services when the majority of the workforce and condoms showed a negative association comprises the ethnic majority.20-21 Some with rating of HSR. The present study revealed literature suggests that minorities could be that, though the clinics have to provide all five discriminated against and marginalized in methods of FP, users generally seek DMPA, health-care delivery and that they could rate implant and IUD services. Service providers HSR poorly; this was not supported in the expect OCP and condom users to use public present study.3,22 health midwives and are not well received at clinics, as providing these FP methods is Being currently employed showed a sometimes treated as adding to the already negative association with the rating of HSR. overburdened clinics.5 This could be explained by the fact that the employed have less time to visit clinics, Satisfaction with the current method was which are generally open only during working positively associated with the HSR rating. As hours.5 This observation is supported in the already shown, this matches the satisfaction literature, which states that respondents in the expressed by this group with all aspects of labour force might have higher non-medical care in clinics.5 The main objective of visiting expectations, especially regarding efficient the clinic was to obtain a method of choice services.22–24 Literature on patient satisfaction that would satisfy the respondent. Therefore, suggests that currently employed respondents obtaining a satisfactory method ‘free at the may be more educated with a higher monthly point of delivery’ in the local context was income and a higher level of expectations.25–27 adequate, and shortcomings in other aspects In our study, the majority of the currently were secondary issues. The literature supports employed had a lower family income level the findings that, irrespective of the level of than those who were employed on a daily expectation, the service seeker’s satisfaction basis, and many had to lose a day’s work to with medication, etc. could lead to a higher seek services. rating.3,24

464 WHO South-East Asia Journal of Public Health 2012;1(4):457-466 W L S P Perera et al. Health systems responsiveness and its correlates: evidence from family planning service provision in Sri Lanka

Having used only one method within The strongest positive association was the previous year was positively associated being satisfied with the overall clinic services.5 with HSR rating. The literature states that The literature emphasizes that patients’ experimenters and frequent changers are assessment of the health-care delivery system the hardest to satisfy during any health-care is mainly dependent on their satisfaction with provision, and that the personality of the the medical care provided.24–25 client plays a role in the satisfaction with the care received.25 This was not taken into consideration in our study. Some respondents References had to change method due to the unavailability 1. World Health Organization. World health report, of another at some point.5 These clients were health systems: improving performance. Geneva: unsatisfied with the general medical services WHO, 2000. of the clinic due to such shortages. 2. World Health Organization. Report on WHO meetings of experts: responsiveness concepts and measurement. Obtaining services from only one clinic Geneva: WHO, 2000. over the previous year was associated with a 3. Murray CJL, Evans DB. Health systems performance positive rating of HSR. The literature suggests assessment: debates, methods and empiricism. 20,28 that doctor shoppers are harder to please. Geneva: World Health Organization, 2001. Local literature also confirms that clients who 4. World Health Organization. World health survey. visit several places are unhappier with their Geneva: WHO - http://www.who.int/healthinfo/ medical and non-medical care.5 In addition, survey/whslka-srilanka.pdf - accessed 16 April regular service seekers are generally from the 2013. surrounding area and know the public health 5. Perera WLSP. Health systems responsiveness of midwife and other health providers, making family plannig services in colombo district. Colombo: the clinic environment more friendly.5 Postgraduate Institute of Medicine, 2011. 6. Perera WLSP, Seneviratne RDA, Fernando HTEI. Attendance by a medical officer showed Development and validation of a tool to assess health a positive association with rating of HSR and systems responsiveness. South East Asia Journal of was an expectation of many when seeking Public Health Bangladesh. 2011; 1(1): 21- 27. any form of health service. In addition, female 7. Sri Lanka, Ministry of Health. Family health report. clients often had more confidence in a male Colombo: Family Health Bureau, 2004. medical officers, who were thought to be more 8. Sri Lanka, Ministry of Health. Family health report. skilled. In relation to invasive procedures such Colombo: Family Health Bureau, 2008. as implants and IUDs, medical officers were 9. Sri Lanka, Ministry of Health Care and Nutrition. believed to be more trained compared with Demographic and health survey 2006/7. Colombo: 5 even the most experienced nursing officer. Department of Census and Statistics, 2008. Intention to use the clinic for FP services 10. Lwanga SK, Lemeshow S. Sample size determination in the future had a positive association with in health studies: a practical manual. 2nd ed. Geneva: HSR. In addition to noting satisfaction, such World Health Organization, 1991. respondents did not wish to make a lower 11. Abramson JH, Abramson ZH. Survey methods in assessment of HSR for fear of discrimination community medicine. London: Churchill Livingston, 1999. in case the health-care providers became aware of their response, despite assurance of 12. Arambepola CK. Abdominal obesity and its association confidentiality.5 with selected risk factors of coronary heart disease in an adult population in the district of Colombo. Colombo: Post Graduate Institute of Medicine, 2004.

WHO South-East Asia Journal of Public Health 2012;1(4):457-466 465 Health systems responsiveness and its correlates: evidence W L S P Perera et al. from family planning service provision in Sri Lanka

13. Hully SB, Cummings SR, Browner WS. Designing 22. De Silva A, Valentine N. Measuring responsiveness: clinical research. 3rd edn. London: Lippincot Williams Results of a key informants survey in 35 countries. and Wilkins, 2001. Geneva: World Health Organization, 2002.

14. De Silva UG. An assessment of quality of care 23. Strengthening health systems to improve health provided by the FP clinics and field services in DPDHS outcomes, everybody’s business. Geneva: World area Colombo. Colombo: Post Graduate Institute of Health Organization, 2007. Medicine, 2004. 24. Valentine N, Darby C, Bonsel GJ. Which aspects 15. Tennakoon S. Patient satisfaction survey in a Sri of non-clinical quality of care are most important? Lankan hospital. Colombo: Post Graduate Institute Results from WHO’s general population surveys of of Medicine, 1990. “health systems responsiveness” in 41 countries. Social Sciences and Medicine. 2008; 66(9): 1939- 16. Jayatissa KLR. An analysis of utilization, quality and 50. costs of maternal and child health services. Colombo: Post Graduate Institute of Medicine, 1995. 25. Mendonca KMPP, Guerra RO. Development and validation of an instrument for measuring patient 17. Jayasekera HDMH. An assessment of quality of life satisfaction with physical therapy. Rev Bras Fisioter. and satisfaction with care in patients diagnosed with 2007; 11: 369-76. some common cancers. Colombo: Post Graduate Institute of Medicine, 2006. 26. Daoud-Marakchi M, Fendri-Elouze S, Ill CH. Development of a Tunisian measurement scale for 18. De Silva P. Functional disability, health related quality patient satisfaction: Study case in Tunisian private of life and health care cost profile of young elderly in clinics. World Academy of Science, Engineering and urban and rural areas of Kalutara district. Colombo: Technology. 2008; 45: 208-219. Post Graduate Institute of Medicine, 2010. 27. Loblaw DA, Bezjak A, Bunston T. Development 19. Darby C, Valentine N, Murray C. Strategy on and testing of a visit-specific patient satisfaction measuring responsiveness. Geneva: World Health questionnaire: The Princess Margaret Hospital Organization, 2003. satisfaction with doctor questionnaire’. Journal of 20. Letkovicova H. Health systems’ responsiveness: Clinical Oncology. 1999; 17(6): 29 -40. analytical guidelines for surveys in multi-country 28. Valentine NB, Bonsel GJ, Murray CJ. Measuring survey study. Geneva: World Health Organization, quality of health care from the user’s perspective 2005. in 41 countries: psychometric properties of WHO’s 21. Peltzer K. Patient experiences and health systems questions on health systems responsiveness. responsiveness in South Africa. Biomed Central. International Journal in Quality Healthcare. 2007; 2009; 9 (2): 117-124. 16(7): 1107-25.

466 WHO South-East Asia Journal of Public Health 2012;1(4):457-466 Policy and practice

Challenges faced by skilled birth attendants in providing antenatal and intrapartum care in selected rural areas of Myanmar Kyaw Ooa, Le Le Wina, Saw Sawa, Myo Myo Mona, Yin Thet Nu Ooa, Thae Maung Maunga, Su Latt Tun Myinta, Theingi Myintb

Background: In Myanmar a large proportion of antenatal and intrapartum care in rural areas is provided by skilled birth attendants (SBAs), this study assessed the coverage by these health workers of all births, their adherence to service guidelines, and community opinion on the antenatal and delivery care they give in two rural health centres in Pathein Township, Ayeyarwaddy Region to identify the challenges and improve antenatal and intrapartum service delivery provided by the SBAs.

Method: A structured questionnaire was used to interview 304 women who had infants under one year of age, and in-depth interviews were held with 12 SBAs and 10 community members.

Results: Of the 304 pregnancies, 93% had received antenatal care (ANC); 97% of these were covered by SBAs at an average 15 weeks’ gestation. The average frequency of ANC visits was 9. Rates of home and hospital deliveries were 84.5% and 13.8% respectively. Among home deliveries, use rate of SBA was 51.4%, while for postnatal care, 31.3% was given by unskilled providers (traditional birth attendants (TBAs) 17.5%, auxiliary midwives (AMWs),13.8%). Multivariate analysis showed that interviewees aged 30 years and below (OR=0.468, P=0.046), with an education at primary level and below that of husband (OR=0.391, P=0.007) or not residing in the village of the rural/station health centre (OR=0.457, P=0.011) were significantly less likely to use SBAs. The categories of supervision, referral, and health education activities of SBAs were not in line with service guidelines. The main reasons were lack of access and community acceptance of TBAs.

Conclusion: Heavy workload, geographical location, transportation and financial concerns were major challenges for SBAs, along with community compliance and mutual coordination. Good communication and service management skills were important to overcome these challenges.

Key words: antenatal care, maternal and child health, Myanmar, skilled birth attendant (SBA), home delivery.

Introduction In Myanmar, maternal and child health, mortality of mothers and children. Safe including newborn care, has been prioritized motherhood initiatives have expanded into with the aim of reducing morbidity and a national movement focusing on the rural aDepartment of Medical Research (Lower Myanmar), Ministry of Health, Yangon, Myanmar. bDepartment of Health, Ministry of Health, Yangon, Myanmar. Correspondence to Yin Thet Nu Oo (e-mail: [email protected])

WHO South-East Asia Journal of Public Health 2012;1(4):467-476 467 Challenges faced by skilled birth attendants in providing antenatal Kyaw Oo et al. and intrapartum care in selected rural areas of Myanmar

population. Approximately 1.3 million women with 10 community members including TBAs give birth each year in Myanmar. The chance and local community leaders. Providers’ that a woman will die from pregnancy-related adherence to service delivery guidelines was causes is 1 in 33 and skilled birth attendants assessed by reviewing records and carrying (SBA) are present at only 40% of deliveries out in-depth interviews with 12 basic health and just over 20% of institutional deliveries.1 staff (BHS) in the selected health centres. Although improvements in the health status Quantitative data were entered by of mothers and children have been made, EpiDatatm and analysed by SPSS version 16. much more needs to be done to sustain the Qualitative data were processed in content gains.2 analysis to address compliance, satisfaction Even though mothers tend to deliver more and choice of services. Matrix analysis was children in rural than in urban areas, 80% of also made to identify existing challenges in deliveries are at home,3 of which the majority skilled attendance. are by unskilled birth attendants. High home delivery rates and low use of skilled providers Findings are addressed in the country’s strategic plans, which comprise improved skills and capacity Profile of study area of providers, family and community practice, Pathein Township had a population of 380 460 and service delivery systems. Information on and an urban–rural ratio of 1.01:1. The number challenges in improving these three major of women of reproductive age was 87 505. Its pillars would support Myanmar’s strategy for nine health centres count 126 BHS spread more effective implementation of a reduction over six rural health centres (RHC) and three in the maternal mortality ratio (MMR). Thus, station hospitals. There were 85 functioning this study tried to identify the challenges and AMWs and 66 trained TBAs, and in 2009, ANC ways to improve antenatal and intrapartum coverage was 59.6%, SBA rate was 51% and service delivery provided by skilled birth referral rate was 5.6%. Delivery rate attended attendants in rural area. by AMWs was 12.6% with a 9.7% referral rate, while deliveries attended by trained TBAs Methodology were 7.3%, with a 7.4% referral rate. Rates for maternal mortality, infant mortality and The study was conducted at Pathein Township neonatal mortality were 2.4, 15.1 and 4.7 in Ayeyarwaddy Region, which was selected per 1000 live births respectively. The abortion because of its low township coverage by SBAs rate was 3.04.5 compared with the national average (51% in Pathein, 63.6% national coverage).4 Two rural health centres – Thalakkhwar Station Health Past obstetric history and Unit and Kyetpaung Rural Health Centre– care-seeking practice were be selected for their relatively high and A total of 304 women were interviewed. low SBA rate respectively. A sample of 300 Their age ranged from 17 to 45 years with a mothers who had given birth in the previous mean (SD) of 29.4 +6.5 years. The average 12 months were interviewed with a structured family size was five, two of whom were questionnaire exploring antenatal and delivery earning members; 30% of women and 60% care-seeking practice. User opinions and of husbands were daily wagers and 42% of community perspectives on services provided women were dependent. About 90% of women were explored by key informant interviews

468 WHO South-East Asia Journal of Public Health 2012;1(4):467-476 Kyaw Oo et al. Challenges faced by skilled birth attendants in providing antenatal and intrapartum care in selected rural areas of Myanmar

and their husbands were of low to middle-level respectively. Antenatal booking rates tapered education. The ethic group Bamar accounted with later order of parity (Figure 1). The for 65% of the population, the remaining being location of the antenatal booking for first Kayin. The ratio of Buddhist to Christian was parity was mostly at health centres or general 2:1. Half of housing is built from wood and practitioner clinics. bamboo, and 35% are made from bamboo and thatch. Recent antenatal, intranatal The route to the nearest health centre for and postnatal care practice 60% of subjects was on foot, and for 24% Antenatal booking rates for last pregnancies by bicycle. The median duration to reach the was 93%, 97% of which occurred with skilled health centre was 20 minutes. Some 50% of health personnel. The average gestational women needed 10–30 minutes to reach the period of antenatal bookings was 15 weeks. nearest health centre. Travel expenses varied The average frequency of ANC visits was 9 from an average 1000 kyats to a maximum of (median=6). Half of women made antenatal 4000 kyats (equivalent to 5 US$). bookings between 12–20 weeks of gestation The median number of children of the and visited 4–10 times for ANC. One third of study group was 2. Of a total 776 previous pregnancies experienced oedema and 18% pregnancies, 84.4% had registered at the noticed high blood pressure. antenatal clinic. Stillbirth and abortion rates The last pregnancies were delivered on of these pregnancies were 0.7% and 1.8% average in the 37th gestational week. Half of

Figure 1: Percentage of antenatal booking by parities

WHO South-East Asia Journal of Public Health 2012;1(4):467-476 469 Challenges faced by skilled birth attendants in providing antenatal Kyaw Oo et al. and intrapartum care in selected rural areas of Myanmar

women delivered between 36 and 38 weeks’ deliveries had changed location between first gestation. The rate of SBAs used was 59.1%. labour pains and birth. Delivery rates at home and hospital were Nearly one-third (31.3%) of postnatal 84.5% and 13.8% respectively. Among home care was provided by unskilled workers deliveries, use rate of SBA was 51.4%. During (TBA=17.5% and AMW=13.8%). Postnatal delivery, 2.6% of cases changed birth place, problems occurred among 5.6% of cases after mainly to hospitals to receive the services of their last deliveries. Most problems were minor skilled providers. The referrals were mainly medical illnesses that were not referred. made by midwives, TBAs, AMWs and medical doctors (Figure 2). The major problem faced after delivery was haemorrhage. SBA rates Factors associated with at the start of labour, during delivery and the use of skilled birth after delivery were 52%, 53% and 52.9%, attendants indicating no significant difference in birth The use of SBA, operationally defined as attendance per phase of labour. Rates of “use at delivery”, was grouped into three skilled attendance at home delivery for all categories: never use, sometimes use and pregnancies was high – 91% at the start always use. Bivariate analysis showed that of labour – showing that most births were the following indicators were associated with planned at home. However, given that 88% higher use of SBAs: better access to a health of deliveries actually occurred at home, 3% of

Figure 2: Proportion of basic health services that need support to meet the set guidelines

470 WHO South-East Asia Journal of Public Health 2012;1(4):467-476 Kyaw Oo et al. Challenges faced by skilled birth attendants in providing antenatal and intrapartum care in selected rural areas of Myanmar

centre, i.e. residing in RHC main village (46% TBAs are over there since long, long ago. vs 35%, P=0.005), higher education level of They can give service for uncomplicated women (58% vs 36%, P=0.002) and husbands cases…Accessibility is their priority. (57% vs 34%, P<0.001), good housing status Deviations from antenatal care guidelines (P=0.04), younger age of woman (44% vs were the inability to cover 100% of cases, 37%, P=0.001) and husband (48% vs 35%, non-provision of some antenatal examinations P=0.003) (Table 1). Multivariate analysis (especially height measurement, urine albumin (multinomial logistic regression) showed that test), and some inadequate antenatal visits, younger women, i.e. 30 years and below exemplified as follows. (OR=0.468, P=0.046), low level of husband’s education, i.e. primary and below (OR=0.391, Many women did not come regularly and P=0.007), not residing near RHC (OR=0.457, for those women, we cannot give antenatal P=0.011) were significant factors for lower use care service for full frequency. of SBA (Table 2). …we have no time to check urine of all women. Thus we check only those who have Adherence to service high blood pressure… Height measuring also delivery guidelines could not be done for all women. Only if we The 12 BHS interviewed had an average total suspect she is short we measure and record. service of 15 years (range 3.5–29 year). Nine …sometimes we have to skip some BHS were graduates and the remaining three procedure stated in the guideline …to get had passed high school. A variety of registries satisfaction of clients, to save their time, and and report books were reviewed during the to get their compliance… interview. Among the three types of BHS, most Activities that deviated most frequently deviations were found in the provision of from the guideline instruction laid down by services by lady health visitors (LHV), and the the Department of Health6 were related to least in the activities of midwives (Figure 6). providing antenatal care, providing delivery Most frequent reasons were deficient facilities/ care, and service management. Others were equipment/materials, lack of accessibility and supervisory, referral and health education low compliance of the client. The next most activities (Figure 2). frequent reasons were time constraints, weak Regarding delivery care, most deviations supervision system and community compliance were due to difficult access and community or use of TBA services (Figure 3). acceptance and use of TBA services, as Regarding ANC service provision, improving illustrated below. management skills (including planning, We could not go to them because they are prioritizing and adapting to the situation) is staying in the middle of a paddy field alone the first priority for BHS in order to be able and there was no road. to face and overcome challenges. Secondly, improved skills in cooperation mechanisms Some area in my coverage is far from me with TBAs, AMWs and local authorities were and nearer to another midwife. She wishes to considered important (Figure 4). get service of that midwife. We could not avoid TBA because they are more accessible and have better relationship

WHO South-East Asia Journal of Public Health 2012;1(4):467-476 471 Challenges faced by skilled birth attendants in providing antenatal Kyaw Oo et al. and intrapartum care in selected rural areas of Myanmar

Table 1: Factors associated with use of SBA in previous and last pregnancies (bivariate analysis)

Use of skilled birth attendant Factor Total χ2 (df) P Never Sometimes Always Residing at or near to HC Residing at subcentre Freq 60 28 48 136 10.574 % 44.1 20.6 35.3 100 0.005 (2) Residing at HC Freq 44 47 77 168 % 26.2 28.0 45.8 100 Age (years) <= 30 Freq 69 30 77 176 13.604 % 39.2 17.0 43.8 100 0.001 (2) 31+ Freq 35 45 48 128 % 27.3 35.2 37.5 100 Age (years) husband <= 30 Freq 54 25 72 151 11.363 % 35.8 16.6 47.7 100 0.003 (2) 31+ Freq 50 50 53 153 % 32.7 32.7 34.6 100 Education level Illit+Pri Freq 87 58 80 225 12.867 % 38.7 25.8 35.6 100 0.002 (2) Middle or above Freq 16 17 45 78 % 20.5 21.8 57.7 100 Husband's education level Illit+Pri Freq 84 50 68 202 17.934 % 41.6 24.8 33.7 100 0.000 (2) Middle or above Freq 20 23 57 100 % 20.0 23.0 57.0 100 Housing status Bamboo Freq 49 49 62 160 % 30.6 30.6 38.8 100 12.151 Wooden or Brick Freq 8 5 20 33 0.016 (4) % 24.2 15.2 60.6 100 Other Freq 46 21 40 107 % 43 19.6 37.4 100

Freq=frequency; HC= health centre; Illit+Pri= grouping from illiterate to primary education. Other non-significant factors not described in the table are health centre (P=0.996); occupation (P=0.657); race (P=0.708); religion (P=0.537); family monthly income (P=0.74); husband’s occupation (P=0.21)

472 WHO South-East Asia Journal of Public Health 2012;1(4):467-476 Kyaw Oo et al. Challenges faced by skilled birth attendants in providing antenatal and intrapartum care in selected rural areas of Myanmar

Table 2: Multivariate analysis of significant factors for use of SBA (multinomial logistic regression)

Use of SBA P 95% CI of OR OR (Ref: Never use) value Lower Upper Age 0.046 0.5 0.2 1.0 (<=30 vs >30) Age of husband 0.104 1.8 0.9 3.8 (<=30 vs >30) Education 0.163 0.6 0.3 1.2 (<=primary vs >primary) Education of husband 0.006 0.4 0.2 0.8 Always (<=primary vs >primary) used Housing 0.718 0.8 0.3 2.2 (bamboo vs wood/brick) Housing 0.284 0.6 0.2 1.6 (other vs wood/brick) Health Centre 0.585 1.2 0.7 2.1 (RHC vs SHU) Residence 0.011 0.5 0.2 0.8 (sub-centre vs main)

OR, odds ratio; RHC, rural health centre; SHU, station health unit; Results for “sometimes used” category were omitted since the factors for use of SBA were similar to the results for “always used” category.

Figure 3:Reasons for deviation of service provision from the guideline

WHO South-East Asia Journal of Public Health 2012;1(4):467-476 473 Challenges faced by skilled birth attendants in providing antenatal Kyaw Oo et al. and intrapartum care in selected rural areas of Myanmar

with local people. In hard to reach areas, we to geographical barriers, or transportation have to cooperate with them. If we cooperate, difficulties, occurring seasonally in the risk cases will be more access to us and be outreach areas covered by midwives. Seasonal referred. fluctuations affect access to roads and diurnal variation of waterways in the delta area. Thus, Community opinion towards in an emergency, the inability to travel means that clients cannot use a skilled provider even skilled providers if they wanted to. Community informants felt that midwives A further challenge for skilled providers could cover about two thirds of all deliveries was poor cooperation of local people and and almost all antenatal activities and authorities to encourage clients to use skilled immunization, including tetanus toxoid personnel and to comply with referrals. They vaccination. The remaining deliveries were stated that community support was low, carried out by AMWs and sometimes, especially especially in delivery care. in hard-to-reach areas, by TBAs. Community informants noted that midwives would try to The last challenge, essential for referral, assume all deliveries in their responsible area was coordination between skilled providers but are unable to cover everything, which the of health centres from different levels. Poor community appreciated. coordination by the providers and financial concerns of the health centre makes the We see midwives frequently in our village provider reluctant to refer, and leads to client ... especially for vaccination. non-compliance. People know that midwives are more skilful Usually, local voluntary providers cooperate but could not always be available …[they] stay with midwife ANC and immunization services far away and thus more costly. by making arrangements for the location, Key informants estimated home delivery transportation and organization of clients. In to be about 90%, with hospital delivery some instances, cooperation between AMWs, only in cases of emergency. All respondents trained TBAs and midwives was supportive for agreed that the main challenge for midwives midwives to develop a close relationship with was overburden of their workload. A midwife local authorities and to gain client compliance. cannot rapidly reach all patients if she is Cooperation was indirectly supportive for responsible for 7–10 villages. In addition, midwives’ intrapartum care services in being midwives sometimes need to go to town for informed of problem cases, compliance with a meeting or training. referral, transportation assistance and the resolution of financial issues. Client access to health providers was low. Half of villages under SBA coverage were too far away from the sub-centre to reach Discussion in time. Emergency situations made clients Unpredictable obstetric complications are more inclined to use a health provider who developing in 15–20% of pregnancies. was nearer and could respond rapidly whether A major proportion of deaths resulting from or not s/he was skillful enough. Pregnant these complications occur either at home or women who were residing in less accessible in transit.4 Increased access to, and quality areas could not use the skilled provider due and quantity of health facilities in the country

474 WHO South-East Asia Journal of Public Health 2012;1(4):467-476 Kyaw Oo et al. Challenges faced by skilled birth attendants in providing antenatal and intrapartum care in selected rural areas of Myanmar

could reduce maternal mortality dramatically, percentage of births attended by skilled according to SriLanka experiences.7,8 Though health workers is considered to be the safe motherhood is a high priority programme, most appropriate process indicator. Current 90% of deliveries still occur at home and 50% estimates indicate that globally, only 56% of are attended by TBAs.2 births are assisted by a skilled person.9

Specific challenges remain for Myanmar to Although communities suggested more reduce its MMR. Low education and economic skilled staff in their areas, manpower and status, and lack of access influence the client’s resources are scarce. The country’s strategic decision to use skilled providers in pregnancy plans should therefore focus on improving care. Providers also highlighted accessibility as the capacity of current health providers, a challenging factor to increase SBA use. This especially in management and communication; challenge was reiterated by community key and improving family and community informants for skilled providers to be able to practice through close coordination with provide full coverage. Accessibility means all community authorities and health volunteers. geographical, managerial and financial aspects Regarding cooperation, the delegation of both by clients and providers. For some some services, accompanied by close and midwives, their area of responsibility and the supportive supervision, good communication size of the population were too large for them and motivation with positive attitudes, would to meet their targets for full coverage of ANC be effective. This study highlights the major and SBA. Geographical barriers can prevent challenges and suggests priorities and ways both clients and providers from meeting their to overcome them. The objective would be to needs. Deviations of skilled providers’ services implement a more effective national strategic from the guideline were mainly caused by plan, with the ultimate goal of reducing having responsibility for too wide area and MMR. too large population size, little time and lack of facilities. Service manuals were not Acknowledgements found in all BHS hands. Many descriptions in the manuals confuse them. The role of This study was conducted with funding lady health visitors in rural areas should be from WHO. The Directors Generals of the emphasized with practical service-provision Department of Medical Research (Lower and supervisory activities over the midwives, Myanmar) and the Department of Health especially for coordination between BHS and gave their kind permission and supervision the community. for the field work. The health authorities of Ayeyarwaddy Regional Health Division and Furthermore, all perspectives considered Pathein Township Health Department are that clients’ habitual use of unskilled providers specially thanked for their hospitality and kind was due to a low level of education and because assistance in field data collection. unskilled providers were more accessible than skilled providers. References Conclusion 1. Myanmar, Ministry of Health. Health services in Skilled attendance for all births is the only Myanmar. Yangon: MOH - http://www.whomyanmar. way to ensure emergency obstetric care org/files/Health_in_Myanmar_2006/06_Health_ Services_in_Myanmar_edited.pdf -accessed 17 April for all pregnancies with complications. The 2013.

WHO South-East Asia Journal of Public Health 2012;1(4):467-476 475 Challenges faced by skilled birth attendants in providing antenatal Kyaw Oo et al. and intrapartum care in selected rural areas of Myanmar

2. Myanmar, Ministry of Health. Health in Myanmar 6. Myanmar, Ministry of Health. Service delivery 2009. Yangon: MOH, 2009. pp. 44 guideline for basic health staff. Yangon: Department of Health, 2006. 3. Kumara Rai N, Sanu Maiyan Dali. Making pregnancy safer in South-East Asia. Regional Health Forum WHO 7. Seneviratne HR, Rajapaksa LC. Safe motherhood in South-East Asia Region. 2002; 6(11): 912-6 Sri Lanka: a 100-year march. International Journal of Gynecology and Obstetrics. 2000; 70: 113–124. 4. Myanmar, Ministry of Health. Reproductive health statistics 2008. Yangon: Department of Health 8. Pathmanathan I, Liljestrand J, Martins JM, Rajapaksa Planning, 2008. LC, Lissner C, de Silva A, et al. Investing in maternal health: learning from Malaysia and Sri Lanka. 5. Myanmar, Ministry of Health. Township health profile Washington, DC: World Bank, 2003. 2009. Yangon: Pathein Township Health Department, 2009. 9. UNFPA. Skilled attendance at birth - http://www. unfpa.org/public/mothers/pid/4383 -accessed 17 April 2013.

476 WHO South-East Asia Journal of Public Health 2012;1(4):467-476 Public health classic

Epidemiologic investigation of excess maternal and neonatal deaths and evidence-based low-cost public health interventions – Ignaz Semmelweis: the etiology, concept and prophylaxis of child bed fever1 S D Guptaa

Introduction When I look back upon the past, I can only to accept that they could be responsible for dispel the sadness which fall upon me by spreading infection. Semmelweis’s claims gazing into that happy future when the were thought to lack scientific basis and infection will be banished …. The conviction his practice of antiseptic use only gained that such a time must inevitably sooner than acceptance years after his death, when Louis later arrive will cheer my dying hour. So wrote Pasteur3 invented the germ theory of disease Ignaz Semmelweis in his last days while he in 1862, and Joseph Lister4 invented methods suffered depression due to continued criticism for antiseptic surgery in 1867. by medical professionals and colleagues. Ignaz Semmelweis’s investigations of Semmelweis suggested that simple hand- maternal and neonatal deaths laid the washing with chlorinated water could prevent foundation of modern public health and the spread of infection and save the lives of epidemiological investigations, and prevention women in maternity.2 His methods showed of avoidable deaths through hand-washing with that maternal deaths due to puerperal fever chlorine and lime before clinical examination of reduced from 12.2% to 2.4% in maternity patients. Semmelweis’s classic approach not wards. He expanded his methods to washing only contributed to our knowledge of causal of instruments, which altogether removed hypothesis, but discovery of appropriate puerperal infection in the hospital. However, interventions supported by evidence that his observations were in serious conflict led to a reduction in maternal and neonatal with contemporary scientific and medical deaths. opinions, which rejected his observations. His suggestion that doctors should wash their Semmelweis’s insights are a brilliant hands with chlorinated lime even offended example of translation of research into some doctors and his colleagues. They refused action for wider human well-being. Today,

aIndian Institute of Health Management Research, Jaipur, India. Correspondence to S D Gupta (e-mail: [email protected])

WHO South-East Asia Journal of Public Health 2012;1(4):477-484 477 Epidemiologic investigation of excess maternal and neonatal deaths S D Gupta and evidence-based low-cost public health interventions

the prevention of puerperal sepsis and Investigations of childbed infections resulting from deliveries conducted fever (puerperal fever) in unhygienic conditions and unsafe abortions Semmelweis’s investigation on excess is a major approach to reduce maternal maternal and neonatal deaths is a testimony and neonatal mortality all over the world. of modern concepts and approaches of The impact of Semmelweis’s discovery on epidemiological investigations. His work human health was so immense that millions demonstrated how simple observations and of maternal deaths have been averted since comparison of frequency of events could the initiation of such a simple intervention help develop causal hypothesis and testing as the adoption of aseptic procedures during to draw conclusive evidence of cause and childbirth. Globally, the estimated 526 000 effect relationships. He laid the foundation maternal deaths in 1980 are reported to of the modern-day epidemiologic approach: have been reduced to 342 000 in 2008, a identify and define problem; define research reduction of almost one third. Yet a large question and develop hypothesis causation number of maternal deaths, about 90%, are and mode of transmission; collect and analyse easily preventable through simple technical data; establish cause and effect relationship; and social interventions. Incidentally, 99% of develop and design intervention; and evaluate maternal deaths globally occur in developing impact. countries where home deliveries are still common and infection control measures are Semmelweis observed excess maternal not adhered to.5 Puerperal sepsis thus is deaths among women who gave birth in still prevalent in developing countries and Vienna Maternity Hospital.1 Of the two continues to present a significant risk of maternity clinics operated by the hospital, he obstetric morbidity and mortality for women found that maternal deaths were higher in in these regions.6 the first. His data collected data on maternal deaths over a period of six years since 1841 We briefly describe below the investigations showed that, on average, the mortality rate of Ignaz Semmelweis and prophylaxis was three times higher in the first clinic over measures that enabled him to demonstrate the second clinic. He also noted that the the significant reduction in maternal and mortality rate was five times higher in the first neonatal deaths. clinic in 1846 (Table 1). Study setting Semmelweis further observed that the deaths in the first clinic would have been Semmelweis’s work was based in Vienna higher as the serious patients were transferred General Hospital in 1846 where he worked to Vienna General Hospital. When these as an assistant after receiving a degree in patients died, they were included in the medicine with specialization in midwifery, mortality data of general hospital rather than in parallel to studying statistics. The Vienna the first clinic. He also observed that the General Hospital operated two maternity referral rates of patients in first clinic were clinics, and was a major centre for midwifery higher than the second clinic due to higher training in Europe, attracting a large number frequency of serious patients. of foreign medical and midwifery students along with the local students. He ascribed these deaths to childbed fever. He defined childbed fever “as a disease

478 WHO South-East Asia Journal of Public Health 2012;1(4):477-484 S D Gupta Epidemiologic investigation of excess maternal and neonatal deaths and evidence-based low-cost public health interventions

Table 1: Annual births, deaths, and maternal mortality rates at the first and second clinics of the Vienna Maternity Hospital from 1841 to 1846

First clinic Second clinic Year Mortality Mortality Births Deaths Births Deaths rate rate 1841 3 036 237 7.8 2 442 86 3.5 1842 3 287 518 15.8 2 659 202 7.6 1843 3 060 274 9.0 2 739 164 6.0 1844 3 157 260 8.2 2 956 68 2.3 1845 3 492 241 6.9 3 241 66 2.0 1846 4 010 459 11.4 3 754 105 2.8 Total 20 042 1989 17 791 691 Average 3 340 331.5 9.9 2 965 115 4.0 Source1.

characteristic of and limited to maternity excess maternal and newborn deaths? What patients, for whose origin the puerperal state was the source and how did these women and specific causal moments were necessary’. acquire that causative agent? He also observed higher mortality among In order to address his research questions, newborns delivered by women and also died Semmelweis explored whether the women in the first clinic due to similar clinical and admitted to the first clinic were different to pathological characteristics (Table 2).

Table 2: Annual births, deaths, and mortality rates for newborns at the first and second clinics of the Vienna Maternity Hospital from 1841 to 1846

First clinic Second clinic Year Mortality Mortality Births Deaths Births Deaths rate rate 1841 2813 177 6.3 2252 91 4.0 1842 3037 279 9.2 2414 113 4.7 1843 2828 195 6.9 2570 130 5.1 1844 2917 251 8.6 2739 100 3.7 1845 3201 260 8.1 3017 97 3.2 1846 3533 235 6.7 3398 86 2.5 Total 18 329 1397 16 390 617 Average 3 055 233 7.6 2 732 103 3.9

Research questions and those in the second clinic (now called selection hypotheses bias). He reviewed the admission process of The prime research question he had in his the maternity hospital selection bias, if any. mind was why there was higher maternal The admission of maternity patients was mortality in the first clinic? What caused regulated as follows: Monday afternoon

WHO South-East Asia Journal of Public Health 2012;1(4):477-484 479 Epidemiologic investigation of excess maternal and neonatal deaths S D Gupta and evidence-based low-cost public health interventions

at four o’clock admissions began in the since 1940 when it started training only first clinic and continued until Tuesday obstetricians, while the second clinic trained afternoon at four. Admission then began only midwives, and wondered whether this in second clinic and continued until could be the reason for the discrepancy. Did Wednesday afternoon at four o’clock. obstetricians cause more damage to the At that time admissions resumed in the birth canal during examination? As most of first clinic until Thursday afternoon, the physicians were foreigners, it was also etc. On Friday afternoon at four o’clock posited whether they were rougher than the admissions began in the first clinic and natives? Incidentally, mortality rates declined continued through forty-eight hours until when foreigner doctors were excluded from Sunday afternoon, at which time again examinations in late 1846 and early 1847; the admissions began in the second clinic. however, they remained higher in the first Admissions alternated between two clinics clinic, and climbed back to higher levels in through twenty-four periods, and only the months that followed (Figure 1). It did not once a week did admissions continue in seem logical that examination by obstetricians, the first clinic for forty-eight hours1. especially foreigners, was associated with higher mortality in the first clinic. He found that the admissions were done regularly without any predetermined criteria; These observations astounded the hospital he concluded that there was no selection bias authorities and Semmelweis. The mystery that would render women admitted to the first of higher maternal deaths in the first clinic clinic more susceptible. remained unresolved.

He considered several factors that could It is interesting to learn how exceptional possibly be associated with excess maternal incidents can play a crucial role in deaths in the first clinic, included various epidemiological investigations of causes clinical conditions, prolonged dilation and and mode of transmission of a disease. other delivery-related characteristics. He Semmelweis left for England to study English also considered environmental factors and in the winter of 1846, returning to Vienna the geographic location of the clinics and in March 1847. On his return, he learnt that arrangement of maternity beds in the wards. Professor Jakob Kolletschka, who he admired, He found no conclusive difference in maternal had died of symptoms similar to those of deaths. maternity patients in his clinic. He described his case history as follows: The authorities were also concerned with disturbing differences in maternal deaths in Kolletschka, Professor of Forensic Medicine, the two clinics. In order to reduce deaths, often conducted autopsies for legal several measures were adopted such as purposes in the company of students. transfer of patients to the general hospital, During one such exercise, his finger was change of physician and medical procedures, pricked by a student with same knife that environmental conditions and rooms from was being used in the autopsy… Professor time to time. Various other measures were Kolletschka contracted lymphangitis also adopted. However, the maternity patients and phlebitis in the upper extremity… continued to die of childbed fever. He died of bilateral pleurisy, pericarditis, peritonitis, and meningitis. A few days Semmelweis then discovered that mortality before he died, a metastasis also formed in was consistently greater in the first clinic

480 WHO South-East Asia Journal of Public Health 2012;1(4):477-484 S D Gupta Epidemiologic investigation of excess maternal and neonatal deaths and evidence-based low-cost public health interventions

Figure 1: Monthly mortality rates at the first clinic of the Vienna Maternity Hospital during 1846 and 1847

20 18 Chloring Handwash 16 14

ate 12

yr 10 8

Mortalit 6 4 2 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

1846 1847

one eye. ... I was haunted by Kolletschka’s In Kolletschka, the specific causal factor was disease and was forced to recognize, cadaverous particles that were introduced even more decisively, that the disease into his vascular system. I was compelled from which Kolletschka died was identical to ask whether cadaverous particles had to that from which so many maternity been introduced into vascular systems of patients died. those patients whom I had seen die of this identical disease. I was forced to answer It proved to be clinching evidence for the affirmatively. cause of childbed fever. In order to answer yet another important Semmelweis had earlier shown from question, i.e. how cadaver particles were autopsies that the cause of newborn deaths transferred to maternity patients, Semmelweis were identical to those of women who had analysed the chain of sequences and died from childbed fever, and thus concluded circumstances in which these particles could that they died of the same disease. The same have contaminated the wounds. results were found in Kolletschka’s autopsy, confirming that he died of the same disease,. The obstetrics students were assigned Semmelweis thus notes: only to the first clinic and the midwives to the second clinic. The obstetrics students were The exciting cause of Professor Kolletschka’s required to carry out anatomical orientation death was known; it was the wound by on cadavers as a part of their training. They autopsy knife that had been contaminated by washed their hands with soap and water, but cadaverous particles ... Kolletschka was not this did not destroy the cadaverous particles the first to have died in this way. I was forced on their hands. They then examined maternity to admit that if his disease was identical with patients and thus contaminated the genitals the disease that killed so many maternity of pregnant and delivering patients with patients, then it must have originated from the cadaverous particles. These particles were same cause that brought it on in Kolletschka.

WHO South-East Asia Journal of Public Health 2012;1(4):477-484 481 Epidemiologic investigation of excess maternal and neonatal deaths S D Gupta and evidence-based low-cost public health interventions

subsequently introduced into the vascular practice began in middle of May 1847 systems of maternity patients, producing a … Both students and I were required to similar disease to that which killed Kolletschka. wash (hands) before examinations. After On the other hand, the trainee midwives in a time I ceased to use Chlorina Liquida the second clinic were not required to carry because its high price, and I adopted the out anatomical interventions on cadavers less expensive chlorinated lime. In May and hence did not visit the mortuary. The 1847, during the second half of which assistants, too, in the second clinic seldom chlorine washings were first introduced, visited the mortuary. These observations 36 patients died – this was 12.24 percent led Semmelweis to infer that the students of 294 deliveries. In the remaining seven whose hands were contaminated with cadaver months of 1847, the mortality rate was particles were responsible for transmission of below that of the second clinic. childbed fever. Monitoring and evaluation Intervention of intervention History was made. The cause of childbed Semmelweis evaluated the effect of chlorine fever (puerperal sepsis) was discovered which washing of hands and equipment, and still kills tens of thousands of women during demonstrated that chlorine prevented the pregnancy and childbirth in developing countries transmission and spread of childbed fever as today. Semmelweis went forward to discover observed from reduction in mortality rates an intervention to prevent the spread of the among maternity patients in the first maternity disease. He hypothesized that, if the cadaverous clinic compared with the mortality levels in the particles are destroyed chemically, the disease second maternity clinic (Table 3). must reduce or be prevented. He instituted In these seven months, of 1841 maternity hand-washing with chlorine solution before patients cared for, 56 died (3.04%). In examining maternity patients (Figure 1): 1846, before washing with chlorine was To destroy cadaverous matter adhering introduced, 4010 patients cared for, 459 to hands, I used Chlorina Liquida. This died (11.4%). In the second clinic in

Table 3: Annual births, deaths, and mortality rates for all maternity patients at the clinics of the Vienna Maternity Hospital for 1839, 1840, 1846–1948

First clinic Second clinic

Year Births Deaths Rate Births Deaths Rate

Pre-training year 1839 2781 151 5.4 2010 91 4.5

Obstetric training permitted 1840 2789 267 9.6 2073 55 2.7

Before chlorine washing 1846 4010 459 11.4 3754 105 2.8

Chlorine washing introduced 1847 3490 176 5.0 3306 32 1.0

After chlorine washing 1848 3556 45 1.3 3219 43 1.3

482 WHO South-East Asia Journal of Public Health 2012;1(4):477-484 S D Gupta Epidemiologic investigation of excess maternal and neonatal deaths and evidence-based low-cost public health interventions

1846, of 3754 patients, 105 died (2.7%). and health care in general. Infection control In 1847, when approximately the middle measures, especially disinfection of hands and of May I instituted washing with chlorine, equipment have saved the lives of millions in the first clinic of 3490 patients, 176 died of pregnant and childbearing women who (5%). In the second clinic, of 3306 patients may have died of puerperal sepsis all over 36 died (1%). In 1948, chlorine washings the world, thanks to the great discovery of were employed throughout the year and Semmelweis. Hospital infection control is the of 3556 patients, 45 died (1.3%). In the hallmark of patient safety today. second clinic in the year 1948, of 3219 Semmelweis’s seminal observations led patients 43 died (1.3%). the way for modern epidemiologic thinking and approach. Observations of disease Replication of interventions trends and making comparisons are critical Semmelweis was convinced that hand-washing components of modern-day epidemiological with chlorinated lime before examining the approaches for initiating investigations and patient by doctors would reduce maternal and testing causal hypotheses. Semmelweis neonatal deaths. He suffered ignominy at the systematically observed maternal death hands of his fellow colleagues and seniors. trends in the maternity hospital and developed He left the Vienna hospital as he could no a causal hypothesis. Though he could not longer endure the frustrations of dealing with explain the causal agent, he successfully the Viennese medical establishment, and investigated the source of infection and returned to Pest in Hungary. Despite being mode of spread of puerperal fever among criticized by the medical community and women who gave birth in his hospital. He also fellow colleagues, Semmelweis replicated his successfully demonstrated that if the chain intervention by introducing hand-washing of spread was blocked, the disease could be with chlorine water in Pest’s small St Rochus prevented. Nevertheless, Semmelweis could Hospital where childbed fever was rampant. not provide scientific evidence to prove his From 1851 to 1855, only eight patients died of hypothesis that cadaveric particles produced childbed fever out of 933 births (0.85%). The puerperal fever. He was severely criticized by results he demonstrated in Vienna Hospital his contemporary medical professionals and were thus reproducible. He repeated the same scientists for bruising their ego and for his interventions in 1857 when he took over as statements that doctors were responsible for Professor at the University of Pest maternity the spread of infection. clinic, where the results were once again Semmelweis was reluctant to publish 7 impressive. his work and disseminate the results among the scientific community until 1861 when he Research for action published his book on Concepts, Etiology and Prophylaxis of Childbed Fever, 16 years after Research remains impotent unless it is his discovery. Dissemination of research is used to protect and improve human health critical to create awareness, advocacy and and promote welfare through public health acceptance of its results, in order to further action. Semmelweis’s research is a salient its applications as evidence for policy and example of how research was employed to action. However, there are several barriers improve human health. His observations and that obstruct meaningful dissemination and experiments formed the basis of modern- uptake of the results. Research evidence day infection control measures in hospitals is one of many inputs to acceptance and

WHO South-East Asia Journal of Public Health 2012;1(4):477-484 483 Epidemiologic investigation of excess maternal and neonatal deaths S D Gupta and evidence-based low-cost public health interventions

decision-making by peers, fellow professional 3. Pasteur L. Scientific papers: The germ theory and its groups and clinicians, health-care managers, applications to medicine and surgery. Vol. 38, Part 7. and health-care policy-makers. Translating The Harvard Classics. New York: P.F. Collier & Son, 1909-14; Bartleby.com, 2001 - http://www.bartleby. research into action requires intellectual com/38/7/1.html - accessed 26 April 2013. rigour, discipline, creativity, clinical judgment 4. Lister J. On the antiseptic principle of the practice of and skills, a favourable organizational climate, surgery. Vol. 38, part 6. The Harvard Classics. New and above all endurance and tolerance York: P.F. Collier & Son, 1909-14; Bartleby.com, 2001 towards contemporary scientific reviews and - http://www.bartleby.com/38/6/1.html - accessed comments.8–13 Semmelweis was academically July 26, 2012.

brilliant and made a seminal discovery, but 5. World Health Organization. Trends in maternal lacked leadership skills. He was arrogant, mortality: 1990-2008. Estimates developed by WHO, severe on his critics, calling them murderers UNICEF, UNFPA and The World Bank. Geneva: WHO, and medical ‘Neros’, and accused his superiors 2010 - http://www.who.int/reproductivehealth/ of causing the deaths of mothers. He became publications/monitoring/9789241500265/en/ - accessed 16 April 2012. frustrated and angry with himself, but refused to publish his monumental discovery. 6. Abouzahr C, Aaahman E, Guidotti R. Puerperal sepsis and other puerperal infections. In: Health dimensions He was denied recognition in his lifetime of sex and reproduction: the global burden of sexually and finished his days in a mental asylum in transmitted diseases, maternal conditions, perinatal disorders, and congenital anomalies, eds. CJL Murray ignominy, dying at the very young age of 47 and AD Lopez. Geneva: WHO, 1998. years. His work was recognized only after about two decades of his demise with the advent 7. Carter, K. Codell; Carter, Barbara R. Childbed fever: a scientific biography of Ignaz Semmelweis, Transaction of germ theory and the value of antiseptic Publishers. Rev. Edn. Praeger. 2005. techniques in preventing spread of infection. He was described as the saviour of mothers. 8. Lavis J, Davies H, Oxman A, Denis JL, Golden-Biddle K, Ferlie E. Towards systematic reviews that inform His discovery of hand-washing with antiseptic health care management and policy-making. J Health solution left a legacy in modern medicine Serv Res Policy. 2005 Jul; 10 Suppl 1:35–48. and infection control, thus saving the lives of 9. Kitson A, Ahmed LB, Harvey G, Seers K, Thompson millions of women; not only averting deaths DR. From research to practice: one organizational from nosocomial infections, but reducing the model for promoting research-based practice. J Adv cost of health care the world over. The Austrian Nurs. 1996 Mar; 23(3): 430-40.

Government issued a commemorative 50 10. Pang T. Evidence to action in developing world. What euro coin in 2008 in his honour; and libraries, evidence is needed? Bull World Health Organ. 2007; museums and universities have been named 85: 247. Semmelweis in recognition of his discovery 11. World Health Organization. World report on knowledge and contribution to medical science. for better health: strengthening health systems. Geneva: WHO, 2004.

12. Haines A, Kuruvilla S, Borchert M. Bridging the References implementation gap between knowledge and action 1. Ignaz Semmelweis. Etiology, concept and prophylaxis for health. Bull World Health Organ. 2004 Oct; of childbed fever. Translated by T. Codell Carter. 82(10): 724-31. Wisconsin: University of Wisconsin Press, Ltd., 13. Dobbins M, Cockerill R, Barnsley J. Factors affecting 1983. the utilization of systematic reviews. A study of public 2. Hani. Semmelweis’ germ theory. Explorable, 2010 - health decision makers. Int J Technol Assess Health http://www.experiment-resources.com/semmelweis- Care. 2001 Spring; 17(2): 03-14. germ-theory.html - accessed 16 April 2013.

484 WHO South-East Asia Journal of Public Health 2012;1(4):477-484 Recent WHO publications

Achieving the health-related millennium development goals in the South-East Asia Region: measuring indicators 2012

Publisher: World Health Organization, Regional Office for South-East Asia. New Delhi: 2013. 179 pages. ISBN: 9789290224297

This publication presents by countries in the past 10 years and the gaps that the achievements on have to be bridged in the remaining five years, along MDGs by Member States with ascertaining the likelihood of reaching each of of the WHO South-East the targets by 2015. The report also highlights the Asia Region gauged only bottlenecks, the most intense challenges and the at the national level. It constraints faced by each country in its strategic depicts the road covered actions and interventions.

South-East Asia Regional Strategic Framework for Improving Neonatal & Child Health and Development

Publisher: World Health Organization, Regional Office for South-East Asia. New Delhi: 2012. 83 pages. DOC NO.- SEA/CAH/09

Despite a significant improve- implementation requires multisectoral collabora- ment in child mortality, the tion. Member States are encouraged to use local South-East Asia (SEA) Region situation assessment, improve convergence within is unlikely to achieve the Mil- Health Programmes, and to strengthen health lennium Development Goal systems and enhance thecapacity in planning and (MDG) 4 target and needs management of the child health programmes for significant improvement in the scaling up of evidence-based interventions maternal, newborn and child for child health and development. The Framework health (MNCH) programmes presents a brief Regional situational analysis that in Member States. The broad has been used to form the basis for appropriate objective of the WHO-UNICEF strategies for child and neonatal health. A brief joint South-East Asia Regional Strategic Frame- analysis of the broader determinants beyond work for Improving Neonatal & Child Health and the health system such as safe drinking-water, Development is to guide and assist Member States sanitation and hygiene, education, gender, is also to develop or strengthen their national strategies included, as well as an analysis of coverage levels and plans to improve newborn and child survival, of existing interventions that helps identify gaps reduce the burden of child morbidity and disability and missed opportunities for strengthening health and promote child health and development. The systems. The Framework is based upon principles Strategic Framework encourages Member States of child rights and equity. It highlights the need to to incorporate an equitable child health and de- think about the whole child, not just their health. velopment perspective into policies and actions The Framework uses the principle of continuum of within the health system and emphasizes that the care across the life-course, from pre-conception

WHO South-East Asia Journal of Public Health 2012;1(4):485-486 485 through pregnancy, childbirth, the postnatal period, nuity across different levels of health-care delivery infancy, childhood and adolescence. Such a holistic from home/community to first-level health centres approach is important since maternal, neonatal and referral hospitals. This Framework is intended and child health are closely linked with each other, to be the basis for subsequent joint WHO-UNICEF not only intrinsically, but also programmatically. It country support for Member States in the WHO emphasizes that the services must be organized SEA Region for newborn and child health and de- through a process that preserves functional conti- velopment.

486 WHO South-East Asia Journal of Public Health 2012;1(4):485-486 WHO South-East Asia Journal of Public Health WHO South-East Asia Journal of Public Health Guidelines for contributors The WHO South-East Asia Journal of Public Health (WHO-SEAJPH) is a Volume 1, Issue 4, October–December 2012, 359-486 peer-reviewed, indexed (IMSEAR), open access quarterly publication Original research articles on public health, primary health Text Advisory Board care, epidemiology, health administration, health systems, of the World Health Organization, Regional Office for South-East Asia. The main text of the paper should begin with title and an health economics, health promotion, public health nutrition, Samlee Plianbangchang abstract which should be structured under the following The Journal provides an avenue to scientists for publication of original communicable and non-communicable diseases, maternal and (Chair) headings: title, background, methods, results, conclusions, research work so as to facilitate use of research for public health child health, occupational and environmental health, social Richard Horton and 5-8 key words for indexing (http://www.ncbi.nlm.nih. and preventive medicine are invited which have potential action. Hooman Momen gov/entrez/query.fcgi?db=mesh). It should not exceed 250 to promote public health in the South-East Asia Region. We David Sanders words. Follow the IMRD format, i.e. Introduction, Methods, also publish editorial commentaries, perspectives, state of Editorial Process Fran Baum Results and Discussion. Manuscript must be written in a the art reviews, research briefs, report from the field, policy manner, which is easy to understand, and should be restricted All manuscripts are initially screened by editorial panel for scope, relevance and Lalit M Nath and practice, letter to the editor and book reviews etc. scientific quality. Suitable manuscripts are sent for peer review anonymously. Chitr Sitthi-amorn to the topic being presented. The discussion should end in Submitted articles are peer reviewed anonymously conclusion statement. Each table and figure should be on Recommendations of at least two reviewers are considered by the editorial Editorial Board and confidentially. The manuscript should be original which separate page at the end of the manuscript. panel for making a decision on a manuscript. Accepted manuscripts are edited has neither been published nor is under consideration for for language, style, length etc. before publication. Authors must seek permission Poonam K Singh publication in any substantial form in any other publication. Acknowledgement from the copyright holders for use of copyright material in their manuscripts. (Chair) All published manuscripts are the property of the World Place it as the last element of the text before references. Submissions are promptly acknowledged and a decision to publish is usually Sattar Yoosuf Health Organization. communicated within three months. Nyoman Kumara Rai Written permission of person/agency acknowledged should Aditya P Dash Submit article to the Chief Editor, as an electronic copy be provided. through CD ROM or email [email protected], World Disclaimer Sangay Thinley Health Organization, Regional Office for South-East Asia, Athula Kahandaliyanage References The designations employed and the presentation of the material in this Indraprastha Estate, Mahatma Gandhi Road, New Delhi, 110 Quazi Munirul Islam publication do not imply the expression of any opinion whatsoever on the part 002, India. Tel. 91-11-23309309, Fax. 91-11-23370197. Only original references should be included. Signed permission of the World Health Organization concerning the legal status of any country, Nata Menabde is required for use of data from persons cited in personal territory, city or area or its authorities, or concerning the delimitation of its Maureen E Birmingham communication. ANSI standard style adapted by the National Manuscript preparation frontiers or boundaries. Dotted lines on maps represent approximate border Mahmudur Rahman Library of Medicine should be followed (http://www.nlm.nih. lines for which there may not yet be full agreement. Chencho Dorji Type the article in double space (including references) gov/bsd/uniform_requirements.html). References should be Hasbullah Thabrany using 12 point font with at least one inch margin on all numbered and listed consecutively in the order in which they The mention of specific companies or of certain manufacturers’ products Nay Soe Maung sides. Uniform Requirements for Manuscripts submitted to are first cited in the text and should be identified in the text, does not imply that they are endorsed or recommended by the World Health Biomedical Journals should be consulted before submission tables and legends by Arabic numerals as superscripts. The Suniti Acharya Organization in preference to others of a similar nature that are not mentioned. of the manuscript (http://www.icmje.org ). All articles should full list of references placed at the end of the main text of the Rohini de A Seneviratne All reasonable precautions have been taken by the World Health Organization mention how human and animal ethical aspects of the study paper should include surnames and initials of all authors up to verify the information contained in this publication. However, the published Phitaya Charupoonphol have been addressed. When reporting experiment on human to six (if more than six, only the first six are given followed material is being distributed without warranty of any kind, either expressed or Mariam Cleason subjects indicate whether the procedures followed were in by et al.); the title of the paper, the journal title abbreviation implied. The responsibility for the interpretation and use of the material lies Kenneth Earhart accordance with the Helsinki Declaration (http://www.wma. according to the style of Index Medicus, year of publication; with the reader. In no event shall the World Health Organization be liable for Jacques Jeugmans net). volume number; first and last page number. Reference of books should give the surnames and initials of the authors, the damages arising from its use. The named authors alone are responsible for Guidelines for preparation of manuscripts for various sections Editorial Team book title, place of publication, publisher and year; citation of the views expressed in this publication. of WHO SEAJPH are available on website (http:www.searo. chapters in a multiple author book should include the surname Aditya P Dash who.int/who-seajph) and initials of the authors, chapter title, and names and (Chief Editor) Permissions initials of editors, book title, place of publication, publisher Charles Raby © World Health Organization 2013. All rights reserved. and year and first and last page numbers of the chapter. Title page (Coordinator) For citing website references, mention author surname and Requests for permission to reproduce or translate WHO publications – whether Rajesh Bhatia This should contain the title of the manuscript (not more than initials, title of the article, the website address, and the date for sale or for non-commercial distribution should be addressed to Information (Associate Editor) 150 characters including spaces), the name of all authors on which website was accessed. Management and Dissemination Unit, World Health Organization, Regional Nyoman Kumara Rai (first name, middle initials, and surname), a short title (not Office for South-East Asia. We encourage other users to link to articles on our Sunil Senanayake more than 40 characters including spaces), name(s) of the Tables and figures web site, provided they do not frame our content; there is no need to seek our Sudhansh Malhotra institution(s) where the work has been carried out, and the address of the corresponding author including telephone, fax permission. Richard Brown The tables/figures must be self explanatory and must not and e-mail. One of the authors should be identified as the Nihal Abeysinghe duplicate information in the text. Each table and figure must corresponding author and guarantor of the paper who will have a title and should be numbered with Arabic numerals. Prakin Suchaxaya Editorial Office take responsibility of the article as a whole. Briefly mention Each table should be typed in double space on a separate WHO South-East Asia Journal of Public Health Roderico Ofrin the contribution of each author in a multi-author article. Also sheet of paper. Figures should be of the highest quality, i.e. World Health Organization, Regional Office for South-East Asia, Renu Garg mention conflict of interest, the source of support in the form glossy photographs or drawn by artist or prepared using Indraprastha Estate, Mahatma Gandhi Road, New Delhi 110 002, India. of grants, equipment, drugs etc. Word count of the abstract standard computer software. A descriptive legend must Tel. 91-11-23309309, Fax. 91-11-23370197 and main text, number of references, figures and table should accompany each figure which should define all abbreviations e-mail: [email protected] also be mentioned on the bottom of title page. used. All tables and figures should be cited in the text. Web site: www.searo.who.int/publications/journals/seajph WHO South-East Asia Journal of Public Health ISSN 2224-3151 Volume 1, Issue 4, October—December 2012, 359-486

Volume 1, Issue 4, October–December 2012, 359-486 Volume 1, Issue 4, October–December 2012, 359-486 Inside Editorial Research brief Success of tuberculosis and HIV collaboration Evidence of HPV subtypes linked with cervical Iyanthi Abeyewickreme, Neelamanie cancer in Nepal Punchihewa, Amaya Maw-Naing 359 Chop L Bhusal, Sulochana Manandhar, Meeta Singh, Aarati Shah, Perspective Sushharma Neupane, Dibesh Karmacharya, Janani Suraksha Yojana: the conditional cash Kate Cuschieri, Heather Cubie, transfer scheme to reduce maternal mortality in Duncan C Gilbert, Sameer M Dixit 441 India – a need for reassessment Rajesh Kumar Rai, Policy and practice Prashant Kumar Singh 362 Institutionalizing district level infant death review: an experience from southern India Pandemic influenza preparedness planning: Sanjeev Upadhyaya, Sudeep Shetty, lessons from Cambodia Selva S Kumar, Amol Dongre, Anthony C Huszar, Tom Drake, Teng Srey, Pradeep Deshmukh 446 Sok Touch, Richard J Coker 369 Health systems responsiveness and its correlates: Review evidence from family planning service A review of Japanese encephalitis in Uttar provisionin Sri Lanka WHO Pradesh, India W L S P Perera, Lillian Mwanri, Roop Kumari and Pyare L Joshi 374 Rohini de A Seneviratne, Thushara Fernando 457 Original research A study on delay in treatment of kala-azar Challenges faced by skilled birth attendants in patients in Bangladesh providing antenatal and intrapartum care in South-East Asia Syed M Arif, Ariful Basher, Mohammad R selected rural areas of Myanmar Rahman, Mohammad A Faiz, 396 Kyaw Oo, Le Le Win,Saw Saw,Myo Myo Mon, WHO South-East Asia Journal of Public Health Yin Thet Nu Oo, Thae Maung Maung, Factors associated with high prevalence of Su Latt Tun Myint, Theingi Myint 467 pulmonary tuberculosis in HIV-infected people Public health classic Journal of visiting for assessment of eligibility for highly active antiretroviral therapy in Kathmandu, Epidemiologic investigation of excess maternal Nepal and neonatal deaths and evidence-based Bishnu R Tiwari, Surendra Karki, Prakash low-cost public health interventions – Ghimire, Bimala Sharma, Sarala Malla 404 Ignaz Semmelweis: the etiology, concept and prophylaxis of child bed fever Public Health Compliance of off-premise alcohol retailers with S D Gupta, MD 477 the minimum purchase age law Recent WHO Publications 485 Areekul Puangsuwan, Kannapon Phakdeesettakun, Thaksaphon Thamarangsi, Surasak Chaiyasong 412

External quality assessment in blood group serology in the World Health Organization South-East Asia Region Patravee Soisangwan 423

Prevalence of group A genotype human rotavirus among children with diarrhoea in Nepal, 2009–2011 Jeevan B Sherchand, W William Schluter, Jatan B Sherchan, Sarmila Tandukar, Jyoti R Dhakwa, Ganga R Choudhary, Chandeshwar Mahaseth 432

ISSN 2224315-1

World Health House Indraprastha Estate, Mahatma Gandhi Marg, New Delhi-110002, India www.searo.who.int/who-seajph www.searo.who.int/ publications/journals/seajph