Integrated Management of Childhood Illness (IMCI) in Bangladesh: Early findings from a cluster-randomised study

Shams El Arifeen; Lauren S. Blum; D. M. Emdadul Hoque; Enayet K. Chowdhury; Rasheda Khan; Prof Robert E. Black; Prof Cesar G. Victora; Jennifer Bryce

doi:10.1016/S0140-6736(04)17312-1

Integrated Management of Childhood Illness (IMCI) in Bangladesh: Early findings from a cluster-randomised study

Shams El Arifeen, Lauren S. Blum, D. M. Emdadul Hoque, Enayet K. Chowdhury, Rasheda Khan, Prof Robert E. Black, Prof Cesar G. Victora, Jennifer Bryce

Bloomberg School of Public Health

Research output: Contribution to journal › Article › peer-review

Abstract

Background We report the preliminary findings from a continuing cluster randomised evaluation of the Integrated Management of Childhood Illness (IMCI) strategy in Bangladesh. Methods 20 first-level outpatient facilities in the Matlab sub-district and their catchment areas were randomised to either IMCI or standard care. Surveys were done in households and in health facilities at baseline and were repeated about 2 years after implementation. Data on use of health facilities were recorded. IMCI implementation included health worker training, health systems support, and community level activities guided by formative research. Findings 94% of health workers in the intervention facilities were trained in IMCI. Health systems supports were generally available, but implementation of the community activities was slow. The mean index of correct treatment for sick children was 54 in IMCI facilities compared with 9 in comparison facilities (range 0-100). Use of the IMCI facilities increased from 0·6 visits per child per year at baseline to 1·9 visits per child per year about 21 months after IMCI introduction. 19% of sick children in the IMCI area were taken to a health worker compared with 9% in the non-IMCI area. Interpretation 2 years into the assessment, the results show improvements in the quality of care in health facilities, increases in use of facilities, and gains in the proportion of sick children taken to an appropriate health care provider. These findings are being used to strengthen child health care nationwide. They suggest that low levels of use of health facilities could be improved by investing in quality of care and health systems support.

Original language	English (US)
Pages (from-to)	1595-1602
Number of pages	8
Journal	Lancet
Volume	364
Issue number	9445
DOIs	https://doi.org/10.1016/S0140-6736(04)17312-1
State	Published - Oct 30 2004

ASJC Scopus subject areas

Medicine(all)

Access to Document

10.1016/S0140-6736(04)17312-1

Cite this

Integrated Management of Childhood Illness (IMCI) in Bangladesh : Early findings from a cluster-randomised study. / Arifeen, Shams El; Blum, Lauren S.; Emdadul Hoque, D. M.; Chowdhury, Enayet K.; Khan, Rasheda ; Black, Prof Robert E.; Victora, Prof Cesar G.; Bryce, Jennifer.

In: Lancet, Vol. 364, No. 9445, 30.10.2004, p. 1595-1602.

Research output: Contribution to journal › Article › peer-review

@article{24c656d4687446c7afbfdffd6682bcf5,

title = "Integrated Management of Childhood Illness (IMCI) in Bangladesh: Early findings from a cluster-randomised study",

abstract = "Background We report the preliminary findings from a continuing cluster randomised evaluation of the Integrated Management of Childhood Illness (IMCI) strategy in Bangladesh. Methods 20 first-level outpatient facilities in the Matlab sub-district and their catchment areas were randomised to either IMCI or standard care. Surveys were done in households and in health facilities at baseline and were repeated about 2 years after implementation. Data on use of health facilities were recorded. IMCI implementation included health worker training, health systems support, and community level activities guided by formative research. Findings 94% of health workers in the intervention facilities were trained in IMCI. Health systems supports were generally available, but implementation of the community activities was slow. The mean index of correct treatment for sick children was 54 in IMCI facilities compared with 9 in comparison facilities (range 0-100). Use of the IMCI facilities increased from 0·6 visits per child per year at baseline to 1·9 visits per child per year about 21 months after IMCI introduction. 19% of sick children in the IMCI area were taken to a health worker compared with 9% in the non-IMCI area. Interpretation 2 years into the assessment, the results show improvements in the quality of care in health facilities, increases in use of facilities, and gains in the proportion of sick children taken to an appropriate health care provider. These findings are being used to strengthen child health care nationwide. They suggest that low levels of use of health facilities could be improved by investing in quality of care and health systems support.",

author = "Arifeen, {Shams El} and Blum, {Lauren S.} and {Emdadul Hoque}, {D. M.} and Chowdhury, {Enayet K.} and Rasheda Khan and Black, {Prof Robert E.} and Victora, {Prof Cesar G.} and Jennifer Bryce",

note = "Funding Information: Our results show that the introduction of IMCI is associated with improvements in the quality of health care for children in first-level facilities, a more than threefold increase in use of first-level facilities for the care of sick children, and steady increases in the proportion of sick children reported by their caretakers to have been taken to a health facility or health worker for care. Differences between the IMCI intervention and comparison areas were even more pronounced for children assessed by health workers or perceived by caretakers to have severe illness. No differences in careseeking practices were noted between the sexes, despite other reports of gender inequalities in Bangladesh for many child health indicators. 16–20 It should be noted that the timing of the data for use of health facilities presented here does not fully capture the impact of the community intervention, which started in earnest in late 2003. Consequently, we expect that use of these facilities will continue to increase as the community interventions reach higher levels of coverage. These positive effects of IMCI are tempered by other findings reported here. Qualitative sub-studies showed low rates of referral completion among children with severe illness sent to local hospitals for care. Despite increases in use of health facilities and improvements in careseeking, the estimated number of visits by under-5s to first-level facilities for illness care in this area of Bangladesh remained at about two per child per year, and only 19% of children reported by their caretakers to have been ill in the 2 weeks before the survey were taken to an appropriate care provider. Even the careseeking rate documented in the most recent round of monitoring was substantially lower than that observed in the Tanzania multi-country evaluation site, where 41% of children who were sick in the previous 2weeks were taken first to trained providers at baseline. 21 However, in Tanzania the improved quality of care did not result in increased use of health facilities, perhaps because baseline rates of use were much higher in Tanzania than in Bangladesh. Higher proportions of sick children in the study area and in Bangladesh must receive basic but effective treatments for common life-threatening illnesses if rapid reductions in mortality rates are to be achieved. This study has limitations characteristic of large-scale assessments of public health programmes. 22 In these studies interventions are made available to the health services, and although their delivery or rates of compliance are not under the direct control of the investigators the results are likely to show a positive bias, or “best practice”. This is true of the way IMCI is being implemented in the study intervention areas in Bangladesh, although all aspects of implementation reflect agreement between the study team and the Government of Bangladesh that the set of interventions and delivery strategies being assessed fall well within the policies and resource availability that would determine how IMCI was scaled-up to nationwide coverage. A second limitation derives from the fact that these results are reported before the larger study is complete. The evaluation design is prospective, and full results on mortality effects and cost effectiveness will be available only in 2007. This delay shows the time needed after implementation to allow IMCI to have a biological effect, as well as the time needed for measurement of the final indicators of effect. 4 However, in view of the importance to public health and policy of these interim findings on the quality of care, use, and careseeking, it is important that these results be made available now—especially since another study within the multi-country evaluation has shown a plausible association between IMCI case management training and child mortality and nutritional status. 23 This study highlights the importance of qualitative research and monitoring to ensure continued improvement of interventions and delivery strategies. The active collaboration between researchers and the GOB permitted flexibility in our setting, leading to implementation strategies that evolved over time in response to identified problem areas. The implication for country teams implementing IMCI is that they must have the authority and understanding to deviate from current guidelines by identifying critical problem areas, and the creativity to design approaches appropriate to the local social and health context. Our experience to date suggests that full implementation of the IMCI strategy, with interventions directed at improvement of health worker skills, health system support for child health care, and family and community practices, is feasible and can lead to changes in careseeking practices and increases in the use of public health facilities. Achieving and expanding on this success, however, requires full and active collaboration among multidisciplinary teams of scientists and government health decision makers and a willingness to improve key elements of the health system. In the context of this study, existing supervisory staff were trained and supported to provide more frequent supervision and to incorporate activities into each visit that targeted quality of care. The standard information system forms used by the Government of Bangladesh were modified to further reinforce correct functioning in health workers Essential drugs and equipment needed to provide quality health care to children are being provided and maintained. Levels of intervention coverage are high, if not universal, and sustained over the study period. The result has been improved health care for all children. The study team continues to work closely with the Government of Bangladesh to incorporate lessons and experiences from the multi-country evaluation study into nationwide implementation of IMCI. Tools and methods developed and implemented in Matlab form an integral part of the IMCI intervention already implemented in 21 of 460 upazillas in Bangladesh by the Government of Bangladesh. Contributors S E Arifeen, R E Black, C G Victora, and J Bryce contributed to the study conception and design. S E Arifeen coordinated fieldwork and supervised intervention implementation and field data collection with LBlum, D M E Hoque, E K Chowdhury, and R Khan. L Blum and RKhan were responsible for the design, data collection, analysis, and write-up of the formative research. D M E Hoque contributed to the intervention description. S E Arifeen and E K Chowdhury assessed data. S E Arifeen, L Blum, and J Bryce wrote the first draft of the article. All authors critically revised the first draft for content and contributed to the final draft. Conflict of interest statement Cesar G Victora and Jennifer Bryce work as part-time consultants for the WHO, one of the institutions involved in implementing IMCI worldwide. No other conflict of interest declared for any other author. Acknowledgments We thank our colleagues and partners in the Government of Bangladesh for their help in designing and implementing the intervention, and the managers, health and family planning workers, and data collectors of Matlab for responding to the needs of the study and making possible a near-to-ideal implementation of IMCI. This study was done at the ICDDR,B: Centre for Health and Population Research, with funding from the Bill and Melinda Gates Foundation through a grant to the WHO Department of Child and Adolescent Health and Development and of Cooperative Agreement #388-A-00–97–00032–00 from the United States Agency for International Development. ICDDR,B acknowledges with gratitude their commitment to the Centre's research effort. This paper is part of the Multi-Country Evaluation of IMCI Effectiveness, Cost and Impact, which is arranged, coordinated, and funded by the Department of Child and Adolescent Health and Development of the WHO, and with the financial support of the Bill and Melinda Gates Foundation and the US Agency for International Development. Copyright: Copyright 2021 Elsevier B.V., All rights reserved.",

year = "2004",

month = oct,

day = "30",

doi = "10.1016/S0140-6736(04)17312-1",

language = "English (US)",

volume = "364",

pages = "1595--1602",

journal = "The Lancet",

issn = "0140-6736",

publisher = "Elsevier Limited",

number = "9445",

}

TY - JOUR

T1 - Integrated Management of Childhood Illness (IMCI) in Bangladesh

T2 - Early findings from a cluster-randomised study

AU - Arifeen, Shams El

AU - Blum, Lauren S.

AU - Emdadul Hoque, D. M.

AU - Chowdhury, Enayet K.

AU - Khan, Rasheda

AU - Black, Prof Robert E.

AU - Victora, Prof Cesar G.

AU - Bryce, Jennifer

N1 - Funding Information: Our results show that the introduction of IMCI is associated with improvements in the quality of health care for children in first-level facilities, a more than threefold increase in use of first-level facilities for the care of sick children, and steady increases in the proportion of sick children reported by their caretakers to have been taken to a health facility or health worker for care. Differences between the IMCI intervention and comparison areas were even more pronounced for children assessed by health workers or perceived by caretakers to have severe illness. No differences in careseeking practices were noted between the sexes, despite other reports of gender inequalities in Bangladesh for many child health indicators. 16–20 It should be noted that the timing of the data for use of health facilities presented here does not fully capture the impact of the community intervention, which started in earnest in late 2003. Consequently, we expect that use of these facilities will continue to increase as the community interventions reach higher levels of coverage. These positive effects of IMCI are tempered by other findings reported here. Qualitative sub-studies showed low rates of referral completion among children with severe illness sent to local hospitals for care. Despite increases in use of health facilities and improvements in careseeking, the estimated number of visits by under-5s to first-level facilities for illness care in this area of Bangladesh remained at about two per child per year, and only 19% of children reported by their caretakers to have been ill in the 2 weeks before the survey were taken to an appropriate care provider. Even the careseeking rate documented in the most recent round of monitoring was substantially lower than that observed in the Tanzania multi-country evaluation site, where 41% of children who were sick in the previous 2weeks were taken first to trained providers at baseline. 21 However, in Tanzania the improved quality of care did not result in increased use of health facilities, perhaps because baseline rates of use were much higher in Tanzania than in Bangladesh. Higher proportions of sick children in the study area and in Bangladesh must receive basic but effective treatments for common life-threatening illnesses if rapid reductions in mortality rates are to be achieved. This study has limitations characteristic of large-scale assessments of public health programmes. 22 In these studies interventions are made available to the health services, and although their delivery or rates of compliance are not under the direct control of the investigators the results are likely to show a positive bias, or “best practice”. This is true of the way IMCI is being implemented in the study intervention areas in Bangladesh, although all aspects of implementation reflect agreement between the study team and the Government of Bangladesh that the set of interventions and delivery strategies being assessed fall well within the policies and resource availability that would determine how IMCI was scaled-up to nationwide coverage. A second limitation derives from the fact that these results are reported before the larger study is complete. The evaluation design is prospective, and full results on mortality effects and cost effectiveness will be available only in 2007. This delay shows the time needed after implementation to allow IMCI to have a biological effect, as well as the time needed for measurement of the final indicators of effect. 4 However, in view of the importance to public health and policy of these interim findings on the quality of care, use, and careseeking, it is important that these results be made available now—especially since another study within the multi-country evaluation has shown a plausible association between IMCI case management training and child mortality and nutritional status. 23 This study highlights the importance of qualitative research and monitoring to ensure continued improvement of interventions and delivery strategies. The active collaboration between researchers and the GOB permitted flexibility in our setting, leading to implementation strategies that evolved over time in response to identified problem areas. The implication for country teams implementing IMCI is that they must have the authority and understanding to deviate from current guidelines by identifying critical problem areas, and the creativity to design approaches appropriate to the local social and health context. Our experience to date suggests that full implementation of the IMCI strategy, with interventions directed at improvement of health worker skills, health system support for child health care, and family and community practices, is feasible and can lead to changes in careseeking practices and increases in the use of public health facilities. Achieving and expanding on this success, however, requires full and active collaboration among multidisciplinary teams of scientists and government health decision makers and a willingness to improve key elements of the health system. In the context of this study, existing supervisory staff were trained and supported to provide more frequent supervision and to incorporate activities into each visit that targeted quality of care. The standard information system forms used by the Government of Bangladesh were modified to further reinforce correct functioning in health workers Essential drugs and equipment needed to provide quality health care to children are being provided and maintained. Levels of intervention coverage are high, if not universal, and sustained over the study period. The result has been improved health care for all children. The study team continues to work closely with the Government of Bangladesh to incorporate lessons and experiences from the multi-country evaluation study into nationwide implementation of IMCI. Tools and methods developed and implemented in Matlab form an integral part of the IMCI intervention already implemented in 21 of 460 upazillas in Bangladesh by the Government of Bangladesh. Contributors S E Arifeen, R E Black, C G Victora, and J Bryce contributed to the study conception and design. S E Arifeen coordinated fieldwork and supervised intervention implementation and field data collection with LBlum, D M E Hoque, E K Chowdhury, and R Khan. L Blum and RKhan were responsible for the design, data collection, analysis, and write-up of the formative research. D M E Hoque contributed to the intervention description. S E Arifeen and E K Chowdhury assessed data. S E Arifeen, L Blum, and J Bryce wrote the first draft of the article. All authors critically revised the first draft for content and contributed to the final draft. Conflict of interest statement Cesar G Victora and Jennifer Bryce work as part-time consultants for the WHO, one of the institutions involved in implementing IMCI worldwide. No other conflict of interest declared for any other author. Acknowledgments We thank our colleagues and partners in the Government of Bangladesh for their help in designing and implementing the intervention, and the managers, health and family planning workers, and data collectors of Matlab for responding to the needs of the study and making possible a near-to-ideal implementation of IMCI. This study was done at the ICDDR,B: Centre for Health and Population Research, with funding from the Bill and Melinda Gates Foundation through a grant to the WHO Department of Child and Adolescent Health and Development and of Cooperative Agreement #388-A-00–97–00032–00 from the United States Agency for International Development. ICDDR,B acknowledges with gratitude their commitment to the Centre's research effort. This paper is part of the Multi-Country Evaluation of IMCI Effectiveness, Cost and Impact, which is arranged, coordinated, and funded by the Department of Child and Adolescent Health and Development of the WHO, and with the financial support of the Bill and Melinda Gates Foundation and the US Agency for International Development. Copyright: Copyright 2021 Elsevier B.V., All rights reserved.

PY - 2004/10/30

Y1 - 2004/10/30

N2 - Background We report the preliminary findings from a continuing cluster randomised evaluation of the Integrated Management of Childhood Illness (IMCI) strategy in Bangladesh. Methods 20 first-level outpatient facilities in the Matlab sub-district and their catchment areas were randomised to either IMCI or standard care. Surveys were done in households and in health facilities at baseline and were repeated about 2 years after implementation. Data on use of health facilities were recorded. IMCI implementation included health worker training, health systems support, and community level activities guided by formative research. Findings 94% of health workers in the intervention facilities were trained in IMCI. Health systems supports were generally available, but implementation of the community activities was slow. The mean index of correct treatment for sick children was 54 in IMCI facilities compared with 9 in comparison facilities (range 0-100). Use of the IMCI facilities increased from 0·6 visits per child per year at baseline to 1·9 visits per child per year about 21 months after IMCI introduction. 19% of sick children in the IMCI area were taken to a health worker compared with 9% in the non-IMCI area. Interpretation 2 years into the assessment, the results show improvements in the quality of care in health facilities, increases in use of facilities, and gains in the proportion of sick children taken to an appropriate health care provider. These findings are being used to strengthen child health care nationwide. They suggest that low levels of use of health facilities could be improved by investing in quality of care and health systems support.

AB - Background We report the preliminary findings from a continuing cluster randomised evaluation of the Integrated Management of Childhood Illness (IMCI) strategy in Bangladesh. Methods 20 first-level outpatient facilities in the Matlab sub-district and their catchment areas were randomised to either IMCI or standard care. Surveys were done in households and in health facilities at baseline and were repeated about 2 years after implementation. Data on use of health facilities were recorded. IMCI implementation included health worker training, health systems support, and community level activities guided by formative research. Findings 94% of health workers in the intervention facilities were trained in IMCI. Health systems supports were generally available, but implementation of the community activities was slow. The mean index of correct treatment for sick children was 54 in IMCI facilities compared with 9 in comparison facilities (range 0-100). Use of the IMCI facilities increased from 0·6 visits per child per year at baseline to 1·9 visits per child per year about 21 months after IMCI introduction. 19% of sick children in the IMCI area were taken to a health worker compared with 9% in the non-IMCI area. Interpretation 2 years into the assessment, the results show improvements in the quality of care in health facilities, increases in use of facilities, and gains in the proportion of sick children taken to an appropriate health care provider. These findings are being used to strengthen child health care nationwide. They suggest that low levels of use of health facilities could be improved by investing in quality of care and health systems support.

UR - http://www.scopus.com/inward/record.url?scp=7444255554&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=7444255554&partnerID=8YFLogxK

U2 - 10.1016/S0140-6736(04)17312-1

DO - 10.1016/S0140-6736(04)17312-1

M3 - Article

C2 - 15519629

AN - SCOPUS:7444255554

VL - 364

SP - 1595

EP - 1602

JO - The Lancet

JF - The Lancet

SN - 0140-6736

IS - 9445

ER -

Integrated Management of Childhood Illness (IMCI) in Bangladesh: Early findings from a cluster-randomised study

Abstract

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this