Linking data sources for measurement of effective coverage in maternal and newborn health: what do we learn from individual- vs ecological-linking methods?

Exploring approaches to weighting estimates of facility readiness to provide health services used for estimating input-adjusted effective coverage: a case study using data from Tanzania

The ideal approach for calculating effective coverage of health services using ecological linking requires accounting for variability in facility readiness to provide health services and patient volume by incorporating adjustments for facility type into estimates of facility readiness and weighting facility readiness estimates by service-specific caseload. The aim of this study is to compare the ideal caseload-weighted facility readiness approach to two alternative approaches: (1) facility-weighted readiness and (2) observation-weighted readiness to assess the suitability of each as a proxy for caseload-weighted facility readiness. We utilised the 2014-2015 Tanzania Service Provision Assessment along with routine health information system data to calculate facility readiness estimates using the three approaches. We then conducted equivalence testing, using the caseload-weighted estimates as the ideal approach and comparing with the facility-weighted estimates and observation-weighted estimates to test for equivalence. Comparing the facility-weighted readiness estimates to the caseload-weighted readiness estimates, we found that 58% of the estimates met the requirements for equivalence. In addition, the facility-weighted readiness estimates consistently underestimated, by a small percentage, facility readiness as compared to the caseload-weighted readiness estimates. Comparing the observation-weighted readiness estimates to the caseload-weighted readiness estimates, we found that 64% of the estimates met the requirements for equivalence. We found that, in this setting, both facility-weighted readiness and observation-weighted readiness may be reasonable proxies for caseload-weighted readiness. However, in a setting with more variability in facility readiness or larger differences in facility readiness between low caseload and high caseload facilities, the observation-weighted approach would be a better option than the facility-weighted approach. While the methods compared showed equivalence, our results suggest that selecting the best method for weighting readiness estimates will require assessing data availability alongside knowledge of the country context.

Association between antenatal care facility readiness and provision of care at the client level and facility level in five low- and middle-income countries

BACKGROUND

Despite growing interest in monitoring improvements in quality of care, data on service quality in low-income and middle-income countries (LMICs) is limited. While health systems researchers have hypothesized the relationship between facility readiness and provision of care, there have been few attempts to quantify this relationship in LMICs. This study assesses the association between facility readiness and provision of care for antenatal care at the client level and facility level.

METHODS

To assess the association between provision of care and various facility readiness indices for antenatal care, we used multilevel, multivariable random-effects linear regression models. We tested an inflection point on readiness scores by fitting linear spline models. To compare the coefficients between models, we used a bootstrapping approach and calculated the mean difference between all pairwise comparisons. Analyses were conducted at client and facility levels.

RESULTS

Our results showed a small, but significant association between facility readiness and provision of care across countries and most index constructions. The association was most evident in the client-level analyses that had a larger sample size and were adjusted for factors at the facility, health worker, and individual levels. In addition, spline models at a facility readiness score of 50 better fit the data, indicating a plausible threshold effect.

CONCLUSIONS

The results of this study suggest that facility readiness is not a proxy for provision of care, but that there is an important association between facility readiness and provision of care. Data on facility readiness is necessary for understanding the foundations of health systems particularly in countries with the lowest levels of service quality. However, a comprehensive view of quality of care should include both facility readiness and provision of care measures.

Evaluations of effective coverage of maternal and child health services: A systematic review

Conventionally used coverage measures do not reflect the quality of care. Effective coverage (EC) assesses the extent to which health care services deliver potential health gains to the population by integrating concepts of utilization, need and quality. We aimed to conduct a systematic review of studies evaluating EC of maternal and child health services, quality measurement strategies and disparities across wealth quantiles. A systematic search was performed in six electronic databases [MEDLINE, EMBASE, Cumulative Index of Nursing and Allied Health (CINAHL), Scopus, Web of Science and Maternity and Infant Care] and grey literature. We also undertook a hand search of references. We developed search terms having no restrictions based on publication period, country or language. We included studies which reported EC estimates based on the World Health Organization framework of measuring EC. Twenty-seven studies, all from low- and middle-income settings (49 countries), met the criteria and were included in the narrative synthesis of the results. Maternal and child health intervention(s) and programme(s) were assessed either at an individual level or as an aggregated measure of health system performance or both. The EC ranged from 0% for post-partum care to 95% for breastfeeding. When crude coverage measures were adjusted to account for the quality of care, the EC values turned lower. The gap between crude coverage and EC was as high as 86%, and it signified a low quality of care. The assessment of the quality of care addressed structural, process and outcome domains individually or combined. The wealthiest 20% had higher EC of services than the poorest 20%, an inequitable distribution of coverage. More efforts are needed to improve the quality of maternal and child health services and to eliminate the disparities. Moreover, considering multiple dimensions of quality and the use of standard measurements are recommended to monitor coverage effectively.

A rapid systematic review and evidence synthesis of effective coverage measures and cascades for childbirth, newborn and child health in low- and middle-income countries

BACKGROUND

Effective coverage measures aim to estimate the proportion of a population in need of a service that received a positive health outcome. In 2020, the Effective Coverage Think Tank Group recommended using a 'coverage cascade' for maternal, newborn, child and adolescent health and nutrition (MNCAHN), which organises components of effective coverage in a stepwise fashion, with each step accounting for different aspects of quality of care (QoC), applied at the population level. The cascade outlines six steps that increase the likelihood that the population in need experience the intended health benefit: 1) the population in need (target population) who contact a health service; 2) that has the inputs available to deliver the service; 3) who receive the health service; 4) according to quality standards; 5) and adhere to prescribed medication(s) or health workers instructions; and 6) experience the expected health outcome. We examined how effective coverage of life-saving interventions from childbirth to children aged nine has been defined and assessed which steps of the cascade are captured by existing measures.

METHODS

We undertook a rapid systematic review. Seven scientific literature databases were searched covering the period from May 1, 2017 to July, 8 2021. Reference lists from reviews published in 2018 and 2019 were examined to identify studies published prior to May 2017. Eligible studies reported population-level contact coverage measures adjusted for at least one dimension of QoC.

RESULTS

Based on these two search approaches this review includes literature published from 2010 to 2021. From 16 662 records reviewed, 33 studies were included, reporting 64 effective coverage measures. The most frequently examined measures were for childbirth and immediate newborn care (n = 24). No studies examined measures among children aged five to nine years. Definitions of effective coverage varied across studies. Key sources of variability included (i) whether a single effective coverage measure was reported for a package of interventions or separate measures were calculated for each intervention; (ii) the number and type of coverage cascade steps applied to adjust for QoC; and (iii) the individual items included in the effective coverage definition and the methods used to generate a composite quality measure.

CONCLUSION

In the MNCAHN literature there is substantial heterogeneity in both definitions and construction of effective coverage, limiting the comparability of measures over time and place. Current measurement approaches are not closely aligned with the proposed cascade. For widespread adoption, there is a need for greater standardisation of indicator definitions and transparency in reporting, so governments can use these measures to improve investments in MNACHN and implement life-saving health policies and programs.

Operationalising effective coverage measurement of facility based childbirth in Gombe State; a comparison of data sources

Estimating effective coverage of childbirth care requires linking population based data sources to health facility data. For effective coverage to gain widespread adoption there is a need to focus on the feasibility of constructing these measures using data typically available to decision makers in low resource settings. We estimated effective coverage of childbirth care in Gombe State, northeast Nigeria, using two different combinations of facility data sources and examined their strengths and limitations for decision makers. Effective coverage captures information on four steps: access, facility inputs, receipt of interventions and process quality. We linked data from the 2018 Nigerian Demographic and Health Survey (NDHS) to two sources of health facility data: (1) comprehensive health facility survey data generated by a research project; and (2) District Health Information Software 2 (DHIS2). For each combination of data sources, we examined which steps were feasible to calculate, the size of the drop in coverage between steps and the resulting estimate of effective coverage. Analysis included 822 women with a recent live birth, 30% of whom attended a facility for childbirth. Effective coverage was low: 2% based on the project data and less than 1% using the DHIS2. Linking project data with NDHS, it was feasible to measure all four steps; using DHIS2 it was possible to estimate three steps: no data was available to measure process quality. The provision of high quality care is suboptimal in this high mortality setting where access and facility readiness to provide care, crucial foundations to the provision of high quality of care, have not yet been met. This study demonstrates that partial effective coverage measures can be constructed from routine data combined with nationally representative surveys. Advocacy to include process of care indicators in facility summary reports could optimise this data source for decision making.

Measuring facility readiness to provide childbirth care: a comparison of indices using data from a health facility survey in Ethiopia

BACKGROUND

Actionable information about the readiness of health facilities is needed to inform quality improvement efforts in maternity care, but there is no consensus on the best approach to measure readiness. Many countries use the WHO's Service Availability and Readiness Assessment (SARA) or the Demographic and Health Survey (DHS) Programme's Service Provision Assessment to measure facility readiness. This study compares measures of childbirth service readiness based on SARA and DHS guidance to an index based on WHO's quality of maternal and newborn care standards.

METHODS

We used cross-sectional data from Performance Monitoring for Action Ethiopia's 2019 survey of 406 health facilities providing childbirth services. We calculated childbirth service readiness scores using items based on SARA, DHS and WHO standards. For each, we used three aggregation methods for generating indices: simple addition, domain-weighted addition and principal components analysis. We compared central tendency, spread and item variation between the readiness indices; concordance between health facility scores and rankings; and correlations between readiness scores and delivery volume.

RESULTS

Indices showed moderate agreement with one another, and all had a small but significant positive correlation with monthly delivery volume. Ties were more frequent for indices with fewer items. More than two-thirds of items in the relatively shorter SARA and DHS indices were widely (>90%) available in hospitals, and half of the SARA items were widely (>90%) available in health centres/clinics. Items based on the WHO standards showed greater variation and captured unique aspects of readiness (eg, quality improvement processes, actionable information systems) not included in either the SARA or DHS indices.

CONCLUSION

SARA and DHS indices rely on a small set of widely available items to assess facility readiness to provide childbirth care. Expanded selection of items based on the WHO standards can better differentiate between levels of service readiness.

Methodological considerations for linking household and healthcare provider data for estimating effective coverage: a systematic review

OBJECTIVE

To assess existing knowledge related to methodological considerations for linking population-based surveys and health facility data to generate effective coverage estimates. Effective coverage estimates the proportion of individuals in need of an intervention who receive it with sufficient quality to achieve health benefit.

DESIGN

Systematic review of available literature.

DATA SOURCES

Medline, Carolina Population Health Center and Demographic and Health Survey publications and handsearch of related or referenced works of all articles included in full text review. The search included publications from 1 January 2000 to 29 March 2021.

ELIGIBILITY CRITERIA

Publications explicitly evaluating (1) the suitability of data, (2) the implications of the design of existing data sources and (3) the impact of choice of method for combining datasets to obtain linked coverage estimates.

RESULTS

Of 3805 papers reviewed, 70 publications addressed relevant issues. Limited data suggest household surveys can be used to identify sources of care, but their validity in estimating intervention need was variable. Methods for collecting provider data and constructing quality indices were diverse and presented limitations. There was little empirical data supporting an association between structural, process and outcome quality. Few studies addressed the influence of the design of common data sources on linking analyses, including imprecise household geographical information system data, provider sampling design and estimate stability. The most consistent evidence suggested under certain conditions, combining data based on geographical proximity or administrative catchment (ecological linking) produced similar estimates to linking based on the specific provider utilised (exact match linking).

CONCLUSIONS

Linking household and healthcare provider data can leverage existing data sources to generate more informative estimates of intervention coverage and care. However, existing evidence on methods for linking data for effective coverage estimation are variable and numerous methodological questions remain. There is need for additional research to develop evidence-based, standardised best practices for these analyses.

Uterotonics for prevention of postpartum haemorrhage: EN-BIRTH multi-country validation study

BACKGROUND

Postpartum haemorrhage (PPH) is a leading cause of preventable maternal mortality worldwide. The World Health Organization (WHO) recommends uterotonic administration for every woman after birth to prevent PPH. There are no standardised data collected in large-scale measurement platforms. The Every Newborn Birth Indicators Research Tracking in Hospitals (EN-BIRTH) is an observational study to assess the validity of measurement of maternal and newborn indicators, and this paper reports findings regarding measurement of coverage and quality for uterotonics.

METHODS

The EN-BIRTH study took place in five hospitals in Bangladesh, Nepal and Tanzania, from July 2017 to July 2018. Clinical observers collected tablet-based, time-stamped data. We compared observation data for uterotonics to routine hospital register-records and women's report at exit-interview survey. We analysed the coverage and quality gap for timing and dose of administration. The register design was evaluated against gap analyses and qualitative interview data assessing the barriers and enablers to data recording and use.

RESULTS

Observed uterotonic coverage was high in all five hospitals (> 99%, 95% CI 98.7-99.8%). Survey-report underestimated coverage (79.5 to 91.7%). "Don't know" replies varied (2.1 to 14.4%) and were higher after caesarean (3.7 to 59.3%). Overall, there was low accuracy in survey data for details of uterotonic administration (type and timing). Register-recorded coverage varied in four hospitals capturing uterotonics in a specific column (21.6, 64.5, 97.6, 99.4%). The average coverage measurement gap was 18.1% for register-recorded and 6.0% for survey-reported coverage. Uterotonics were given to 15.9% of women within the "right time" (1 min) and 69.8% within 3 min. Women's report of knowing the purpose of uterotonics after birth ranged from 0.4 to 64.9% between hospitals. Enabling register design and adequate staffing were reported to improve routine recording.

CONCLUSIONS

Routine registers have potential to track uterotonic coverage - register data were highly accurate in two EN-BIRTH hospitals, compared to consistently underestimated coverage by survey-report. Although uterotonic coverage was high, there were gaps in observed quality for timing and dose. Standardisation of register design and implementation could improve data quality and data flow from registers into health management information reporting systems, and requires further assessment.

Linking household surveys and facility assessments: a comparison of geospatial methods using nationally representative data from Malawi

BACKGROUND

Linking facility and household surveys through geographic methods is a popular technique to draw conclusions about the relationship between health services and population health outcomes at local levels. These methods are useful tools for measuring effective coverage and tracking progress towards Universal Health Coverage, but are understudied. This paper compares the appropriateness of several geospatial methods used for linking individuals (within displaced survey cluster locations) to their source of family planning (at undisplaced health facilities) at a national level.

METHODS

In Malawi, geographic methods linked a population health survey, rural clusters from the Woman's Questionnaire of the 2015 Malawi Demographic and Health Survey (MDHS 2015), to Malawi's national health facility census to understand the service environment where women receive family planning services. Individuals from MDHS 2015 clusters were linked to health facilities through four geographic methods: (i) closest facility, (ii) buffer (5 km), (iii) administrative boundary, and (iv) a newly described theoretical catchment area method. Results were compared across metrics to assess the number of unlinked clusters (data lost), the number of linkages per cluster (precision of linkage), and the number of women linked to their last source of modern contraceptive (appropriateness of linkage).

RESULTS

The closest facility and administrative boundary methods linked every cluster to at least one facility, while the 5-km buffer method left 288 clusters (35.3%) unlinked. The theoretical catchment area method linked all but one cluster to at least one facility (99.9% linked). Closest facility, 5-km buffer, administrative boundary, and catchment methods linked clusters to 1.0, 1.4, 21.1, and 3.3 facilities on average, respectively. Overall, the closest facility, 5-km buffer, administrative boundary, and catchment methods appropriately linked 64.8%, 51.9%, 97.5%, and 88.9% of women to their last source of modern contraceptive, respectively.

CONCLUSIONS

Of the methods studied, the theoretical catchment area linking method loses a marginal amount of population data, links clusters to a relatively low number of facilities, and maintains a high level of appropriate linkages. This linking method is demonstrated at scale and can be used to link individuals to qualities of their service environments and better understand the pathways through which interventions impact health.

Effective coverage measurement in maternal, newborn, child, and adolescent health and nutrition: progress, future prospects, and implications for quality health systems

Intervention coverage-the proportion of the population with a health-care need who receive care-does not account for intervention quality and potentially overestimates health benefits of services provided to populations. Effective coverage introduces the dimension of quality of care to the measurement of intervention coverage. Many definitions and methodological approaches to measuring effective coverage have been developed, resulting in confusion over definition, calculation, interpretation, and monitoring of these measures. To develop a consensus on the definition and measurement of effective coverage for maternal, newborn, child, and adolescent health and nutrition (MNCAHN), WHO and UNICEF convened a group of experts, the Effective Coverage Think Tank Group, to make recommendations for standardising the definition of effective coverage, measurement approaches for effective coverage, indicators of effective coverage in MNCAHN, and to develop future effective coverage research priorities. Via a series of consultations, the group recommended that effective coverage be defined as the proportion of a population in need of a service that resulted in a positive health outcome from the service. The proposed effective coverage measures and care cascade steps can be applied to further develop effective coverage measures across a broad range of MNCAHN services. Furthermore, advances in measurement of effective coverage could improve monitoring efforts towards the achievement of universal health coverage.

What the percentage of births in facilities does not measure: readiness for emergency obstetric care and referral in Senegal

Introduction

Increases in facility deliveries in sub-Saharan Africa have not yielded expected declines in maternal mortality, raising concerns about the quality of care provided in facilities. The readiness of facilities at different health system levels to provide both emergency obstetric and newborn care (EmONC) as well as referral is unknown. We describe this combined readiness by facility level and region in Senegal.

Methods

For this cross-sectional study, we used data from nine Demographic and Health Surveys between 1992 and 2017 in Senegal to describe trends in location of births over time. We used data from the 2017 Service Provision Assessment to describe EmONC and emergency referral readiness across facility levels in the public system, where 94% of facility births occur. A national global positioning system facility census was used to map access from lower-level facilities to the nearest facility performing caesareans.

Results

Births in facilities increased from 47% in 1992 to 80% in 2016, driven by births in lower-level health posts, where half of facility births now occur. Caesarean rates in rural areas more than doubled but only to 3.7%, indicating minor improvements in EmONC access. Only 9% of health posts had full readiness for basic EmONC, and 62% had adequate referral readiness (vehicle on-site or telephone and vehicle access elsewhere). Although public facilities accounted for three-quarters of all births in 2016, only 16% of such births occurred in facilities able to provide adequate combined readiness for EmONC and referral.

Conclusions

Our findings imply that many lower-level public facilities—the most common place of birth in Senegal—are unable to treat or refer women with obstetric complications, especially in rural areas. In light of rising lower-level facility births in Senegal and elsewhere, improvements in EmONC and referral readiness are urgently needed to accelerate reductions in maternal and perinatal mortality.

"What gets measured better gets done better": The landscape of validation of global maternal and newborn health indicators through key informant interviews

Background

A large number of indicators are currently used to monitor the state of maternal and newborn health, including those capturing dimensions of health system and input, care access and availability, care quality and safety, coverage and outcomes, and impact. Validity of these indicators is a key issue in the process of assessing indicator performance and suitability. This paper aims to understand the meaning of indicator validity in the field of maternal and newborn health, and to identify key recommendations for future research.

Methods

This qualitative study used purposive sampling to identify key informants until thematic saturation was achieved. We interviewed 32 respondents from a variety of backgrounds using semi-structured interviews covering five themes: the meaning of indicator validity, methodological approaches to assessing validity, acceptable levels of indicator validity, gaps in validation research, and recommendations for addressing these gaps. Interview transcripts were analysed data using thematic content approach.

Results

Three conceptually different definitions of indicator validity were described by respondents. They considered indicator validity to encompass meaning and potential to spur action, going beyond diagnostic validity. Indicator validation was seen as an ongoing process of building and synthesising a wide range of evidence rather than a one-size-fits-all cut-off in diagnostic validity tests. Gaps identified included assessing validity of indicators of quality of care and indicators based on facility-level data, as well as expanding studies to a broader range of global settings. The key recommendation was to develop a coordinated approach to summarising and evaluating research on indicator validity, including capacity building in appraising and communicating the available evidence for country-specific needs.

Conclusion

The findings will inform future recommendations around indicator testing and validation.

Effective coverage of facility delivery in Bangladesh, Haiti, Malawi, Nepal, Senegal, and Tanzania

Background

The persistence of preventable maternal and newborn deaths highlights the importance of quality of care as an essential element in coverage interventions. Moving beyond the conventional measurement of crude coverage, we estimated effective coverage of facility delivery by adjusting for facility preparedness to provide delivery services in Bangladesh, Haiti, Malawi, Nepal, Senegal, and Tanzania.

Methods

The study uses data from Demographic and Health Surveys (DHS) and Service Provision Assessments (SPA) in Bangladesh (2014 DHS and 2014 SPA), Haiti (2012 DHS and 2013 SPA), Malawi (2015–16 DHS and 2013–14 SPA), Nepal (2016 DHS and 2015 SPA), Senegal (2016 DHS and 2015 SPA), and Tanzania (2015–16 DHS and 2014–15 SPA). We defined effective coverage as the mathematical product of crude coverage and quality of care. The coverage of facility delivery was measured with DHS data and quality of care was measured with facility data from SPA. We estimated effective coverage at both the regional and the national level and accounted for type of facility where delivery care was sought.

Findings

The findings from the six countries indicate the effective coverage ranges from 24% in Haiti to 66% in Malawi, representing substantial reductions (20% to 39%) from crude coverage rates. Although Malawi has achieved almost universal coverage of facility delivery (93%), effective coverage was only 66%.vSuch gaps between the crude coverage and the effective coverage suggest that women delivered in health facility but did not necessarily receive an adequate quality of care. In all countries except Malawi, effective coverage differed substantially among the country’s regions of the country, primarily due to regional variability in coverage.

Interpretation

Our findings reinforce the importance of quality of obstetric and newborn care to achieve further reduction of maternal and newborn mortality. Continued efforts are needed to increase the use of facility delivery service in countries or regions where coverage remains low.

Scoping review to identify and map the health personnel considered skilled birth attendants in low-and-middle income countries from 2000-2015

INTRODUCTION

The "percentage of births attended by a skilled birth attendant" (SBA) is an indicator that has been adopted by several global monitoring frameworks, including the Sustainable Development Goal (SDG) agenda for regular monitoring as part of target 3.1 for reducing maternal mortality by 2030. However, accurate and consistent measurement is challenged by contextual differences between and within countries on the definition of SBA, including the education, training, competencies, and functions they are qualified to perform. This scoping review identifies and maps the health personnel considered SBA in low-to-middle-income-countries (LMIC).

METHODS AND ANALYSIS

A search was conducted inclusive to the years 2000 to 2015 in PubMed/MEDLINE, EMBASE, CINAHL Complete, Cochrane Database of Systematic Reviews, POPLINE and the World Health Organization Global Index Medicus. Original primary source research conducted in LMIC that evaluated the skilled health personnel providing interventions during labour and childbirth were considered for inclusion. All studies reported disaggregated data of SBA cadres and were disaggregated by country.

RESULTS

The search of electronic databases identified a total of 23,743 articles. Overall, 70 articles were included in the narrative synthesis. A total of 102 unique cadres names were identified from 36 LMIC countries. Of the cadres included, 16% represented doctors, 16% were nurses, and 15% were midwives. We found substantial heterogeneity between and within countries on the reported definition of SBA and the education, training, skills and competencies that they were able to perform.

CONCLUSION

The uncertainty and diversity of reported qualifications and competency of SBA within and between countries requires attention in order to better ascertain strategic priorities for future health system planning, including training and education. These results can inform recommendations around improved coverage measurement and monitoring of SBA moving forward, allowing for more accurate, consistent, and timely data able to guide decisions and action around planning and implementation of maternal and newborn health programmes.

Linking household survey and health facility data for effective coverage measures: a comparison of ecological and individual linking methods using the Multiple Indicator Cluster Survey in Côte d'Ivoire

Background

Population-based measures of intervention coverage are used in low- and middle-income countries for program planning, prioritization, and evaluation. There is increased interest in effective coverage, which integrates information about service quality or health outcomes. Approaches proposed for quality-adjusted effective coverage include linking data on need and service contact from population-based surveys with data on service quality from health facility surveys. However, there is limited evidence about the validity of different linking methods for effective coverage estimation.

Methods

We collaborated with the 2016 Côte d'Ivoire Multiple Indicator Cluster Survey (MICS) to link data from a health provider assessment to care-seeking data collected by the MICS in the Savanes region of Côte d'Ivoire. The provider assessment was conducted in a census of public and non-public health facilities and pharmacies in Savanes in May-June 2016. We also included community health workers managing sick children who served the clusters sampled for the MICS. The provider assessment collected information on structural and process quality for antenatal care, delivery and immediate newborn care, postnatal care, and sick child care. We linked the MICS and provider data using exact-match and ecological linking methods, including aggregate linking and geolinking methods. We compared the results obtained from exact-match and ecological methods.

Results

We linked 731 of 786 care-seeking episodes (93%) from the MICS to a structural quality score for the provider named by the respondent. Effective coverage estimates computed using exact-match methods were 13%-63% lower than the care-seeking estimates from the MICS. Absolute differences between exact match and ecological linking methods were ±7 percentage points for all ecological methods. Incorporating adjustments for provider category and weighting by service-specific utilization into the ecological methods generally resulted in better agreement between ecological and exact match estimates.

Conclusions

Ecological linking may be a feasible and valid approach for estimating quality-adjusted effective coverage when a census of providers is used. Adjusting for provider type and caseload may improve agreement with exact match results. There remain methodological questions to be addressed to develop guidance on using linking methods for estimating quality-adjusted effective coverage, including the effect of facility sampling and time displacement.

A prospective validation study in South-West Nigeria on caregiver report of childhood pneumonia and antibiotic treatment using Demographic and Health Survey (DHS) and Multiple Indicator Cluster Survey (MICS) questions

Background

Childhood pneumonia is the single largest infectious cause of death in children under five worldwide. Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys (MICS) provide health information on care sought for sick children in resource poor settings. Despite not being primarily designed to identify childhood pneumonia, there are concerns that reported episodes of "symptoms of acute respiratory infection" in DHS and MICS are often interpreted by other groups as a "proxy" for childhood pneumonia. Using DHS5 and MICS5 survey tools, this study aimed to assess how accurately caregivers report of "symptoms of acute respiratory infection" reflect pneumonia episodes and antibiotic use in children under five.

Methods

Children aged 0 to 59 months presenting with cough and/or difficult breathing were recruited from four study hospitals in Ibadan, Nigeria from August 2015 to March 2017. Children were assessed using World Health Organization (WHO) standard criteria by study physicians to identify whether they had pneumonia. Three hundred and two matched children in each category of 'pneumonia' and "no pneumonia" were followed up at home, either two or eight weeks later, using questions from DHS5 and MICS5 surveys to assess the accuracy of caregiver recall of pneumonia.

Results

The specificity of DHS5 and MICS5 questions for identifying childhood pneumonia were 87.4 (95% confidence interval (CI) = 83.1-91.0) and 86.1 (95% CI = 81.7-89.8) respectively and the sensitivity of questions were 37.1 (95% CI = 31.6-42.8) and 37.1 (95% CI = 31.6-42.8). Correct recall of antibiotic treatment was poor (kappa statistic = 0.064) but improved with the use of medicine pill boards (kappa statistic = 0.235).

Conclusions

DHS5 and MICS5 survey questions are not designed to identify childhood pneumonia and this study confirms that they do not accurately discern episodes of childhood pneumonia from cough/cold in children under five. The proportion of pneumonia episodes appropriately treated with antibiotics cannot be accurately assessed using current DHS and MICS surveys. If these results are used to guide programmatic decisions, it is likely to encourage overuse and inappropriate prescribing of antibiotics for episodes of cough/cold. International agencies who continue to use these household data to monitor the proportion of children with pneumonia who receive antibiotic treatment should be discouraged from doing this as these data are likely to mislead national and global programmes. Medicine pill boards are used in a number of DHS surveys and should be promoted for wider use in national population surveys to improve the accuracy of antibiotic recall.