Comparison of World Health Organization and Demographic and Health Surveys data to estimate sub-national deworming coverage in pre-school aged children

Abstract

Background: The key metric for monitoring the progress of deworming programs in controlling soil-transmitted helminthiasis (STH) is national drug coverage reported to the World Health Organization (WHO). There is increased interest in utilizing geographically-disaggregated data to estimate sub-national deworming coverage and equity, as well as gender parity. The Demographic and Health Surveys (DHS) offer a potential source of sub-national data. This study aimed to compare deworming coverage routinely reported to WHO and estimated by DHS in pre-school aged children to inform global STH measurement and evaluation.

Methodology: We compared sub-national deworming coverage in pre-school aged children reported to WHO and estimated by DHS aligned geospatially and temporally. We included data from Burundi (2016-2017), Myanmar (2015-2016), and the Philippines (2017) based on data availability. WHO provided data on the date and sub-national coverage per mass drug administration reported by Ministries of Health. DHS included maternally-reported deworming status within the past 6 months for each child surveyed. We estimated differences in sub-national deworming coverage using WHO and DHS data, and performed sensitivity analyses.

Principal findings: We compared data on pre-school aged children from 13 of 18 districts in Burundi (N = 6,835 in DHS), 11 of 15 districts in Myanmar (N = 1,462 in DHS) and 16 of 17 districts in the Philippines (N = 7,594 in DHS) following data exclusion. The national deworming coverages estimated by DHS in Burundi, Myanmar, and the Philippines were 75.5% (95% CI: 73.7%-77.7%), 47.0% (95% CI: 42.7%-51.3%), and 48.0% (95% CI: 46.0%-50.0%), respectively. The national deworming coverages reported by WHO in Burundi, Myanmar, and the Philippines were 80.1%, 93.6% and 75.7%, respectively. The mean absolute differences in district-level coverage reported to WHO and estimated by DHS in Burundi, Myanmar, and the Philippines were 9.5%, 41.5%, and 24.6%, respectively. Across countries, coverage reported to WHO was frequently higher than DHS estimates (32 of 40 districts). National deworming coverage from DHS estimates were similar by gender within countries.

Conclusions and significance: Agreement of deworming coverage reported to WHO and estimated by DHS data was heterogeneous across countries, varying from broadly compatible in Burundi to largely discrepant in Myanmar. DHS data could complement deworming data reported to WHO to improve data monitoring practices and serve as an independent sub-national source of coverage data.

Fig 1. Timeline of national deworming campaigns reported to WHO for pre-school aged children and DHS data collection periods in study countries.

We obtained sub-national deworming coverage data reported to WHO and estimated by DHS across comparable timelines. We used WHO data on programmatic deworming to determine national deworming campaign start dates (dates on left-hand side and time-axis in month/year, e.g. 06/2016 is June 2016) and estimated coverage for Burundi (panel A), Myanmar (panel B) and the Philippines (panel C). We used DHS data from each country to estimate deworming coverage by matching respondents to the corresponding deworming campaigns, based on the survey date and recall period in the orange/gray bars (base, 6-month recall; alternatively, 4- or 8-months based on the sensitivity analysis) that represents the accepted time period (months) after deworming campaign start.

Fig 2. Inclusion and exclusion criteria for DHS and WHO deworming data of pre-school aged children for study countries.

The data inclusions and exclusion are provided for the three study countries: (A)Burundi; (B) Myanmar; and (C) Philippines. Key definitions: N_OBS: Number of pre-school aged children sampled in DHS; N_CAMPAIGN: Number of deworming campaigns reported to WHO; Pre-SAC: Pre-school aged children (12–59 months). In DHS observations, the response of “yes/no” refers to children who were or were not dewormed, while “maybe” refers to mothers who were unable to recall.

Fig 3. District-level sample size of DHS observations used to estimate deworming coverage in study countries.

We determined the sample size of DHS respondents at a district-level and estimated the proportion eligible for one or two deworming campaigns based on the maternal recall period used for DHS observation inclusion. These data are visualized for Burundi (panel A), Myanmar (panel B), and the Philippines (panel C) each with a base case recall period of 6 months based on formulation of the original DHS question for deworming receipt.

Fig 4. Comparison of district-level deworming coverage in pre-school aged children reported to WHO and estimated by DHS data for study countries.

We obtained WHO data on district-level deworming coverage reported by national Ministries of Health based on health facilities data. We estimated district-level deworming coverage from DHS data corresponding to the known WHO deworming campaigns using maternally-reported data on child deworming receipt. The comparison of deworming coverage using WHO and DHS data are shown for a base case recall period of 6-month in Burundi (panel A), Myanmar (panel B) and the Philippines (panel C) based on the DHS question formulation.

Fig 5. Comparison of national-level deworming coverage amongst girls and boys of pre-school age estimated by DHS data in study countries.

We estimated the national-level deworming coverage using maternally-reported child deworming status from DHS stratified by child gender (girls and boys). We compared the estimated coverage by gender using a base recall period of 6 months in Burundi, Myanmar, and the Philippines as based on the DHS question formulation.