Estimation of population coverage for antibiotic distribution for trachoma control: a comparison of methods

Trachoma control includes mass drug administration (MDA) with antibiotics targeting coverage of at least 80%. Coverage is traditionally calculated by dividing doses distributed by population estimate, which is unreliable. We compared a verifiable coverage assessment method against self-reported participation and the traditional calculation, and examined factors associated with MDA participation in Akobo County, Southern Sudan. During MDA, recipients were marked with indelible ink and followed-up using a two-stage household survey: 25 clusters from three districts, and 10 households per cluster. All household members were enumerated; asked about self-reported participation and observed for indelible marks. Household heads were interviewed to assess factors associated with MDA. Overall 11 419 treatments were given and 1358 residents from 247 households surveyed. By traditional methods MDA coverage was 20.9% (95% CI: 20.6-21.3); 61.5% (95 % CI: 49.4-73.6) by self-reporting; and 37.5% (95% CI: 25.1%-49.9%) from observed ink marks. Controlling for other factors, presence of a health worker (OR 2.3, 95% CI: 1.5-3.6); head of household knowledge of azithromycin (OR 1.6, 95% CI: 1.5-3.1); and head of household advance notice of MDA (OR 34.9, 18.1-66.3) were independent predictors of participation. Cluster randomised surveys can estimate MDA coverage better than the traditional method and implementation of indelible marking merits investigation.