Reproducibility of the mark-resight method to assess vaccination coverage in free-roaming dogs

The mark-resight methodology is a widely used tool to assess both population size and vaccination coverage of free-roaming dogs. Evaluating coverage is often critical to determine the progress and impact of vaccination interventions. The aim of this study was to assess the reproducibility of the mark-resight method during a rabies vaccination campaign. A total of 90 areas in Goa, India, were surveyed on two occasions following a vaccination campaign, during which vaccinated dogs were marked. The estimated vaccination coverage for both confined and free-roaming dogs in the study area after a first round of vaccination was 68.9% (95% CI, 66.7-71.0%). The number of dogs counted in each survey, as well as the proportion of marked dogs, were the main outcomes recorded. Linear and mixed-effects models were used to evaluate the effect of survey order, surveyor, time of day, days between surveys, and weather on the outcomes between repeated surveys. Fewer dogs were likely to be counted in afternoon surveys compared to morning surveys (19.2% fewer dogs, 95%CI 7.5-31.9%) and a 17.2% (95%CI 3.6-32.7%) higher proportion of marked dogs was associated with the repeated survey compared to the first survey. Increased variation of the marked percentage between surveys was associated with using different surveyors (8.3%, 95%CI 0.4-16.2). This study demonstrated that external or personnel factors can influence the number of dogs counted and the estimated vaccination coverage of individual survey areas. Regular staff training, to ensure equal effort and quality of survey techniques between surveyors, could help standardise survey outcomes.