An in-depth report of quality control on kato-katz and data entry in four clinical trials evaluating the efficacy of albendazole against soil-transmitted helminth infections

Background Efforts to control soil-transmitted helminth (STH) infections have intensified over the past decade. Field-survey data on STH prevalence, infection intensity and drug efficacy is necessary to guide the implementation of control programs and should be of the best possible quality. Methodology During four clinical trials designed to evaluate the efficacy of albendazole against STHs in Brazil, Ethiopia, Lao PDR and Tanzania, quality control (QC) was performed on the dupli-cate Kato-Katz thick smears and the data entry. We analyzed datasets following QC on both fecal egg counts (FECs) and data entry, and compared the prevalence of any STH infection and moderate-to-heavy intensity (MHI) infections and the drug efficacy against STH infections. Results Across the four study sites, a total of 450 out of 4,830 (9.3%) Kato-Katz thick smears were re-examined. Discrepancies in FECs varied from ~3% (hookworms) to ~6.5% (Ascaris lum-bricoides and Trichuris trichiura). The difference in STH prevalence and prevalence of MHI infections using the datasets with and without QC of the FECs did not exceed 0.3%, except for hookworm infections in Tanzania, where we noted a 2.2 percentage point increase in MHI infections (pre-QC: 1.6% vs. post-QC: 3.8%). There was a 100% agreement in the classification of drug efficacy of albendazole against STH between the two datasets. In total, 201 of the 28,980 (0.65%) data entries that were made to digitize the FECs were different between both data-entry clerks. Nevertheless, the overall prevalence of STH, the prevalence of MHI infections and the classification of drug efficacy remained largely unaffected. Conclusion/significance In these trials, where staff was informed that QC would take place, minimal changes in study outcomes were reported following QC on FECs or data entry. Nevertheless, imposing QC did reduce the number of errors. Therefore, application of QC together with proper train-ing of the personnel and the availability of clear standard operating procedures is expected to support higher data quality.