Data management for large community trials in Nepal.

The current accessibility and sophistication of hardware and software has made it possible to design high-quality data management systems for community-based trials in resource-poor environments. We designed, implemented, and operated an effective data management system for the Nepal Nutrition Intervention Project Sarlahi (NNIPS), a placebo-controlled community trial to assess the impact of vitamin A supplementation on the preschool mortality of 38,000 children in the Sarlahi district of Nepal. A data center was established in Kathmandu, approximately 8 hr drive from the study area. The trial is now completed and over 200,000 forms have been processed. The majority of potential data discrepancies were identified and corrected by field workers during the interviews. Supervisors and forms editors located at the field office corrected most data errors. Once forms reach Kathmandu, the average number of data entry errors was 3.1 per 10,000 keystrokes for the eight data entry operators employed during the study. Extensive computerized checking of data during data entry found out-of-range, missing, or inconsistent data in only 1% of forms. Timely analysis of field worker performance provided ongoing feedback to supervisors and analysis for the Data Safety and Monitoring Committee, and for publication of results.