A PDA-based data collection system for a large-scale Tuberculosis prevalence survey in Western Kenya

We are conducting a tuberculosis prevalence survey in rural Western Kenya targeting 20,000 participants residing in an area monitored by a continuous demographic surveillance system (DSS). We demonstrate different data collection methods applied in the survey and discuss experiences gained. The survey uses Personal Digital Assistants (PDA) with integrated bar-code scanners for most field data collection, scannable data collection forms for a separate survey, and networked computer data entry with barcode readers for entry of laboratory data entry. Pre-existing DSS data are loaded into the PDA's to allow linking of survey data to DSS data. The PDA application, programmed using visual CE, prompts the interviewer for the need of an additional sputum sample, based on pre-defined combinations of symptoms. Data collected in PDA's are downloaded into a laptop in the field and queried, allowing any data issues to be discussed immediately with the interviewers. Flash drives are used to transport back up data to the data centre. Barcode scanners inserted in the PDA socket are used to read the participants unique study ID from pre-printed barcoded labels. Barcode scanners are used to track participants when they come for chest X-ray, and to track sputum samples and X-rays upon arrival in the laboratory and X-ray store respectively. Networked data entry of sputum smear results is done in the laboratory. The study number includes 1 check digit, to enable tracing of transcription errors when pre-printed barcode labels cannot be used (e.g. on microscopy slides). The data management process allows timely availability of clean data for reporting and provision of test results to study participants.