RESUMEN
PURPOSE: To categorize, quantify and interpret findings documented in feedback letters of monitoring or auditing visits for an investigator-initiated, peer-review funded multicenter randomized trial testing probiotics for critically ill patients. MATERIALS & METHODS: In 37 Canadian centers, monitoring and auditing visits were performed by 3 trained individuals; findings were reported in feedback letters. At trial termination, we performed duplicate content analysis on letters, categorizing observations first into unique findings, followed by 10 pre-determined trial quality management domains. We further classified each observation into a) missing operational records, b) errors in process, and potential threats to c) data integrity, d) patient privacy or e) safety. RESULTS: Across 37 monitoring or auditing visits, 75 unique findings were categorized into 10 domains. Most frequently, observations were in domains of training documentation (180/566 [32%]) and the informed consent process (133/566 [23%]). Most observations were missing operational records (438/566 [77%]) rather than errors in process (128/566 [23%]). Of 75 findings, 13 (62/566 observations [11%]) posed a potential threat to data integrity, 1 (1/566 observation [0.18%]) to patient privacy, and 9 (49/566 observations [8.7%]) to patient safety. CONCLUSIONS: Monitoring and auditing findings predominantly concerned missing documentation with minimal threats to data integrity, patient privacy or safety. TRIAL REGISTRATION: PROSPECT (Probiotics: Prevention of Severe Pneumonia and Endotracheal Colonization Trial): NCT02462590.