Problems and pitfalls in community-based outcomes research

OBJECTIVE: To assess the feasibility of multisite community-based outcomes research. DESIGN: Prospective observational study of variations in treating acute external otitis by otolaryngologists and primary care practitioners. SETTING: Community-based independent otolaryngology practices. PARTICIPANTS: Patients with external otitis treated by otolaryngologists in Project Solo, a nonprofit, grassroots organization of independent physicians united for quality patient advocacy, and cost containment. METHODS: Confidential (bar-coded), disease-specific outcomes questionnaires completed by patients (12 items) and by participating physicians (15 items), Response to treatment was measured with a follow-up patient questionnaire (3 items). RESULTS: Nine patients were recruited from 5 of 29 enrolled otolaryngologists. Primary care practitioners were more likely to prescribe oral antimicrobials than otolaryngologists (100% vs. 44%, p = 0.03), but less likely to insert a wick in the external auditory canal (11% vs, 78%, p = 0.02). Poor recruitment was caused by an overly long and complex survey, data collection at multiple time points, lack of time during office hours, cumbersome data collection requirements, inadequate ongoing communication, and a lack of enthusiasm for the project. CONCLUSIONS: Future efforts at implementing a multisite outcomes study will require shorter questionnaires, smoother integration of the. survey process with regular office flow, simplified procedures for data exchange, frequent communication with data collection sites, and motivational programs for participating physicians and their office staff.