A study to enhance clinical end-user MEDLINE search skills: design and baseline findings.

abstract

OBJECTIVE: To determine if a preceptor and timely, individualized feedback improves the performance of physicians in searching MEDLINE using GRATEFUL MED in clinical settings. DESIGN: Randomized controlled trial. SETTING: A 300 bed primary to tertiary care teaching hospital. Computers were installed in wards and clinics of 6 major clinical services, and the emergency room, intensive care and neonatal intensive care units. SUBJECTS: All physicians and physicians-in-training from the departments of Medicine, Family Medicine, Surgery, Psychiatry, Pediatrics, and Obstetrics and Gynecology were included if they made patient care decisions for at least 8 weeks during the study period. INTERVENTION: All participants were given a 1-hour training class and 1 hour of individualized searching with 1 of the 2 study librarians. After training, participants were randomized to a control group who received no further intervention or to an intervention group in which each person chose a clinical preceptor experienced in MEDLINE searching and received individualized feedback by a study librarian on their first 10 searches, indicating search quality and providing suggestions for improvement. Feedback was mailed the first week day after the search was done. MAIN MEASURES: Baseline characteristics by study group, department and level of training, study participation rates, and searching rates. MAIN RESULTS: 308 of 392 eligible physicians joined the study. Participation was almost 80% with some variation by department and level of training. Excellent balance in the baseline characteristics was achieved for the 2 groups, as well as for the number who did first searches. Intervention group participants searched MEDLINE more often than did controls (3.5 searches per month vs 2.5 per month for controls, P = 0.046). The recall and precision for first searches for both groups was significantly less than that of librarians. The analysis of study data will be completed by September 1991. CONCLUSIONS: Clinicians are willing to do self-service searching of MEDLINE in clinical settings but their precision and recall are less than a trained librarian at baseline. Search skills enhancements are needed and the effect of feedback and preceptors is being tested. SOURCE OF FUNDING: U.S. National Library of Medicine and Ontario Ministry of Health.