Enhancing Evidence-Based Pharmacy by Comparing the Quality of Web-Based Information Sources to the EVInews Database: Randomized Controlled Trial With German Community Pharmacists

Background Self-medication counseling in community pharmacies plays a crucial role in health care. Counseling advice should therefore be evidence-based. Web-based information and databases are commonly used as electronic information sources. EVInews is a self-medication–related information tool consisting of a database and monthly published newsletters for pharmacists. Little is known about the quality of pharmacists’ electronic information sources for evidence-based self-medication counseling. Objective Our aim was to investigate the quality of community pharmacists’ web-based search results for self-medication–related content in comparison with the EVInews database, based on an adjusted quality score for pharmacists. Methods After receiving ethics approval, we performed a quantitative web-based survey with a search task as a prospective randomized, controlled, and unblinded trial. For the search task, participants were instructed to search for evidence-based information to verify 6 health-related statements from 2 typical self-medication indications. Pharmacists across Germany were invited via email to participate. After providing written informed consent, they were automatically, randomly assigned to use either web-based information sources of their choice without the EVInews database (web group) or exclusively the EVInews database (EVInews group). The quality of the information sources that were used for the search task was then assessed by 2 evaluators using a quality score ranging from 100% (180 points, all predefined criteria fulfilled) to 0% (0 points, none of the predefined criteria fulfilled). In case of assessment discrepancies, an expert panel consisting of 4 pharmacists was consulted. Results In total, 141 pharmacists were enrolled. In the Web group (n=71 pharmacists), the median quality score was 32.8% (59.0 out of 180.0 points; IQR 23.0-80.5). In the EVInews group (n=70 pharmacists), the median quality score was significantly higher (85.3%; 153.5 out of 180.0 points; P<.001) and the IQR was smaller (IQR 125.1-157.0). Fewer pharmacists completed the entire search task in the Web group (n=22) than in the EVInews group (n=46). The median time to complete the search task was not significantly different between the Web group (25.4 minutes) and the EVInews group (19.7 minutes; P=.12). The most frequently used web-based sources (74/254, 29.1%) comprised tertiary literature. Conclusions The median quality score of the web group was poor, and there was a significant difference in quality scores in favor of the EVInews group. Pharmacists’ web-based and self-medication–related information sources often did not meet standard quality requirements and showed considerable variation in quality. Trial Registration German Clinical Trials Register DRKS00026104; https://drks.de/search/en/trial/DRKS00026104