The costs outweigh the benefits: seeing side-effects online may decrease adherence to statins.

ABSTRACT

BACKGROUND: The prevalence of medical misinformation on the Internet has received much attention among researchers concerned that exposure to such information may inhibit patient adherence to prescriptions. Yet, little is known about information people see when they search for medical information and the extent to which exposure is directly related to their decisions to follow physician recommendations. These issues were examined using statin prescriptions as a case study. METHODS: We developed and used a tool to rank the quality of statin-related web pages based on the presence of information about side effects, clinical benefits, management of side effects, and misinformation. We then conducted an experiment in which students were presented with a hypothetical scenario in which an older relative was prescribed a statin but was unsure whether to take the medication. Participants were asked to search the web for information about statins and make a recommendation to this relative. Their search activity was logged using a web-browser add-on. Websites each participant visited were scored for quality using our tool, quality scores were aggregated for each participant and were subsequently used to predict their recommendation. RESULTS: Exposure to statin-related benefits and management of side effects during the search was significantly associated with a higher probability of recommending that an older relative adhere to their physician's recommendation. Exposure to misinformation and side effects were not associated, nor were any other participant characteristics. Bigram analyses of the top reasons participants gave for their recommendation mirrored the statistical findings, except that among participants who did not recommend following the prescription order, myriad side effects were mentioned. CONCLUSIONS: Our findings suggest that units of information people see on health-related websites are not treated equally. Our methods offer new understanding at a granular level about the impact of Internet searches on health decisions regarding evidence-based recommended medications. Our findings may be useful to physicians considering ways to address non-adherence. Preventive care should include actively engaging patients in discussions about health information they may find on the web. The effectiveness of this strategy should be examined in future studies.