Do drug-related safety warnings have the expected impact on drug therapy? A systematic review

Purpose The need for drug-related safety warnings is undisputed, and their impact should also be evaluated. This systematic review investigates and assesses the impact of safety warnings on drug therapy. Methods Studies published in English between January 1998 and December 2018 were searched in EMBASE and MEDLINE, complemented by manual search. Randomised controlled trials, cohort studies with a before/after component, and case-control studies were included, selected to predefined criteria, and assessed for their reporting and methodological quality. Results Out of 7454 references identified, 72 studies were included. A total of 28/72 (39%) studies described the impact of safety warnings on drug therapy as being effective, whereas 12/72 (17%) studies did not. Further, 26/72 (36%) studies described a partial implementation of the warnings (one part of the warning had an impact on drug therapy and another did not). Unintended effects were investigated in 6/72 (8%) studies. While 34 (47%) studies examined safety warnings on psychotropic drugs using an interrupted time series (ITS) design (53%), a before/after (26%), and a time series design (21%), 38 (53%) studied other substances using an ITS design (34%), a before/after (40%), and a time series design (26%). The proportion of an effective impact on drug therapy was lower in the "psychotropic drugs" group (23%) than in the "others" group (53%). Conclusion Drug-related safety warnings induce intended and unintended effects. The included studies are of broadly varying methodological quality. To better compare their effectiveness, studies should be conducted using standardised procedures.