Impact of safety-related regulatory action on clinical practice: a systematic review.

BACKGROUND: After market approval, new serious safety issues are regularly identified for drugs that lead to regulatory action to inform healthcare professionals. However, the effectiveness of these safety-related regulatory actions is under question. We currently lack a comprehensive overview of the effects of these drug safety warnings on clinical practice to resolve the debate about their effectiveness.
OBJECTIVE: The aim of this systematic review is to provide an overview of studies that assessed the impact of safety warnings. STUDY SELECTION: A systematic search was performed for articles assessing the impact of Direct Healthcare Professional Communications or 'Dear Doctor' letters, Black Box Warnings and Public Health Advisories on clinical behaviour published between January 1996 and January 2010. The following variables were extracted: publication year, country, name of the drug, safety issue, specific safety warning (Direct Healthcare Professional Communication/Black Box Warning/Public Health Advisory), effect (intended/unintended) of the safety warning, outcome measure and study design. Papers were checked for several quality aspects. Study data were summarized using descriptive analyses.
RESULTS: A total of 50 articles were identified. Two articles assessed two different drugs and were therefore counted twice (n = 52). Thirty-three articles described the impact of safety warnings issued for three drugs and drug groups, i.e. third-generation oral contraceptives, cisapride and selective serotonin reuptake inhibitors. The remaining 19 articles described a broad variety of 14 drugs and drug groups. Twenty-five studies applied an interrupted time series design, 23 a controlled or uncontrolled before/after design, and four articles applied both. None of the articles could rule out the influence of confounding factors. The intended effects were reported in 18 (72%) of the 25 before/after analyses, whereas only 11 (41%) of the 27 interrupted time series analyses reported an impact. Only two (8%) of the before/after analyses against 11 (41%) of the interrupted time series analyses reported mixed impacts. When unintended effects were assessed in case of selective serotonin reuptake inhibitors and third-generation oral contraceptives, these were almost always present: in 19 of 22 and 4 of 5 articles, respectively. Our review shows that safety-related regulatory action can have some impact on clinical practice but firm conclusions are difficult to draw. Evidence is primarily based on three drugs and drug groups. Almost half of the studies had inadequate before/after designs and the heterogeneity in analyses and outcome measures hampered the reporting of overall effect sizes. Studies with adequate interrupted time series design reported a more mixed impact of safety warnings than before/after studies. Furthermore, this review shows the relevance of considering not only the intended but also the unintended effects of safety warnings.
CONCLUSIONS: There is a clear need for further research with appropriate study designs and statistical analyses, with more attention to confounding factors such as media coverage, to understand the impact of safety-related regulatory action.