Completeness of published case reports on suspected adverse drug reactions: evaluation of 100 reports from a company safety database.

摘要

BACKGROUND: Case reports of suspected adverse drug reactions (ADRs) are common in the biomedical literature. Standards for authors and editors for writing, submitting and publishing ADR case reports have been empirically established since the 1980s; however, these recommendations have not been widely disseminated or incorporated into practice. Comprehensive and standardized guidelines on good publication practice have recently been proposed. No study has been performed so far to assess the adherence of published ADR case reports to these guidelines. OBJECTIVE: To describe the current situation with regards to the reliability and completeness of published ADR case reports. METHODS: A random sample of 100 single ADR case reports published between 2005 and 2008 (25 for each year) was retrieved from Pfizer's pharmacovigilance database. Reliability and completeness were assessed by comparing the relevant information contained in the retrieved ADR case reports against the recommendations prescribed by the guidelines. Descriptive statistics and correlation analysis using the Statistical Package for Social Science (SPSS) were undertaken. RESULTS: The patient's medical history relevant to the ADR was reported in 92% of the case reports. Concerning the suspected drug, 11% of the reports included the proprietary name; duration, dosage, route and formulation were reported in 87%, 85%, 37% and 21% of the reports, respectively. Information on concomitant therapies was included in 71% of the reports. The description of the ADR contained details on management (99%), time-course (97%) and diagnostic tests (95%), while final outcome and seriousness were reported in 73% and 52% of the reports, respectively. A discussion on the possible mechanism for the ADR was present in 70% of the case reports. The possible implications for clinical practice of the reported drug-event association were described in 75% of the cases. Causality assessment was reported in 81%, and rating scales to support the causal link were used in 20% of the reports. The major predictive factor for the presence of an objective causality assessment was found to be publication in specialized pharmacoepidemiology or clinical pharmacology journals: 47% specialized versus 11% non-specialized (odds ratio = 6.93; 95% CI 2.37, 20.26). CONCLUSIONS: The findings of this study show that published ADR case reports, especially those coming from non-specialized journals, still lack important information necessary for comprehensive evaluation. As published ADR case reports are expected to be reported to regulatory authorities using the same approach as for spontaneous cases, it is paramount for their effective integration in the pharmacovigilance system that pharmaceutical companies and learned societies actively promote a culture of good publication practices.