Toward improved adverse event/suspected adverse drug reaction reporting
R. Gross,B. Strom
DOI: https://doi.org/10.1002/pds.800
2003-03-01
Pharmacoepidemiology and Drug Safety
Abstract:Despite advances in our ability to test hypotheses relating to adverse medication events at the population level using electronic medical records, and at the patient level using molecular techniques, spontaneous reports remain the primary mechanism for generating hypotheses regarding potential adverse drug effects. Other proposed strategies for hypothesis generation have included mining of various medical databases, but the development of these methodologies has been limited to date. The major advantage of spontaneous reporting systems with respect to hypothesis generation is the sheer number and heterogeneity of exposed individuals potentially included (i.e. all patients seen by clinicians). However, there are also major limitations, including inconsistent recognition and attribution by providers, barriers to reporting the events once recognized, and inaccuracy and incompleteness of data in the reports once they are filed. In this issue of Pharmacoepidemiology and Drug Safety, Rosebraugh et al. report on an intervention they undertook to improve the quality of adverse events reports. They added a teaching tool to a clinical pharmacology curriculum for medical students, randomized subjects to receive the added training or not, and tested the groups’ ability to complete the form immediately after the training session. They compared the quality of the reports of the intervention group to those of the control group using a scorecard developed for this project. The reports were judged by FDA personnel and it was found that medical students who were specifically trained provided better overall reports as compared with students without specific training. In particular, the intervention group were better at describing the adverse event itself and at identifying the suspect medication, although this latter difference did not achieve the traditional level of statistical significance. Yet, the results are not conclusive since the groups were not found to be different in all areas of quality assessment, most notably and importantly in whether FDA personnel believed that the information provided was sufficient to categorize the case as probably, possibly, or unlikely to be caused by the medication in question. The authors are to be congratulated for developing a program that trains medical students in an area in which to date they have been undereducated. In fact, high quality reports, as defined by the authors in the article, are particularly valuable for hypothesis generation and therefore must be a priority for any plans to improve the current system. However, as noted by the authors, the durability of the effect would be uncertain since studies have found that education to improve physician behavior have a relatively short duration of effect. Therefore, even if this intervention were effective, other strategies would likely be needed to continue to reinforce it. Furthermore, it is not clear what the efficacy of this intervention will be with respect to increasing the quality of reports actually received at FDA since improving an individual’s quality of reporting is only relevant in the setting in which that individual is likely to recognize and actually file a report. Unfortunately, the barriers to recognizing an adverse event as caused by a drug and then reporting the event may be more difficult to overcome than training reporters in filing quality reports. Two separate approaches must be considered, one focusing on increasing physicians’ reporting and another potentially parallel mechanism using ancillary medical personnel. To increase physician reporting, it would be important to increase the prominence of adverse effects recognition and reporting at all stages of medical training: in medical school, post-graduate training, and continuing medical education programs. The authors’ educational session did include a segment