Identifying potential adverse effects by patients' ratings: a proof-of-concept study of a novel approach.

OBJECTIVE Methods to evaluate adverse effects of medications are significantly underdeveloped compared to those for efficacy. In this pilot proof-of-concept study, we preliminarily compared a novel approach-the Symptom Assessment Tool (SAT)-to a systematic and detailed assessment by a physician for identifying symptoms that were potentially adverse effects (sensitivity) and excluding symptoms that were unlikely to be adverse effects (specificity). METHODS A symptom inventory and rating of symptom severity were completed before starting a psychotropic medication (or increasing its dose), and again 2 weeks later. Each symptom was systematically assessed by the patient-rated SAT and by a physician and was classified as either a potential or unlikely adverse effect. The primary analysis compared the classification of symptoms by the SAT to that by the physician. Potential adverse effects were also subcategorized as possible or probable adverse effects. RESULTS A sample of 193 symptoms from 15 adults was evaluated, only 37.3% of which were considered potential adverse effects by the physician. Sensitivity of the SAT compared to physician's assessment was 90.3% for potential adverse effects and 97.5% for the subgroup of probable adverse effects. The SAT correctly identified 63.6% of the symptoms as unlikely adverse effects (specificity), and its negative predictive value was 91.7%. CONCLUSIONS The SAT, appropriate for its intended use as a screening tool, had high sensitivity and moderate specificity and could present physicians with a limited number of potential adverse effects for further assessment and intervention. Further evaluation and refinement of this approach is warranted.