Validity of self-reports of medication treatment of depression in general population surveys.

Abstract

To assess the concurrent validity of a single question on medication use for depression in a general population survey. Using data from 2015 to 2016 and 2017 to 2018 National Health and Nutrition Examination Survey, we compared responses to a single question on medication use for depression with responses to a detailed questionnaire confirmed by inspecting medication packages or pharmacy printouts. There was a strong agreement (96.4%) between response to a single question about using medication for depression and responses to questions about using antidepressants or other psychiatric medications for depression on the detailed questionnaire. The single-question assessment had excellent sensitivity (93.8%) and specificity (96.7%), positive predictive value (71.5%), and kappa (0.79). Psychometric properties were mostly consistent across population subgroups. Single-question assessments of medication use for depression have acceptable concurrent validity against more detailed assessments and provide an efficient method for assessing medication treatment of depression in population health surveys.