Accuracy of the Structured Medication History Assessment Tool (MedHAT) Compared with Recorded Real‐Time Medication Use

Abstract

Medication histories are collected to measure drug exposure in epidemiologic studies, to identify adverse drug events and in clinical practice. The objective of this study was to compare the accuracy of a structured medication history obtained by using the Medication History Assessment Tool (MedHAT) with a medication diary in which subjects recorded real-time medication use. Prospective observational study. Outpatient research center. Sixty-seven adult healthy volunteers. Subjects were provided diaries to record the product name, dose quantity, and time that they used a prescription or nonprescription medication, supplement, or vitamin. After a minimum of 30 continuous days of diary use, a final interview was conducted, and medication history data were collected by using the MedHAT. We compared the medications reported during the interview with the medications recorded in the diary for each day to determine the accuracy of the verbal history. The primary outcome was the proportion of matches for each day for each subject. Overall accuracy for medication use for the day preceding the interview was 90%, and accuracy during the 30 days preceding the interview was 76%. Recall was higher for subjects taking prescription medications, those who had more recent medication use, older subjects, and subjects taking a higher proportion of prescription medications. The MedHAT system provided relatively high accuracy for immediate past use and for prescription medications and may offer better accuracy than other methods. Medication histories are often inaccurate, however, and may represent an important source of misclassification in epidemiologic studies. Thus medication histories alone should be used cautiously when measuring associations between drug exposure and health outcomes.