Objective: Alcohol screening and brief intervention (BI) are recommended preventive health practices. Veterans Health Administration (VA) uses a performance measure to incentivize BI delivery. Concerns have been raised about the validity of the BI performance measure, which relies on electronic health record (EHR) documentation. Our objective was to assess concordance between EHR-based documentation and patient-reported receipt of BI, and to examine correlates of concordance.
Method: Patients with a documented positive screen for unhealthy alcohol use at VA Greater Los Angeles primary care clinics were surveyed (within 15 days on average) in 2013-2014. Documented BI was indicated by an EHR note that the patient was advised to drink within recommended limits or reduce or abstain from drinking. Patient-reported receipt of BI corresponded to an affirmative response to questions on whether a VA provider advised the patient to drink less or abstain. Patient report and documentation were assessed over the same period.
Results: Documented and patient-reported receipt of BI had low concordance. Almost all patients who reported receiving BI had documentation of BI (93%; 95% CI [90%, 95%]), but only 63% [59%, 67%] of patients with documented BI reported receiving it. BI concordance was associated with more severe unhealthy alcohol use and drinking-related consequences, mental health comorbidity, and greater readiness-to-change alcohol use.
Conclusions: Discrepancies between EHR documentation and patient-reported BI raise concerns about performance measure validity. Patient-reported receipt of BI could be an alternative or complementary measure of BI.