Does Alcohol Screening Score Predict Alcohol-Related Health Outcomes?

A validated surrogate marker that predicts alcohol-related health outcomes would be useful for patient monitoring, research, and program evaluation. The 3-question Alcohol Use Disorder Identification Test-Consumption (AUDIT-C) is increasingly available in electronic health records. This retrospective cohort study analyzed data from 486,115 outpatients at 24 Veterans Affairs health care systems from 2004–2007 to determine the AUDIT-C’s predictive validity for high-density lipoprotein (HDL, an alcohol biomarker) and 2 alcohol-related health outcomes: gastrointestinal (GI) hospitalizations and physical trauma (any fractures or hospital discharge diagnosis of trauma) over the subsequent year.

• A baseline AUDIT-C score of 0 was associated with a subsequent mean HDL of 41.4 mg/l and a score of 12 with an HDL of 53.5 mg/l in the follow-up year.
• The probability of GI hospitalization increased from 0.49% for a baseline AUDIT-C score of 0, to 1.8% for a score of 12. Similarly, the probability of physical trauma increased from 3% for an AUDIT-C score of 0, to 6% for a score of 12.
• Compared with stable patients, those whose AUDIT-C score increased over time experienced increased HDL, and vice-versa. Probabilities of GI hospitalization and trauma increased with increases in AUDIT-C.

Comments:

Although this descriptive study suggests that changes in AUDIT-C correlate with changes in some alcohol-related health outcomes, measures of association, explanatory power and predictive ability are not presented. AUDIT-C might be a useful surrogate outcome for research and evaluation, but the prognostic importance of AUDIT-C, beyond its role as an indicator of alcohol consumption, remains difficult to infer for individual patients.

Peter D. Friedmann, MD, MPH

Reference:

Bradley KA, Rubinsky AD, Lapham GT, et al. Predictive validity of clinical AUDIT-C alcohol screening scores and changes in scores for three objective alcohol-related outcomes in a Veterans Affairs population. Addiction. 2016;111:1975–1984.