Although clinical trials demonstrate that screening for alcohol use in primary care is efficacious in identifying alcohol use disorders, the performance of screening in real-world settings has not been assessed. Using a national sample of 6861 Veterans Affairs (VA) outpatients, investigators compared AUDIT-C* results obtained during clinical care with AUDIT-C results obtained via mailed survey. A score ≥5 was considered positive.
- Results were discordant in 561 patients (8% overall): 468 of 765 patients (61%) who screened positive on the mailed survey had a negative clinical screen, while 93 of 390 patients (24%) who screened positive in the clinic screened negative on the survey.
- Factors independently associated with discordance in multivariable analysis included black/African American race (odds ratio [OR], 2.1), VA network site (highest OR, 11.7), and mailed survey score (compared with a score of 0, a score of 3–4=OR, 8.9; a score of 5– 7=OR, 69; and a score of 8–12=OR, 46).
These results suggest scaling up alcohol screening in real-world settings is not foolproof. Although this study did not capture details of specific implementation strategies at the sites, it importantly reminds us that high rates of screening are not equivalent to high-quality screening. Hillary Kunins, MD, MPH, MS
Bradley KA, Lapham GT, Hawkins EJ, et al. Quality concerns with routine alcohol screening in VA clinical settings. J Gen Int Med. September 22, 2010 [E-pub ahead of print].