Substance use disorders (SUDs) are underdiagnosed, leading to fewer opportunities for treatment. To screen for SUDs in primary care, the U.S. Preventive Services Task Force recommends the use of brief, validated substance use screens that ask about a person's frequency of cannabis and other drug use. Researchers designed a cross-sectional study to evaluate the psychometric properties of a standardized substance use symptom checklist when used routinely in primary care patients reporting daily use of cannabis, other drugs, or both.

Adult patients in a large integrated health care system who reported "daily or almost daily" or any drug use (that is, a response other than "never") on a behavioral health screen as part of routine care were asked to complete a symptom checklist that included 11 items corresponding to SUD in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition.

A total of 23,304 screens were completed, with 16,140 patients reporting daily cannabis use only, 4,791 reporting other drug use only, and 2,373 reporting both daily cannabis and other drug use. Among patients reporting daily cannabis use only, other drug use only, or both daily cannabis and other drug use, 26.3%, 30.2%, and 51.8%, respectively, reported symptoms consistent with SUD. The symptom checklist measured SUD along a unidimensional continuum with items that discriminated severity of SUD in each subsample. Differential item functioning analyses indicated that patient age, sex, race, and ethnicity had minimal impact on accurately assessing SUD severity.

Some patients may have underestimated or underreported substance use and symptoms, the authors caution. Furthermore, dichotomous endorsement of symptoms may not reflect the reality of how patients experience an SUD, and the symptom checklist didn't assess the frequency of criteria. Finally, because the study population included predominantly non-Hispanic White patients who were insured by an integrated health care system in a state where adult cannabis use is legal, the authors note that their findings may not be generalizable to other patient populations, health systems, and settings.