Substance use checklist demonstrates validity, feasibility in primary care

Key takeaways:

• It is the first study to review a substance use checklist of Diagnostic and Statistical Manual for Mental Disorders criteria in primary care.
• The findings support the tools’ use in diverse patient populations.

Recent data showed that an 11-item symptom checklist was effective in identifying substance use disorder severity in primary care, supporting its use as a clinical decision-making tool, researchers reported.

According to Theresa E. Matson, PhD, MPH, a research interventionist at Kaiser Permanente Washington Health Research Institute, and colleagues, 7% of adults in the United States meet the diagnostic criteria for substance use disorders (SUD), defined as at least two measures in the fifth edition of the Diagnostic and Statistical Manual for Mental Disorders (DSM-5). However, the proportion of patients who receive a diagnosis is much lower, ranging from 0.8% to 4.6%.

As a result, “low rates of diagnosis decrease opportunities for patients to receive treatment, despite evidence-based options (eg, pharmacotherapy for opioid use disorder, behavioral treatments for cannabis and stimulant use disorders),” they wrote.

The researchers conducted a cross-sectional study on the 11-item Substance Use Symptom Checklist, which is used to identify cannabis and other drug use, “to better understand whether this tool was a good measure of a patient’s underlying substance use disorder severity and whether performance differed across important groups based on age, sex, race or ethnicity,” Matson told Healio.

To the researchers’ knowledge, their study is the first to review a tool in primary care where all screening items correspond with SUD criteria in DSM-5.

They used item response theory (IRT) to evaluate the tool’s validity. IRT, they wrote, modeled the relationship between unobservable measures with factors like SUD severity and determined unidimensionality.

Matson and colleagues also conducted a differential item functioning (DIF) analysis, which determined how items performed similarly across characteristics like age, sex and demographic.

The study included electronic health record data that were collected during 23,304 primary care screenings performed from March 1, 2015, to March 1, 2020. The patients had a mean age of 38 years and 53.9% were men.

Among patients who reported daily cannabis use, other drug use or both, 26.3%, 30.2%, and 51.8%, respectively, met two or more items on the symptom checklist, according to Matson and colleagues.

The researchers also reported that for all drug and cannabis subtypes, IRT models supported the unidimensionality of the symptom checklist, while checklist items for subsamples had discriminated higher vs. lower SUD severity “as expected, with the probability of endorsing each item increasing as SUD severity increased.”
The DIF analysis further revealed that patient characteristics did not significantly impact total criteria counts, which supported use among diverse patient populations, the researchers wrote.

Notably, “when we implemented the Substance Use Symptom Checklist, we did not know how often patients would report symptoms due to substance use,” Matson said. “We learned early on, however, that patients often appreciated being asked about their substance use.”

Matson and colleagues also noted that the tool demonstrated several positive qualities, which included feasibility, affordability and being acceptable to primary care patients.

“The ability to measure SUD symptoms can help with SUD identification, symptom management, and treatment planning for patients,” they wrote. “The strong psychometric performance identified may help clinicians feel confident in measurement-based tools to support SUD identification and care in general medical settings where they are underrecognized and undertreated.”

They acknowledged multiple limitations to the study. For example, it was unknown which substance contributed most to symptoms among those who reported daily cannabis and other drug use, and patients may have underestimated or underreported substance use.

The researchers concluded that further research is needed “to evaluate test-retest reliability of the symptom checklist and discriminative validity compared with a confidential interview comparison standard for SUD.”