Machine Learning Flags Untreated Alcohol Use Disorder

Most people with substance use disorders never get ID’d when they go in for a checkup. That’s partly why dual diagnosis treatment often starts years later than it should. A new study tests whether the data already in a patient’s chart can close that gap.

Researchers from the northeast trained supervised machine learning models on electronic health record-like data and found the models could reliably identify people with untreated alcohol use disorder.

Machine Learning in Action

Researchers drew on a national US survey to build data to mirror a real health system. After that, they made models to flag patients with drug addictions that were never treated.

The alcohol models worked really well. Those for opioid and cocaine use disorders showed promise but were limited by small sample sizes. 

The main point is that these models can show how many people a health system must screen to find one untreated case compared with screening everyone.

Screenings Miss So Many People

Substance use disorders go widely undetected in health care settings in part because comprehensive screening is difficult to fit into a brief clinical encounter.

Screening tools exist, but a busy clinic has to choose what fits.

Notably, the researchers built and evaluated model versions with and without psychiatric comorbidity. That separates what psychiatric history contributes to detection from the rest of the record.

Connecting Mental & Behavioral Health

Substance use disorders and psychiatric conditions go together. Treating one without addressing the other tends to produce poor results. Depression, anxiety and bipolar disorders all appear at higher rates among patients who use drugs. The relationship goes both ways.

That’s the definition of dual diagnoses. Integrated care teams that treat them together work better than separating them.

A detection tool that IDs untreated drug use in someone already receiving mental health treatment is a dual diagnosis identification tool. 

Treating Dual Diagnoses

Treatment approaches used in integrated programs include:

• Group behavioral therapy to address emotion and thought patterns underlying both substance use and mood or anxiety symptoms
• Trauma-focused counseling to fight the high overlap between PTSD and drug use
• Medications for psychiatric conditions and alcohol or opioid addictions

The authors are careful about the gap between a model that performs well on survey-derived data and a support system in real life. 

For now, if a substance use concern exists alongside a mental health diagnosis, talk about it. Clinicians can’t treat what they don’t know about. Current screening catches only a fraction.

Comprehensive Treatment for All

That’s where our comprehensive treatment directory comes in. Browse our listings by condition treated, level of care and whether the program provides integrated dual diagnosis treatment. 

Feel free to also dial 800-908-4823 (Sponsored) for free and confidential assistance with an expert.

Author

Eric Owens

Read Bio

Eric Owens has been a writer and editor for various businesses as well as his own successful websites. He has extensive experience creating content in the health and wellness space and the sustainability space. He holds a bachelor degree in Philosophy which helped him with presenting complex information in a simple way that all audiences can understand.

Author

Peter Lee, PhD

Read Bio

Peter W.Y. Lee is a writer and historian of American history during the Cold War. His primary focus is the relationship between youth and popular culture and its impact on U.S. society during the twentieth century. He has published widely on how the public has used popular culture as a mechanism to address political and social shifts throughout time