Researchers Build AI To Predict Repeat Psychiatric Hospitalization

Getting help to potential “high utilizers” will save safety net hospitals money.

People who are admitted multiple times a year to inpatient psychiatric hospitals could be better served by preemptive services that not only keep them out of the hospital, but save a bed for someone else. That’s why researchers at the University of Texas Health Science Center at Houston used a machine learning algorithm to find out who is most likely to become a so-called “high utilizer.”

Analyzing data over three years from nearly 10,000 patients at the UTHealth Harris County Psychiatric Center, a safety net hospital, researchers learned that people who did not complete high school, people with schizophrenia, and people with a personality disorder and another disorder, such as an anxiety disorder, were all more likely to be admitted three times or more in a year. The research was published in February in The Journal of Health Care for the Poor and Underserved.

“Rehospitalization in a psychiatric hospital affects patient care and adds financial burden to our current health care system,” said Lokesh Shahani, the hospital’s chief medical officer and co-author of the study, in a press release. He hopes the research will help them create “new treatment modalities to reduce their likelihood of future hospitalization.”

I am a reporter in New York covering mental health, public health, and the things that affect each of them, especially the economy and public policy. My reporting has