…in this instance, screening hospital patients to identify potential for problems with opioid use, the kind that might otherwise go unrecognized by physicians and nursing staff. From an article that appeared in April of last year:
AI screening for opioid use disorder associated with fewer hospital readmissions
NIH-supported clinical trial shows AI tool as effective as healthcare providers in generating referrals to addiction specialists
The screening instrument simply scanned the chart for notes and patient history, and from that alone was able to identify certain patterns that earlier research has associated with the potential for Opioid Use Disorder (OUD). That triggered an alert to the physician to watch for opioid withdrawal symptoms, and also signaled the need for a consult with an addiction specialist, probably from the hospital’s own treatment service.
That’s similar to the procedure now in many hospitals, except there the consultation would depend on the physician’s awareness of, and familiarity with, OUD and its treatment — as well as their willingness to address it.
It’s not always an area of strength for medical practitioners. Some openly resist the idea of dealing with addiction issues at all.
“That’s not what I’m here for, and it’s not what my patient is here for, either,” one GI specialist told our counselor, with a poke in the chest for emphasis. “And I don’t appreciate you people interfering with my care.” Ignoring the fact it had been the patient’s family that requested our help in the first place.
The study found that AI screening was actually superior to human assessment in terms of reducing the risk of an early readmission following discharge. A quick readmission is something hospitals (and insurance companies) seek to avoid.
“The AI screener was associated with fewer 30-day readmissions, with approximately 8% of hospitalized adults in the AI screening group being readmitted to hospital, compared to 14% in the traditional provider-led group.”
My guess: Superior nagging by AI in the form of alerts and reminders resulted in better physician follow-through.
Over the study period, the researchers projected some 16 fewer early readmissions from the AI screened group. That translated to a net savings of almost $7000 per readmission.
The average cost of a 30-day hospital readmission? They estimated it at $16,300, which I admit sounds ridiculously low, based on my own recent (intimate) experience with hospital bills.
“I’d like to visit that hospital,” joked one of my colleagues. “Where is it, Siberia?”
