Researchers predict opioid risks using artificial intelligence

(Edmonton) Alberta researchers seek to predict the risks of opioid prescribing using artificial intelligence.

Physicians already have an established protocol for identifying patients at risk for opioid addiction. But the D^r Dean Eurich, director of the clinical epidemiology program at the University of Alberta, thinks artificial intelligence “could do a better job.”

With this tool, doctors could predict the effects of an opioid on patients and save them unnecessary emergency room visits or even death within 30 days of starting treatment.

The D^r Eurich is the principal investigator of an article published in December in the Journal of the American Medical Association Network, which analyzed medical data from more than 850,000 Albertans anonymously and predicted the best patient outcomes.

The D^D Fizza Gilani of the College of Physicians and Surgeons of Alberta said machine learning could be an effective way to reduce hospitalizations and patient morbidity.

Sometimes, she added, current methods don’t predict the origins of risk and medical solutions could be more complicated than just reducing a patient’s opioid dose.

The algorithm was fed with data related to various health factors to determine a patient’s risk, including history of injury, obesity, depression, diabetes and psychosis. These were combined with diagnoses, visits to hospitals and information concerning, in particular, the patient’s place of residence.

“The idea is not to get doctors to stop prescribing opioids, but to minimize risk after exposure to opioids,” argued Dr.^D Gilani.

The researchers looked at about three million opioid prescriptions a year from various healthcare professionals — doctors, nurse practitioners, dentists — to more than 600,000 Alberta patients. Those who had cancer or were receiving palliative care were excluded.

The D^r Eurich noted that 20% of patients took opioids with other high-risk medications, “increasing the risk of adverse effects.”

The D^r Eurich, who has worked with artificial intelligence for more than three years, said the system can “correctly predict for four out of five patients” who is at high risk and has a higher chance of being hospitalized within the first 30 days. the prescription of the drug.

Moms Stop the Harm co-founder Petra Schulz said most opioid-related deaths in the province are from illicit drugs, not prescriptions.

“This type of artificial intelligence could limit access to safer solutions, she fears. It’s like you’re doing detective work and want to figure out what’s wrong with the patient instead of developing a trusting doctor-patient relationship, which allows the patient to talk openly. »

The D^D Gilani agreed with M’s observation.^me Schulz on the opioid crisis, but recalled that there is an “indirect link” between a host of factors fed into the artificial intelligence system and that the tool could help reduce these deaths based on the data.

According to the D^r Eurich says a “good part” of the negative consequences associated with opioids are not fueled by illicit drugs, but by prescription use ― especially in the beginning.

He explained that patients continue to be exposed to opioids to treat pain and end up using the healthcare system to “shop around their doctor and get massive amounts of opioids.”

He added that his system would provide “good continuity of care” even when patients change doctors.

This article was produced with the financial support of the Facebook Fellowships and The Canadian Press.