AI Can Help Reduce Opioid Prescriptions After Surgery, Research Finds

Artificial intelligence (AI) can help determine which patients are at high risk for severe pain after surgery, as well as which patients may benefit from non-opioid related pain management, a new study finds.

The research, which is being presented at the American Association of Anesthesiologists’ annual meeting, aims to reduce the number of people who are exposed to opioids in an attempt to lower the risk of opioid use disorder.

The U.S. is currently in the middle of an opioid crisis. Opioid overdoses caused more than 42,000 deaths in 2016, and an estimated 40% of opioid overdose deaths involved a prescription opioid, according to the U.S. Department of Health and Human Services (HHS). New research is pushing for medical professionals to prescribe opioids less often.

The study was broken into two parts. In the first, researchers analyzed data from 5,944 patients who underwent different surgeries, including gallbladder removal, hysterectomy, hip replacement, and prostate surgery. Of those patients, 1,287 (22%) had taken a high dose of 90 morphine milligram (mg) equivalent in the first 24 hours after surgery.

The researchers used 163 potential factors to predict who would have high pain after surgery and crafted three different machine learning algorithm models that analyzed the patients' medical records. The factors were then narrowed to those that most accurately predicted the patients' pain severity and potential need for opioid use after their operations.

Some of the most predictive factors of post-surgical pain included:

• Younger age
• Higher body mass index
• Female sex assigned at birth
• Pre-existing pain
• Prior opioid use

For the second part of the study, the researchers compared what their models predicted in terms of who actually needed to use opioids post-op. The researchers found that all three models accurately identified which patients were more likely to have severe pain about 80% of the time.

“This novel tool can help reduce unnecessary prescription of opioids for those who don’t need them and help people gain access to non-opioid treatments after surgery that may be more beneficial,” lead study author Mieke A. Soens, MD, an anesthesiologist at Brigham and Women's Hospital and anesthesiology instructor at Harvard Medical School in Boston, tells Verywell.

How Opioids Are Typically Prescribed

“Currently doctors do not routinely identify patients at higher risk for severe pain and high opioid requirements after surgery,” Soens says. Many doctors fill out questionnaires to try to determine who will be in the most pain after surgery, but the practice is “very time-consuming and cumbersome and impractical for use in every-day practice,” Soens says.

A person’s personal history may also be a factor, Vernon Williams, MD, a pain management specialist and founding director of the Center for Sports Neurology and Pain Medicine at Cedars-Sinai Kerlan-Jobe Institute in Los Angeles, tells Verywell. “If a patient has had previous surgeries, a lot of times they’ll tell the doctor that they had difficulty in the past,” he says. “Sometimes there is information in the chart that tells the doctor this may be more difficult.”

The type of procedure matters, too. “For certain kinds of procedures, we can anticipate that someone may have discomfort for a few days and will be given medications for a few days on an as-needed basis,” Williams says. “But the issue is whether there is an accurate way to predict where a patient may have a much more severe or longer degree of pain, and how you can anticipate that.”

Soens hopes her algorithm can help fill that gap. “The model that we developed uses data already available in the patient’s medical record and machine learning algorithms and is able to quickly and in real-time predict high opioid needs after surgery,” Soens says.

The goal, she says, is to be able to identify patients who will likely need a high dose of opioids after surgery and then work to try to maximize non-opioid options, as well as a “tailored pain management approach” for each patient.

Non-Opioid Treatment Options

Opioids aren’t the only pain management options available to patients. “One strategy is alternating acetaminophen and ibuprofen. This is pretty common,” Jamie Alan, PhD, assistant professor of pharmacology and toxicology at Michigan State University, tells Verywell.

Reassuring a patient and giving them a recovery timeline can be beneficial. “Letting patients know how long to anticipate the pain can be really helpful,” Williams says.

Intramuscular injections, oral anti-inflammatories, and topical treatments, including ice, can help, too, he says.

Nerve blocks and epidurals can reduce pain, Soens says. However, they’re not cheap. “These alternatives can be very costly and sometimes risky,” she says. “Therefore, being able to target the right treatment to the right patients is not only important to reduce opioid use, but also to ensure that patients receive the treatment that is right for them.”

Ultimately, Soens says, the goal is to give patients a more tailored pain management plan. “Many people have become users and over-users of pain medications after surgery, and we want to help reduce that burden for patients, their families, and society at large,” she says. "As physicians, we want to be able to get the right treatments for the right patients and help our colleagues to make the right decisions.”

Soens says her current model is “only the beginning”—she and her team plan to continue perfecting the algorithm.