Listen to Austin 360 Radio

Dell Medical School helps develop artificial intelligence tool for knee surgery

The artificial intelligence tool developed by OM1 and Dell Medical School gives information on a tablet about knee replacement.

If you had osteoarthritis in your knee, how would you know when it's the right time to have knee replacement surgery?

Dell Medical School at the University of Texas helped patients at UT Health Austin's Musculoskeletal Institute answer that question by studying a new artificial intelligence tool it helped develop with health solutions company OM1.

Researchers at Dell Medical School and UT Health Austin did a randomized study of 129 people who came to them to consider total knee replacement surgery because of osteoarthritis. In the study, 69 people were given a computerized survey of questions to answer right before their visit with the doctor. The other 60 people in the study were given educational materials alone. 

Those who took the survey had their answers analyzed using artificial intelligence to compare their case against hundreds of thousands of other cases. The patients were given a decision aid that offered patient education, preference assessment and personalized outcome estimations based on their answers, medical history and social factors. That information also was compared with hundreds of thousands of other people who had had the same operation. The control group received some standard educational materials. 

Mental health care:Texas losing $2.2 billion by not treating mothers' mental health, report says

Coronavirus care:Not a waste: UT Austin researchers are tracking COVID-19 in a valuable way

Health care for the homeless:John Paul DeJoria gives $2 million for Dell Medical School chair on homelessness

In the decision aid, the computer looks at everything from the patient's age, sex, weight and body mass index to other medical conditions, the number of times that patient has been hospitalized, and the patient's physical and emotional wellbeing, says Dr. Prakash Jayakumar, an assistant professor in surgery and peri-operative care at Dell Medical School and director of value based health care and outcome measurement.

The report from the decision aid sent to patients and their doctors is personalized to predict the risks and benefits of the surgery as well as the likelihood of improvement in their pain. 

"We've never had that before," Jayakumar says. "It's not guess work. It's based on hundreds of thousands of patients." 

The OM1 survey uses a patient's answers to create a personalized report about whether knee replacement surgery is recommended.

Jayakumar says that doctors are learning there is evidence that "a higher-than-expected proportion of patients really are unsatisfied and not doing as well as you think" after having had surgery, even though the medical community might be touting the advancement of total knee replacement surgeries. 

This tool can give both the patient and the doctor an idea of how likely that patient might be at getting back on the golf course or other activities without pain after surgery. For some patients, if that's not going to be possible, the surgery might not be worth it. The tool also considers the patient's social and emotional environment. For example, the patient who has anxiety or depression, has difficulty losing weight, or doesn't have the right social support might not see a good outcome after surgery. 

The tool then can help guide the conversation about why surgery might not be recommended. If the risk factor of being a smoker is identified, there can be a conversation around helping that patient with a smoking cessation program. If it's around diabetes management or obesity, then the conversation might be about nutritional changes. 

Patients can work on some of the things the tool recommends and retake the survey to see if the predicted outcomes have changed. Or, they can decide not to have surgery and instead look at pain relief strategies.

For those patients who received the decision aid, it's not just the surgeon tipping the scales toward or away from surgery. Instead, it's data-based. "It all really comes down to a better way of looking at it," Jayakumar says. "It's the right care, the right decision at the right time." 

As the tool analyzes more and more patients and their data as well as their outcomes, it becomes more and more precise in its analysis, Jayakumar says.

The report created using artificial intelligence makes a recommendation about whether or not you should have a knee replacement surgery.

The study, which was published in the Journal of the American Medical Association in February,  found that the group that received the decision aid reported better decision quality, more shared decision making and improved functional outcomes. 

Following the study, UT Health Austin and UT Health San Antonio are continuing to use the tool for their patients looking for knee replacement because of osteoarthritis. They are planning to build and expand the tool for other surgeries such as hip replacements and spinal surgeries. They also will look at developing it for more than osteoarthritis conditions for knee surgery. 

Since the study was released, Jayakumar says they have had a lot of interest in the tool at other locations and in other surgical specialties.  

Nicole Villalpando writes about health care for the American-Statesman. She can be reached at nvillalpando@statesman.com.