This post is authored by Eugenia Lin, 2020-21 Orthopaedic Value-Based Health Care Fellow; Lauren Uhler, assistant director of research; and Prakash Jayakumar, assistant professor of surgery and perioperative care.
Osteoarthritis of the knee is a major public health concern impacting the quality of life of almost one in seven adults in the U.S., with a concerning national treatment cost of $163 billion every year. Treatments for osteoarthritis range from physical therapy and weight-loss to joint injections and total knee replacement (TKR). The rates of TKR are rising and concerns have been raised around whether this procedure is the right choice of treatment for many patients in the first place, with one study finding up to a third of surgeries may be inappropriate. Now is the time for improved decision-making between surgeons and patients considering TKR.
One answer? Artificial intelligence (AI). A team from the UT Health Austin Musculoskeletal Institute (MSKI) developed a patient decision aid in collaboration with a digital health company. The decision aid uses AI and patient-reported outcome measures (PROMs) for enhanced shared decision-making around TKR surgery. AI is the branch of computer science that builds smart machines capable of performing advanced tasks. PROMs are surveys that patients fill out about their health, and how conditions like knee arthritis affect their everyday lives. Combining AI with PROMs, the decision aid aims to enhance shared decision-making — an approach that allows patients to identify preferences, learn about treatment options, and risks and benefits so they can make more informed treatment decisions. For a given patient, the AI considers the patient’s health background and PROM survey scores to create personalized predictions of future health outcomes following TKR.
In a recently published study, MSKI researchers evaluated the impact of using this decision aid for patients with knee osteoarthritis considering TKR. We used a randomized controlled trial to compare outcomes between a group using the AI-enabled decision aid, education and preference-setting tools compared to another group who received just the educational material.
Our study found that the group using the AI-enabled decision aid, education and preference-setting tools had higher levels of decision quality, shared decision-making, patient satisfaction and knee function outcomes compared to patients receiving just the educational tools without the AI-enabled decision aid. Importantly, the study showed that our decision aid produces positive outcomes from patients of all backgrounds — regardless of employment status or type of insurance, and including those that have traditionally felt less empowered to make informed decisions. For clinicians, the decision aid can help clinicians better understand a patients’ know-how and their fears and preferences around surgery, helping them to better manage the patients’ expectations.
The few patient decision aids that have incorporated PROMs show positive impact on functional outcomes of knee osteoarthritis. We believe that the decision aid in this study could also be improving patient engagement, helping to support patients in feeling like equal partners in the decision-making process. This decision aid demonstrates the benefits of a personalized, data-driven approach to shared decision-making for patients considering TKR — a benefit that could have far reaching consequences for individuals and populations suffering from knee osteoarthritis.