Much of orthopedics and sports medicine is rooted in patient history and a good physical exam.
“If a patient comes to see me and says, ‘Dr. Frank, I was playing soccer and twisted my knee awkwardly and I felt a pop,’ in my head I’m already thinking of three likely diagnoses: an ACL tear, a patella dislocation, or a meniscus tear,” says Rachel Frank, MD, professor of orthopedics at the University of Colorado School of Medicine. “Just from that, I know where I’m going with my exam.”
Orthopedics, she says, is uniquely reliant on knowing how the patient was injured, what symptoms and pain they are experiencing, and what a doctor can see and feel.
“You just have to use your brain, training, and your hands,” Frank says of most diagnoses.
Even so, artificial intelligence (AI) is becoming more prominent in all aspects of the specialty, from the initial evaluation all the way through surgery and rehabilitation. For Frank, learning the technology is critical because trained models can yield excellent results, but so is double-checking the results of that technology.
“I'm always going to do my exam. I'm always going to look at the imaging. Even if I can get AI to tell me what the answer is, I still feel that I need to verify,” she says.
In a 2024 literature review of AI applications for orthopedic surgeons that Frank co-authored, researchers wrote that AI “enhances surgeons’ performance in many different areas, including diagnosis, and precision of surgery.”
“This in turn aims to improve the outcome of patients care,” they concluded. “The ability of surgeons to keep pace with this rapidly evolving field will be vital to exploiting future technological developments.”
From clinic notes to rare diagnoses
AI’s subsets, such as machine learning (ML) and large language models (LLMs), are useful for recognizing patterns in data and making conclusions about those patterns. In many aspects of medicine, these tools have become powerful diagnostic resources, but for now, Frank’s biggest use of AI is in the form of notetaking during an evaluation.
“From a clinical perspective, we're using AI to help in clinic for notetaking. With patient permission, we let the application record during the evaluation and I talk to the patient, discussing the exam, the imaging, the diagnosis, and treatment options as normal. At the end, I hit ‘stop’ and it generates a note,” Frank says.
Unlike a regular transcription program, which may just turn an audio recording into a transcript, this resource uses AI to synthesize the history and exam, and produces a plan. It’s not perfect, Frank says, but it does increase efficiency and helps free up more time in her day, so instead of creating all the notes, she only has to tweak them to fit her preferred style.
Much of AI’s potential in sports medicine is focused on diagnostic assistance and surgical support. In many ways, AI models have revolutionized imaging, and robotics are commonplace in orthopedic surgeries to help create precise bone cuts in some surgeries.
Still, Frank says it’s the human touch that stands out in her field.
“I really believe that there always needs to be a human factor in medicine, though AI will likely be good for the majority of normal and typical abnormal findings,” Frank says. “However, when it comes to the zebras, or the outlier cases, there may not be enough data to detect those. If we're exclusively relying on AI to read an x-ray, read an MRI, or read a pathology specimen, there’s a possibility that something gets missed.”
Improving care, reducing obstacles
Patients with sports injuries shouldn’t expect AI to take over their doctor’s job anytime soon. They may see more instances of it in their care, however. Tools powered by AI can be especially helpful in physician-patient interaction and help analyze a patient’s progress post-surgery.
For example, the development of wearable technologies has become a popular use of AI. Tools, some of which can be integrated into smart watches, can make daily predictions, alert doctors if the tool detects possible speed bumps in recovery, and deliver virtual techniques for rehabilitation.
These uses of AI in the evaluation, surgery, and in recovery can be helpful. However, Frank says physicians should be mindful of whether these tools are improving the workflow and who is picking up the tab for the technology.
“We've seen, in some instances, despite fancy technology, surgeries can take longer without any improvement in outcomes or any decrease in complications,” she says. “If that's the end result, what's the point? We should be asking: Will this use of AI improve outcomes? Will it reduce complications, or a minimum, result in no difference than the standard of care, while improving efficiency? If it does that, great, we should adopt it and promote it in our work.”
There’s also the cost of the technology.
“If these technologies, which currently can be very expensive, don't result in improved outcomes, maybe no worse, but no better, and cost more, who is footing the bill, and is that the patient, the physician, the hospital system, or the insurance company? Understanding the cost relative to the benefit of all of this is crucial. We all have different measures of benefit, so there is a need to focus on these questions and what that means for implementation.”
Without proper consideration and research, disparities could be the outcome.
“If cost is an issue, access to care could be an issue, because if there's a patient that wants to pay for the advanced technology over another patient that can't afford it, that will create disparities and barriers,” she says. “If there's a medical practice that can afford it versus another practice that can't afford it, that will also create disparities and barriers.”
For Frank and others in her field, there’s no denying that AI is changing the game. In a lot of circumstances it’s for the better, but big questions are still ahead of physicians and researchers.
“This is the world that we're in, and AI is expanding in every aspect of what we do, and sports medicine is no different,” Frank says. “I think our tasks as physicians is to make sure, number one, trust, but verify. We should embrace technological advances and figure out how to make it best for our patients in the safest, most efficient way.”
