A feeling for the patient

Training as an orthopaedic surgeon forever changed people-watching for me. In my younger life, I could extrapolate romantically about the lives of people I observed. Was the young couple in love? That harassed-looking man — were his worries about work or family?

Now, instead, I note the subtle coxalgic gait of a young person walking toward me or the early planovalgus hindfoot alignment of a middle-aged woman a few steps of me ahead on the escalator, and the diagnostic motors in my mind start whirring. Although my furtive scanning became a family joke, this awareness of the musculoskeletal system is the gift of experience. That experience has value, even as new technologies emerge that show promise to replicate our diagnostic and surgical skills.

A feeling for the diagnosis
In college, I was introduced to Barbara McClintock, PhD, via Evelyn Fox Keller’s lyrical biography, “A Feeling for the Organism.” A Nobel laureate for her work in the field of genetics, Dr. McClintock tracked traits using time- and labor-intensive study of maize generations.

In relating the accomplishments of this remarkable scientist, Keller described the intuitive connection Dr. McClintock felt to the maize, evolving from her deep study of the organism. This resonated with me on first read as an undergraduate, and later, realizing the art of medicine mattered as much as the science.

Rather than seeing each trait as an isolated study subject, Dr. McClintock saw the interconnectedness of the organism, which allowed her to identify subtle changes and ultimately extrapolate into her description of epigenetics.

For a physician, seeing a musculoskeletal malady within the context of the whole patient enhances our diagnostic skill.

The death of Aunt Minnie?
Aunt Minnie was alive and well in radiology training when I was in medical school. If you missed her, “Aunt Minnie” referred to the traits that allow you to recognize your aunt out of a sea of other women, just as a constellation of specific findings on an image should lead a radiologist to a diagnosis. Artificial intelligence (AI) threatens to replace our radiology colleagues as algorithms get better and faster at spotting Aunt Minnie.

However, studies assessing the use of AI in radiology note multiple issues that compromise reads, including application to patients who differ from those used to train the system (i.e., dataset shift); age, race, and other patient factors; and the presence of secondary conditions such as degenerative changes.

Clinician-AI collaboration has been identified as one way to improve the technology, providing the vital component of context. Seeing someone who looks like Aunt Minnie at a family wedding makes a more plausible identification than spying her doppelgänger at a professional meeting. Bone edema with night pain and no history of trauma should similarly change the differential diagnosis.

Integration of AI-driven technologies across various facets of clinical care provides additional context and a powerful basis to get the right diagnosis. An AI tool can not only produce a consult note after transcribing an clinical encounter but also review the entire electronic health record to cull pertinent findings from imaging, lab studies, and prior visits.

As AI evolves, recognition of the need for ethical guardrails has evolved to guard against bias and ensure transparency and accountability. But a true feeling for the patient transcends diagnosis.

A recently presented study out of Johns Hopkins assessed the ability of AI to interpret body language, comparing performance between AI and humans in reading images and videos. The AI model came up short.

Communicating subtly with the patient, watching their family dynamics, and interpreting unspoken anxieties not only impact diagnosis but also direct the treatment plan. Technology can spit out a diagnosis and recommend a total knee replacement, but empathy determines whether surgery is really the best answer for a specific patient.

Aunt Minnie is not dead; however, even if AI can recognize her, as long as it cannot really understanding who she is, the technology still remains incomplete.

A feeling for the procedure
For pedicle screws, a meta-analysis demonstrated that robotic-assisted techniques beat freehand for perfect screw placement. In trauma cases, surgical navigation demonstrates high accuracy, low complications, and decreased radiation use.

Articles in this issue of AAOS Now highlight multiple aspects of robotics and navigation. The article by Khan et al reviews the impact of robotic arthroplasty techniques on surgeons in training. Robotic technology may facilitate teaching knee arthroplasty alignment and gap balancing, but the authors also note concerns about failure of trainees to develop manual skills.

Even if a robot can beat my accuracy, I believe there is important value to the combination of the subtle change of audible pitch and palpable resistance that tells me to ease up as the drill comes close to the far cortex. If surgeons are never developing these skills in the first place, what happens when the technology goes down?

Beyond technical skill, the feeling for a diagnosis and the best treatment plan that we experience in the exam room extrapolates to a “feeling for the case” in the OR. There is no substitute for appropriate preoperative planning, but all surgeons know that even with plans A, B, and C, unexpected findings sometimes demand creativity. Until AI, robotics, and navigation are trained on datasets as robust as all the cases of residency, fellowship, and practice, the art of orthopaedics matters.

Advancing technologies offer excitement and great promise for surgeons and their patients. The trope of the orthopaedic-surgeon-as-carpenter has some basis. We are more sophisticated than treating every musculoskeletal problem as a nail to match our hammer, but temptation exists to let the tool drive the treatment.

However, interpreting results and implementing new surgical techniques should be recognized as tools to enhance, not replace, assessment and treatment of our orthopaedic patients. Retaining a “feeling for the patient” can provide balance.

Alexandra E. Page, MD, FAAOS, is a foot and ankle specialist in private practice in San Diego, California, and the deputy editor of AAOS Now.

References

1. Keller EF. “A Feeling for the Organism.” WH Freeman and Company; 1983.
2. Shapira J, Diulus SC, Rosinsky PJ, Maldonado DR, Lall AC, Domb BG. Robotics and navigation as learning tools for fellows training in hip arthroplasty. J Am Acad Orthop Surg. 2021;29(4):176-181. doi: 10.5435/JAAOS-D-20-00357
3. Garcia K, McMahon E, Conwell C, Bonner MF, Isik L. Modeling dynamic social vision highlights gaps between deep learning and humans. Poster presented at: International Conference on Learning Representations 2025; April 24-28, 2025; Singapore. Accessed July 2, 2025. https://iclr.cc/virtual/2025/poster/27867
4. Rajpurkar P, Lungren MP. The current and future state of AI interpretation of medical images. N Engl J Med. 2023;388(21):1981-1990. doi: 10.1056/NEJMra2301725
5. Fatima N, Massaad E, Hadzipasic M, Shankar GM, Shin JH. Safety and accuracy of robot-assisted placement of pedicle screws compared to conventional free-hand technique: a systematic review and meta-analysis. Spine J. 2021;21(2):181-192. doi: 10.1016/j.spinee.2020.09.007
6. Karkenny AJ, Mendelis JR, Geller DS, Gomez JA. The role of intraoperative navigation in orthopaedic surgery. J Am Acad Orthop Surg. 2019;27(19):e849-e858. doi: 10.5435/JAAOS-D-18-00478