A typical measure of success for spine surgeons is normally measured by a clinic visit at 30 days, 60 days, 90 days and maybe a two-year follow-up survey.
Jang Yoon, MD, director of minimally invasive spine surgery at the UC San Francisco, said that model is missing most of the story, and he has the phone data to prove it.
During the height of COVID-19, with elective spine cases paused, Dr. Yoon and a group of colleagues began looking for a way to measure recovery that did not depend on patients showing up to an office. He said the answer was sitting in their pockets.
“We use a smartphone app that we developed, install it in people’s phones, and then we get their [historical] activity data. Then we analyze it throughout longitudinally, before and after the surgery,” Dr. Yoon said.
The team has since published six peer-reviewed papers on the work, from a foundational 2021 study through machine learning models released in 2024.
‘A huge range’
Smartphones already log years of movement history passively. A patient enrolling today can contribute a decade of baseline data instantly; Dr. Yoon demonstrated this on himself, pulling 10 years of his own cellphone data. His studies typically use a one-year window before and after surgery to avoid unrelated events, such as an old injury or a cardiac event, that could muddy the baseline.
The data found that patients’ movement declined about three months before they ever saw a surgeon, living with the problem for an average of 12 weeks before booking an appointment. The team also found that the algorithm could distinguish a sudden injury from a slow, age-related decline with 92% accuracy, just from the pattern of activity over time.
Dr. Yoon said the most notable finding was that return to normal activity ranged from 5.6 weeks to nearly 30 weeks depending on the patient — a difference that a single 90-day follow-up date cannot capture.
“That’s a huge range,” Dr. Yoon said. “So to say that our measurement of people following up in 30 days, 60 days, 90 days, and then two-year follow-up, and asking them to fill out a survey — that’s actually a very crude way. We’re missing a lot of the information in between the clinic visits.”
The reimbursement angle
Dr. Yoon ties this to how CMS and payers judge surgical value: 30-day readmissions, 90-day reoperations and ER visits.
“Imagine that people’s lives don’t stop at 90 days, right? They continue on for many, many years,” he said. “So I think our policies and the payer policies, and the way we assess how well the patients are doing, are very crude — and that’s on both sides. That’s on the physician side, and that’s on the healthcare CMS side.”
In other words, fixed follow-up windows may be poorly matched to how patients actually recover, especially after less invasive procedures for which faster, more variable recovery is a primary goal.
What’s next
UCSF is now building AI models that predict a patient’s likely recovery curve before surgery even happens, using their preoperative activity data. The goal is to eventually counsel patients with a personalized recovery timeline instead of a population average.
The work has produced several patents and is ongoing, with collaborators from the University of Pennsylvania in Philadelphia, where Dr. Yoon previously served as director of minimally invasive spine surgery, also contributing.
At the Becker’s 32nd Annual Meeting: The Business and Operations of ASCs, taking place October 29-31 in Chicago, ASC leaders, surgeons and healthcare executives will explore strategies to drive growth, enhance operational performance, navigate reimbursement challenges and prepare for the future of ambulatory surgery. Apply for complimentary registration now.
