AI-Driven HSS Study Honored for Uncovering Risk Factors Behind Severe Postoperative Pain After Knee Replacement
A groundbreaking study by researchers at Hospital for Special Surgery (HSS) that harnesses artificial intelligence (AI) to identify patients at risk of experiencing severe pain after knee replacement surgery has received national recognition for its scientific excellence. The study, which used machine learning to classify pain archetypes and highlight critical predictors of postoperative pain, was named one of the Best of Meeting abstracts at the 50th Annual Meeting of the American Society of Regional Anesthesia and Pain Medicine (ASRA).
This prestigious honor is awarded to three abstracts among the top 10 highest-scoring submissions, as determined by the ASRA Research Committee. The award recognizes innovation and quality in pain research and highlights the impact of new scientific approaches on clinical care.
“It is an honor to have one of the top professional organizations in the field of regional anesthesia and pain medicine highlight the collaborative work of our department’s Pain Prevention Research Center,” said Alexandra Sideris, PhD, director of the Pain Prevention Research Center at HSS. “This award reflects our team’s dedication to advancing the science of pain prevention and underscores the broader recognition our work is receiving in the medical community.”
The Growing Challenge of Postoperative Pain After Knee Replacement
Total knee replacement surgery is one of the most common orthopedic procedures in the United States, with over one million procedures performed annually—a number that continues to grow due to an aging population and increasing rates of osteoarthritis. While the surgery is highly effective at restoring mobility and reducing chronic joint pain, a significant proportion of patients experience moderate to severe pain in the days and weeks following the procedure.
Managing this pain remains a clinical challenge, as standard approaches do not always account for individual differences in how patients perceive and process pain. Some patients recover quickly with minimal discomfort, while others endure persistent pain that affects sleep, delays physical therapy progress, and diminishes overall quality of life. This variability has driven researchers to explore new methods for identifying high-risk patients before surgery, allowing for more targeted and effective pain management strategies.
“There is a critical need to better understand the diverse pain trajectories of patients undergoing knee replacement,” said Dr. Sideris. “One of the most promising developments in this area is the application of artificial intelligence to large clinical datasets. With our robust patient database at HSS, we can now apply machine learning techniques to analyze patient demographics, health conditions, and medication histories to predict which individuals are most likely to experience severe postoperative pain.”
A Data-Driven Approach to Personalized Pain Management
The study was designed with several ambitious goals in mind: to identify distinct pain response patterns (referred to as “pain archetypes”) following total knee arthroplasty, to determine which patient characteristics were most strongly associated with severe postoperative pain, and to develop a predictive model to classify patients according to their risk of experiencing high levels of pain in the immediate postoperative period.
To achieve these aims, the research team conducted a retrospective analysis of 17,200 patients who underwent total knee replacement surgery at HSS between April 1, 2021, and October 31, 2024. This large dataset allowed the team to apply advanced machine learning methods, including both unsupervised and supervised learning algorithms.
“Using unsupervised machine learning, we identified two distinct clusters—or archetypes—of pain responses among the patients in our study,” explained Justin Chew, MD, PhD, a clinical fellow at HSS and the study’s lead presenter at the ASRA meeting on May 1. “One group consisted of patients who experienced well-controlled pain after surgery, while the second group included those who reported severe and difficult-to-manage pain. These archetypes provide a meaningful framework for understanding the range of postoperative experiences.”
After establishing these archetypes, the team used supervised machine learning techniques to determine which preoperative factors were most predictive of belonging to the severe pain group. Several variables emerged as significant predictors, including younger patient age, higher body mass index (BMI), preexisting physical or mental impairments, and the use of opioids or gabapentinoids prior to surgery.
“These findings challenge some conventional assumptions,” Dr. Chew said. “For example, we typically think of older adults as being more vulnerable, but in our study, younger patients were actually more likely to report severe pain after surgery. This could be due to differences in pain perception, expectations, or underlying psychological factors.”
Implications for Clinical Practice and Future Research
The insights gleaned from this study have immediate and long-term implications for clinical care. By identifying patients who are likely to face greater pain challenges after surgery, healthcare teams can design customized pain management protocols that begin even before the operation. These might include prehabilitation programs, mental health support, early intervention with non-opioid pain management strategies, or the use of nerve blocks and regional anesthesia.
“AI allows us to move beyond a one-size-fits-all approach,” Dr. Sideris noted. “Instead, we can develop tailored care pathways that take into account a patient’s unique risk profile. Ultimately, this means better outcomes, faster recoveries, and a more patient-centered model of care.”
Moreover, the award-winning study represents just one part of a broader research initiative at HSS focused on the role of artificial intelligence in pain prevention and surgical recovery. According to Dr. Sideris, the Pain Prevention Research Center plans to extend its work by tracking patients over longer periods of time to examine how pain trajectories evolve in the weeks and months after surgery.
“We are currently developing models that can predict not just immediate postoperative pain, but also long-term pain and function outcomes,” she said. “This kind of work can inform the design of interventions that span the entire surgical journey—from preoperative preparation through rehabilitation.”
Recognition as a Milestone in Pain Research
Being recognized by ASRA—one of the leading societies in the field of pain and regional anesthesia—reinforces the importance of this research and its potential to change the way pain is managed in orthopedic surgery.
“This award is a validation of the direction we’re heading,” Dr. Chew said. “It shows that integrating advanced data science with clinical insight can produce meaningful results that resonate with both researchers and practitioners.”
As the healthcare system increasingly embraces personalized and predictive medicine, studies like this one from HSS are at the forefront of transforming patient care. By leveraging artificial intelligence and large-scale data analysis, the HSS team is not only advancing the science of pain management but also setting new standards for what patients can expect in their surgical recovery.
The HSS Pain Prevention Research Center continues to invest in AI-driven research to improve patient outcomes across a range of procedures. Future studies are expected to incorporate additional data sources, such as electronic health records, wearable devices, and patient-reported outcomes, to create even more robust and accurate models for predicting and managing pain.
In the long term, the goal is to translate these insights into practical tools that clinicians can use at the point of care—whether that’s a risk calculator integrated into the surgical planning process or a digital app that provides personalized recovery guidance.
“With each study, we’re building a deeper understanding of how pain works, why it affects patients differently, and what we can do about it,” Dr. Sideris said. “AI is helping us unlock that knowledge in ways that were not possible before.”
