An award-winning paper from Hospital for Special Surgery (HSS) in New York has looked at pain archetypes for total knee arthroplasty (TKA) and hoped to answer whether those archetypes are related to patient-specific pre-op factors (demographics, comorbidities) and intra-op factors (surgical and anesthesia technique). Or something else.
The unpublished work, “Classification and stratification of patient pain archetypes following total knee arthroplasty: a machine learning approach,” won the Best of Meeting award at the 50th Annual Meeting of the American Society of Regional Anesthesia and Pain Medicine.
This is an award which recognizes excellence in scientific research and it won by only three of the top ten highest-scoring abstracts chosen by the organization’s Research Committee.
“TKA is one of the most common orthopedic procedures performed in the United States (over one million annually), and post-op pain in particular can be a serious issue for many patients,” explained co-author Justin Chew, M.D., Ph.D. to OTW.
The research team used machine learning:
- to identify the different pain archetypes after surgery,
- to determine the factors that contribute most to post-op pain, and
- to predict how a given patient might do after TKA in terms of pain.
Dr. Chew, a regional anesthesiology and acute pain medicine fellow in the department of anesthesiology at HSS, told to OTW, “We know that after TKA, some patients tend to do much better in terms of post-op pain than others, but the extent of how these patients differ from each other has not been closely examined in a systematic way.”
“For instance, some patients might initially have significant pain that improves with treatment, and others might have persistently difficult to control pain throughout their entire hospitalization, and yet others may never have any major issues with post-op pain.”
Methodology
For this study, the researchers collected and analysed data for 9,970 patients who had been treated with total knee arthroplasty by HSS surgeons between January 2021 through January 2023.
The research team found four clusters of patients/pain archetypes:
- patients with consistently low pain scores (n=1,805),
- patients with low pain scores that experienced breakthrough pain which resolved appropriately after intervention (n=3,845),
- patients with initially low pain scores who then experienced worsening pain that was difficult to resolve (n=1,308), and
- patients with high pain scores throughout that did not respond to interventions (n=3,012).
The patients had correspondingly proportionate amounts of opioid consumed throughout hospitalization to discharge.
Preditors of Difficult Post-Op Pain
“The most significant predictors for difficult post-op pain,” Dr. Chew told OTW, “were younger age, higher body mass index/obesity, procedures planned as outpatient surgery, pre-op opioid or gabapentinoid use, and worse pre-op physical functional status.”
OTW asked Dr. Chew to explain the extent to which tailored/personalized pain management might work going forward.
“Currently, it remains a challenge to identify the patients at highest risk of having difficult to control post-op pain,” said Dr. Chew, “but with machine learning, our goal is to build a model that can take into account many individualized factors for a given patient, including patient comorbidities, patient demographics, and surgical factors.:
“It ultimately remains to be seen how personalized we are able to make our pain management plans with machine learning, as its effectiveness highly depends on the quality of the available data in the electronic medical record, and thus more work needs to be done before we are able to fully realize the power of machine learning in this area.”
Dr. Chew also told OTW, “Larger studies need to be performed before we are able to generalize our results to other patient populations—our patients tend to be younger and healthier than in many other hospital systems. Nevertheless, our findings are largely consistent with other work that has looked at risk factors for difficult post-op pain, and we hope that these results can inform clinical thinking for others in terms of tailoring individualized pain regimens for patients who are at highest risk for developing difficult to control post-op pain.”