Justin Cooper , 2025-05-12 09:30:00
Key takeaways:
- Community factors were more important than race in models predicting 90-day odds of readmission, revision and mortality.
- Living environments should be considered when planning care, researchers said.
Social determinants of health at the community level, such as median income and walkability, have a larger influence on hip replacement outcomes than an individual patient’s race, according to data.
The study was inspired by “persistent disparities observed in total hip arthroplasty outcomes, often attributed to individual factors such as race,” study author Bella Mehta, MD, MBBS, MS, a rheumatologist at Hospital for Special Surgery, told Healio.
“Although racial disparities are well-documented, we wanted to explore whether community-level social determinants of health, such as socioeconomic factors and community characteristics, might play a more significant role in total hip arthroplasty (THA) outcomes,” she said. “The study aimed to address the gap in understanding how these collective community factors compare to individual patient characteristics, including race.”
To accomplish this, Mehta and colleagues retrospectively analyzed patient data from the Pennsylvania Health Care Cost Containment Council database. Their analysis, published in Arthritis Care & Research, included 105,336 patients who underwent unilateral primary elective THA between 2012 and 2018.
The researchers additionally used Patient ZIP codes to glean data about their community-level social determinants of health, with a focus on factors “known to be relevant or to correlate with THA outcomes,” including median household income, percentage of householders with computer access and the walkability of the area, they wrote.
Individual factors under consideration included demographic variables, such as age, sex and race; home discharge vs. non-home discharge; and comorbidity burden, assessed via the Elixhauser comorbidity index.
Then, using a type of machine learning model called an “explainable boosting machine,” Mehta and colleagues assessed the likelihood of hospital readmission, revision surgery and death.
“This model is useful because it balances complexity and interpretability, allowing us to see which factors had the most impact on outcomes,” Mehta said. “Typical machine learning models do not allow us to see what is going on inside the model. This one is a ‘glass box’ model, whereby we can see how each feature is influencing outcomes.”
The cohort’s median length of hospital stay was 2 days (interquartile range: 1-3). Within 90 days, 8% of patients were readmitted and 0.3% died, while 1.5% demonstrated revision within 1 year.
The predictive performance of the models, assessed using the area under the receiver operating characteristic curve, was 0.76 for 90-day mortality, 0.66 for 90-day readmission and 0.58 for 90-day revision. A regression model predicting length of stay had a root mean squared error of 0.41 (R2 = 0.2).
According to the researchers, race was the least important factor in the models for 90-day revision and mortality. Community factors, analyzed in aggregate, were also more important than race in the 90-day readmission model.
For length of stay, community was the second-most important factor in the model — behind discharge location — while race was the least important factor.
One surprising result, according to Mehta, was that the walkability of patients’ neighborhoods was inversely correlated with their likelihood of readmission.
“This might seem counterintuitive, because walkable areas are often perceived as healthier environments,” she said. “However, this could reflect the urban nature of these areas, where factors like socioeconomics and health care access vary widely.”
Overall, the study “challenges the traditional focus on race as the primary determinant of THA outcomes and encourages health care systems to pay closer attention to the broader social environment patients come from,” Mehta said.
“This highlights the need for a more holistic approach when planning care, taking into account patients’ living environments rather than focusing solely on demographic factors like race,” she said. “We suggest that health care teams incorporate social support, community resources and discharge planning based on patients’ community contexts to improve outcomes.”
For more information:
Bella Mehta, MD, MBBS, MS, can be reached at drbellamehta@gmail.com; X: @bella_mehta.
Sources/Disclosures
Collapse
Disclosures:
Mehta reports no relevant financial disclosures. See the full study for additional authors’ disclosures.
