Sociodemographic disparities in individuals with total hip (THA) and knee (TKA) joint arthroplasty have been well-documented. Complication rates, readmissions, infections, and revisions are more common among women, low-income patients, and patients who are minoritized. Evidence of potential disparities in longer-term functional outcomes is mixed, and there is very little information available about the role of rehabilitation services in disparities. I first became interested in this topic while working as a physical therapist in a clinic located at the intersection of a high-income, well-resourced neighborhood and a lower-income, medically underserved neighborhood. The purpose of this study was to investigate differences in community-based (home health and outpatient settings) rehabilitation utilization and outcomes in patients post-THA and TKA.  The study used electronic health record data on all patients who received THA and TKA in the UPMC health system between 2017-2021, including home health and outpatient data for those who received community-based rehabilitation services within the system. We examined process-of-care outcomes including the number of visits, the timing of care, and the length of episodes within the home health and outpatient settings. We also examined functional outcomes including an activities of daily living composite measure, calculating items from the OASIS instrument (home health) and the Knee Outcome Survey (outpatient). We examined four sociodemographic variables: (1) sex (dichotomized as male or female); (2) race (limited to studying non-Hispanic White and non-Hispanic Black patients, due to very low proportions of patients of other races and ethnicities in the data set); (3) state-level Area Deprivation Index (ADI), using a tool based on home address and US Census data; and (4) rural vs. non-rural residence, categorized using a Federal Office of Rural Health Policy tool based on zip code. We used regression modeling to examine the associations between each of the four sociodemographic variables, functional outcomes, and rehabilitation utilization, and further examined how the interactions between rehabilitation utilization and sociodemographic variables affect functional outcomes. We found that in both home health and outpatient settings, female and Black patients used more visits over a longer period of time, but had poorer functional outcomes. In low-ADI and rural populations, there were no significant differences in outcomes or utilization in adjusted models.

Challenges in this study included the availability of data. We originally intended to examine the role of socioeconomic status using patient insurance (Medicaid or dual eligibility) but were not able to do so because of misclassification in the dataset. Therefore, we used ADI instead of Medicaid/dual eligibility as a proxy for socioeconomic status. Missing outcomes data in our outpatient rehabilitation files resulted in smaller sample sizes for outpatient functional outcomes. For future studies, investigators should pay close attention to the expected cohort size in their dataset and missingness in key variables to ensure adequate sample size to answer the research question. Our future work will focus on understanding the role that rehabilitation plays in generating disparities in functional outcomes after arthroplasty, and how rehabilitation care can be optimized and personalized to ensure that every patient can achieve the best possible functional outcome.