The Effects of Sociodemographic Factors on Variation in Baseline Patient Reported Outcome Measures in Foot and Ankle Patients

Abstract

Category: Other Introduction/Purpose: Patient Reported Outcome Measures (PROMs) are used to inform treatment, payment, and health policy decisions. As such, it is crucial to develop a robust understanding of extrinsic factors influencing PROM scores. Prior studies suggest sociodemographic factors are associated with variation in outcome scores not only in response to treatment, but also at initial presentation. The purpose of this study was twofold: first, to deepen our understanding of sociodemographic factors affecting baseline PROMs in a foot and ankle population; and second, to assess the extent to which modifiable sociodemographic factors may explain perceived racial and ethnic disparities in outcomes. Increased understanding of the complex relationships between health and social factors may facilitate conscientious use of PROMs and help identify points of intervention for reducing healthcare disparities. Methods: We retrospectively reviewed baseline Foot and Ankle Ability Measure (FAAM) and PROMIS Global-Mental scores of 26,409 foot and ankle patients within our institutional database from 2015 to 2021. Primary predictors included age, sex, race, ethnicity, primary language, education level, median household income, and Charlson Comorbidity Index. All continuous predictors were transformed into categorical variables for analyses. Descriptive analyses, two-tailed T-tests, and Bonferroni post hoc analyses were utilized to assess variation in unadjusted baseline PROM scores stratified by sociodemographic factors. Multivariable modeling was utilized to investigate the independent correlations with and relative importance of the predictors in accounting for variability in outcome scores while adjusting for confounding by all. To examine if the effects of race and ethnicity were modulated by related predictors, we used a series of sequential regression models and evaluated the change in race and ethnicity parameter estimates and R2 values upon including additional predictors in the models. Results: On unadjusted analysis, our data revealed differences in FAAM and PROMIS Global-Mental scores that exceeded the minimal clinically important differences (MCIDs) when stratified by educational level and primary language spoken, but not age, sex, race, ethnicity, median household income, or Charlson Comorbidity Index. Our models accounted for less than 20% of the observed variation in PROM scores. In adjusted models, education level was most prominently associated with baseline scores. Black vs White race had a minor effect in unadjusted models, well below the MCID, and appears to be influenced by differences in education level and household income, which do importantly impact outcomes. Hispanic ethnicity had a larger effect than race in unadjusted analyses and appears to be explained by primary language spoken, education level, and median household income. Conclusion: FAAM and PROMIS Global-Mental baseline scores differed by amounts exceeding the MCID when stratified by educational level and primary language spoken. In adjusted models, race and ethnicity did not independently correlate with clinically significant variations in PROMs scores. Interventions to eliminate disparities among racial and ethnic groups should address modifiable factors such as improving health literacy and removing language and financial barriers to care. To avoid widening health disparities, differences in education level and primary language spoken should be acknowledged when using PROMs as a metric in research, clinical care, or health policy.