Abstract
BACKGROUND: While peripheral nerve blocks (PNBs) are associated with various improved outcomes in patients undergoing total hip or knee arthroplasty (THA/TKA), disparities in PNB utilization have been reported. This study assessed the importance of socioeconomic, demographic, clinical, and hospital determinants in explaining PNB utilization using the population-attributable risk (PAR) framework. Subsequently, we examined the association between PNB use and 3 secondary outcomes: Centers for Medicare and Medicaid Services (CMS)-defined complications, 90-day all-cause readmissions, and length of stay >3 days. METHODS: This retrospective cohort study included 52,926 THA and 94,795 TKA cases from the 5% 2012 to 2021 Medicare dataset. Mixed-effects logistic regression models measured the association between study variables and PNB utilization. Variables of interest were demographic (age, sex), clinical (outpatient setting, diagnosis, prior hospitalizations in the year before surgery, Deyo-Charlson index, obesity, (non)-opioid abuse, smoking), socioeconomic (neighborhood Social Deprivation Index, race and ethnicity) and hospital variables (beds, ownership, region, rurality, resident-to-bed ratio). The model was used for the calculation of variable-specific and variable category-specific PARs (presented in percentages), reflecting the proportion of variation in PNB use explained after eliminating variables (or groups of variables) of interest with all other factors held constant. Subsequently, regression models measured the association between PNB use and secondary outcomes. Associations are presented with odds ratios (ORs) and 95% confidence intervals (95% CIs). RESULTS: Socioeconomic and demographic variables accounted for only a small proportion of variation in PNB use (up to 3% and 7%, respectively). Clinical (THA: 46%; TKA: 34%) and hospital variables (THA: 31%; TKA: 22%) were the primary drivers of variation. In THA, variation by clinical variables was driven by increased PNB use in the inpatient setting (OR, 1.28 [95% CI, 1.07-1.53]) and decreased use in patients with ≥2 prior hospitalizations (OR, 0.72 [95% CI, 0.57-0.90]). Moreover, nonosteoarthritis diagnoses associated with reduced PNB utilization in THA (OR, 0.64 [95% CI, 0.58-0.72]) and TKA (OR, 0.35 [95% CI, 0.34-0.37]).In TKA, PNB use was subsequently associated with fewer complications (OR, 0.82 [95% CI, 0.75-0.90]) and less prolonged length of stay (OR, 0.90 [95% CI, 0.86-0.95]); no association was found for readmissions (OR, 0.98 [95% CI, 0.93-1.03]). In THA, associations did not reach statistical significance. CONCLUSIONS: Among THA and TKA patients on Medicare, large variations exist in the utilization of PNBs by clinical and hospital variables, while demographic and socioeconomic variables played a limited role. Given the consistent benefits of PNBs, particularly in TKA patients, more standardized provision may be warranted to mitigate the observed variation.