Skip to main content
. 2021 May 28;4(5):e2111858. doi: 10.1001/jamanetworkopen.2021.11858

Figure 2. Differences in Adjusted Percentages of Hip and Knee Replacement Use With Comprehensive Care for Joint Replacement (CJR) Model Implementation.

Figure 2.

Differences in adjusted percentages were derived from patient-level multivariable logistic regression models with robust or sandwich estimators of variance (Table 2). The difference for each race/ethnicity dual-eligible (DE) group key represents the percentage point difference in the probability of surgical procedures in the metropolitan statistical areas (MSAs) with CJR model implementation vs those MSAs without CJR. The triple difference (vs non-Hispanic White non–dual-eligible [NDE] beneficiaries) key represents the percentage point difference in the probability of procedures for each race/ethnicity DE group (vs non-Hispanic White NDE group) in MSAs with CJR model implementation vs MSAs without CJR model.

aP < .05.

bP < .01.

cP < .001.