Table 2. Association of the Comprehensive Care for Joint Replacement (CJR) Model With the Probabilities of Total Hip or Total Knee Replacement, 2013-2017a.
| Variable | Unadjusted probabilities, % (N = 18 403 141 observations; 171 MSAs) | Adjusted probabilities, % (N = 18 403 141 observations; 171 MSAs) | ||||||
|---|---|---|---|---|---|---|---|---|
| MSAs with CJR modelb | MSAs without CJR modelc | Percentage point change in MSAs with vs without CJR model (95% CI)d | Percentage point change with respect to White NDE beneficiaries (95% CI)e | |||||
| Before implementation (95% CI) | After implementation (95% CI) | Percentage point change (95% CI) | Before implementation (95% CI) | After implementation (95% CI) | Percentage point change (95% CI) | |||
| THR | ||||||||
| Race/ethnicity by dual eligibility | ||||||||
| Non-Hispanic White NDE beneficiaries | 1.25 (1.24 to 1.26) | 1.31 (1.30 to 1.33) | 0.07 (0.05 to 0.09)f | 1.24 (1.23 to 1.25) | 1.32 (1.30 to 1.33) | 0.08 (0.06 to 0.09)f | 0.05 (0.01 to 0.09)g | 1 [Reference] |
| Non-Hispanic White DE beneficiaries | 0.31 (0.30 to 0.32) | 0.33 (0.31 to 0.36) | 0.02 (−0.004 to 0.05) | 0.40 (0.39 to 0.42) | 0.47 (0.44 to 0.50) | 0.07 (0.03 to 0.10)f | −0.02 (−0.08 to 0.05) | −0.06 (−0.14 to 0.01) |
| Non-Hispanic Black NDE beneficiaries | 0.73 (0.70 to 0.76) | 0.76 (0.71 to 0.80) | 0.03 (−0.03 to 0.08) | 0.64 (0.62 to 0.66) | 0.70 (0.66 to 0.74) | 0.06 (0.01 to 0.11)g | −0.002 (−0.10 to 0.10) | −0.05 (−0.16 to 0.06) |
| Non-Hispanic Black DE beneficiaries | 0.28 (0.26 to 0.31) | 0.33 (0.29 to 0.38) | 0.05 (−0.002 to 0.10) | 0.27 (0.25 to 0.29) | 0.31 (0.27 to 0.35) | 0.04 (−0.01 to 0.09) | 0.06 (−0.05 to 0.17) | 0.01 (−0.11 to 0.13) |
| Hispanic NDE beneficiaries | 0.35 (0.29 to 0.41) | 0.39 (0.29 to 0.50) | 0.05 (−0.07 to 0.17) | 0.26 (0.21 to 0.31) | 0.40 (0.30 to 0.51) | 0.14 (0.02 to 0.26)g | 0.05 (−0.24 to 0.33) | −0.001 (−0.29 to 0.29) |
| Hispanic DE beneficiaries | 0.13 (0.11 to 0.15) | 0.14 (0.11 to 0.17) | 0.01 (−0.03 to 0.05) | 0.11 (0.08 to 0.13) | 0.18 (0.13 to 0.24) | 0.08 (0.01 to 0.14)g | −0.09 (−0.24 to 0.05) | −0.14 (−0.29 to 0.01) |
| TKR | ||||||||
| Race/ethnicity by dual eligibility | ||||||||
| Non-Hispanic White NDE beneficiaries | 2.28 (2.26 to 2.29) | 2.38 (2.36 to 2.41) | 0.11 (0.08 to 0.14)f | 2.44 (2.43 to 2.45) | 2.47 (2.45 to 2.49) | 0.03 (0.003 to 0.05)g | 0.10 (0.05 to 0.15)f | 1 [Reference] |
| Non-Hispanic White DE beneficiaries | 0.77 (0.75 to 0.80) | 0.81 (0.77 to 0.85) | 0.04 (−0.01 to 0.08) | 0.96 (0.94 to 0.99) | 0.98 (0.94 to 1.03) | 0.02 (−0.03 to 0.07) | 0.11 (0.004 to 0.21)g | 0.01 (−0.10 to 0.12) |
| Non-Hispanic Black NDE beneficiaries | 1.65 (1.61 to 1.69) | 1.63 (1.56 to 1.70) | −0.02 (−0.10 to 0.06) | 1.52 (1.49 to 1.56) | 1.68 (1.62 to 1.74) | 0.15 (0.08 to 0.22)f | −0.15 (−0.29 to −0.01)g | −0.25 (−0.40 to −0.10)h |
| Non-Hispanic Black DE beneficiaries | 0.74 (0.70 to 0.78) | 0.68 (0.61 to 0.75) | −0.06 (−0.14 to 0.01) | 0.70 (0.66 to 0.73) | 0.79 (0.72 to 0.85) | 0.09 (0.01 to 0.16)g | −0.18 (−0.34 to −0.01)g | −0.27 (−0.45 to −0.10)h |
| Hispanic NDE beneficiaries | 1.30 (1.18 to 1.41) | 1.41 (1.21 to 1.61) | 0.12 (−0.11 to 0.35) | 1.52 (1.40 to 1.64) | 1.82 (1.60 to 2.05) | 0.31 (0.05 to 0.56)g | 0.05 (−0.55 to 0.66) | −0.04 (−0.65 to 0.57) |
| Hispanic DE beneficiaries | 0.90 (0.86 to 0.95) | 1.06 (0.96 to 1.15) | 0.15 (0.05 to 0.26)h | 1.38 (1.29 to 1.47) | 1.37 (1.22 to 1.52) | −0.01 (−0.19 to 0.16) | 0.32 (−0.001 to 0.64) | 0.23 (−0.10 to 0.55) |
Abbreviations: DE, dual-eligible; MSA, metropolitan statistical area; NDE, non–dual-eligible; THR, total hip replacement; TKR, total knee replacement.
Adjusted probabilities (expressed in percentages) from patient-level multivariable logistic regression models with robust or sandwich estimators of variance. The regression models assessed the association of the CJR model with the probability of surgical procedures for each race/ethnicity dual-eligibility group (vs non-Hispanic White NDE beneficiaries) in MSAs with vs without the CJR model. The regression models controlled for age, sex, comorbidities, calendar year (and interactions with indicators in MSAs with the CJR model and with race/ethnicity dual-eligibility indicator), MSA fixed effects, and MSA weights. The analysis excluded data from 2016 because the CJR model was introduced in April 2016, which precluded the classification of Medicare beneficiaries into pre- and post-CJR model cohorts. For the adjusted columns, the probabilities and changes in probabilities were obtained using Stata margins and lincom commands (StataCorp LLC).
Presents the unadjusted probabilities (expressed in percentages) of surgical procedures in MSAs with CJR model implementation and the percentage point differences with CJR model implementation.
Presents the unadjusted probabilities (expressed in percentages) of surgical procedures in MSAs without CJR model implementation and the percentage point differences with CJR model implementation.
Presents the percentage point differences in the probability of surgical procedures for each race/ethnicity dual-eligibility group in MSAs with vs without CJR model implementation.
Presents the percentage point differences in the probability of surgical procedures for each race/ethnicity dual-eligibility group (vs non-Hispanic White NDE beneficiaries) in MSAs with vs without CJR model implementation (triple difference).
P < .001.
P < .05.
P < .01.