Abstract
Introduction
Bundle payment models have clinical and economic impacts on providers. Despite efforts made to improve care, experience has shown that a few episodes with costs well above a target (bundle busters) can reduce or negate positive performances. The purpose of this study was to identify both the primary episode drivers of cost and patient factors that led to episodes above target.
Methods
A retrospective study of 10,000 joint replacement episodes from a large healthcare system in CJR and a private orthopedic practice in BPCI was conducted. Episodes with costs greater than target price (TP) were designated as bundle busters and sub-divided into 4 groups:
-
1)
< 1 standard deviation (SD) above TP (n=1700)
-
2)
> 1 to 2 SD above TP (n=240)
-
3)
> 2 to 3 SD above TP (n=70)
-
4)
> 3 SD above TP (n=70)
Bundle busters were compared to the control that were at/below the TP (n= 7500). For the CJR/BPCI cohorts, one SD was defined as $10,700/$13,000, respectively.
Two linear regressions assessed the likelihood of factors predicting a bundle buster and the total episode cost. These variables included demographics, acuity classifications, comorbidities, length of stay, readmissions, discharge disposition, post-acute utilization, and episode costs.
Results
Group 1 had 66% higher total episode costs than the control group, 13% higher hospital costs, and 144% higher SNF costs. Costs for Group 2 were at least double that of Group 1.
Significant predictors for bundle busters included, race, age, comorbidities, BMI, and LOS. (P<0.01) Significant predictors for total episode costs included, gender, age, ROM, SOI, comorbidities, and BMI. (P<0.1).
Discussion
Age, Comorbidities, and BMI are among significant predictors of higher total episode costs. Identifying drivers and factors that impact bundle busters supports the importance of care model improvements that mitigate risks associated with these patients.
