Refining Risk Adjustment for the Proposed CMS Surgical Hip and Femur Fracture Treatment Bundled Payment Program

The U.S. Centers for Medicare & Medicaid Services (CMS) has been considering the implementation of a mandatory bundled payment program, the Surgical Hip and Femur Fracture Treatment (SHFFT) model. However, bundled payments without appropriate risk adjustment may be inequitable to providers and may restrict access to care for certain patients. The SHFFT proposal includes adjustment using the Diagnosis-Related Group (DRG) and geographic location. The goal of the current study was to identify and quantify patient factors that could improve risk adjustment for SHFFT bundled payments.

We retrospectively reviewed a 5% random sample of Medicare data from 2008 to 2012. A total of 27,898 patients were identified who met SHFFT inclusion criteria (DRG 480, 481, and 482). Reimbursement was determined for each patient over the bundle period (the surgical hospitalization and 90 days of post-discharge care). Multivariable regression was performed to test demographic factors, comorbidities, geographic location, and specific surgical procedures for associations with reimbursement.

The average reimbursement was $23,632 ± $17,587. On average, reimbursements for male patients were $1,213 higher than for female patients (p < 0.01). Younger age was also associated with higher payments; e.g., reimbursement for those ≥85 years of age averaged $2,282 ± $389 less than for those aged 65 to 69 (p < 0.01). Most comorbidities were associated with higher reimbursement, but dementia was associated with lower payments, by an average of $2,354 ± $243 (p < 0.01). Twenty-two procedure codes are included in the bundle, and patients with the 3 most common codes accounted for 98% of the cases, with average reimbursement ranging from $22,527 to $24,033. Less common procedures varied by >$20,000 in average reimbursement (p < 0.01). DRGs also showed significant differences in reimbursement (p < 0.01); e.g., DRG 480 was reimbursed by an average of $10,421 ± $543 more than DRG 482. Payments varied significantly by state (p ≤ 0.01). Risk adjustment incorporating specific comorbidities demonstrated better performance than with use of DRG alone (r = 0.22 versus 0.15).

Our results suggest that the proposed SHFFT bundled payment model should use more robust risk-adjustment methods to ensure that providers are reimbursed fairly and that patients retain access to care. At a minimum, payments should be adjusted for age, comorbidities, demographic factors, geographic location, and surgical procedure.