Risk-Adjusted Excess Length of Stay for Patients With Heart Failure Across Facilities: A Large US Cohort Study

Area of Science:

• Health Services Research
• Clinical Informatics
• Health Economics

Background:

• Hospital length of stay (LOS) for heart failure (HF) impacts patients, payers, and providers.
• Risk adjustment of LOS is crucial for fair reimbursement, especially for facilities serving complex or marginalized populations.
• Existing methods often require extensive clinical data, limiting broad applicability.

Purpose of the Study:

• To assess sociodemographic, clinical, and facility-level factors influencing HF LOS.
• To develop and validate a risk-adjustment model for HF LOS using commonly available data.
• To explore facility-specific differences in risk-adjusted LOS.

Main Methods:

• A Fine and Gray Cox proportional hazards model was applied to a large US cohort (89,621 HF patients).
• The model controlled for censoring and competing risks of in-hospital mortality.
• Risk adjustment incorporated patient-level (sociodemographic, clinical) and facility-level factors.

Main Results:

• Sociodemographic, clinical, and facility characteristics significantly explained variations in HF LOS.
• The risk-adjustment model demonstrated moderate performance (C-statistics ≈ 0.69) in both derivation and validation cohorts.
• Facility-level factors contributed to idiosyncratic differences in risk-adjusted LOS.

Conclusions:

• Multiple factors at the patient and facility levels contribute to excess HF LOS.
• A novel HF LOS risk-adjustment method is presented, requiring less extensive clinical data.
• This approach supports more equitable facility performance assessment and reimbursement.