Abstract 17472: Impact of Do-Not-Resuscitate Status on Hospital Risk-adjusted Mortality for Heart Failure
Introduction: Risk-adjusted mortality following heart failure hospitalization is used by payers to adjust reimbursement and incentivize improvements. This risk-adjusted measure does not account for variation in hospital rates of do not resuscitate (DNR) on admission, which may impact measure validity.
Methods: California Healthcare Cost and Utilization Project data identifies patients with DNR within 24 hours of admission. We included all adults admitted with primary diagnosis of heart failure during 2011. We determined hospital-level in-hospital mortality, before and after accounting for admission DNR, from hierarchical logistic regression adjusting for Hierarchical Clinical Condition variables. We generated hospital risk-adjusted Predicted-to-Expected mortality ratios (P/E ratios). We determined the effect of accounting for DNR on patient-level mortality risk, and on changes in hospital P/E ratio quartiles.
Results: Among 55,865 hospitalized heart failure patients at 290 hospitals, the overall DNR rate was 12.1% (6,755). Median hospital DNR rate was 8.7% (IQR 3.9% to 20.0%). Median hospital mortality rate was 2.9% (IQR 1.8 - 4.1). At the patient-level, DNR status was associated with increased in-hospital mortality (OR 3.50, 95% CI 3.11 - 3.94, p<0.001). Hospital DNR rates and adjusted P/E ratios were positively correlated (0.204, p<0.001), and did not correlate after adjusting for DNR (-0.079, p<0.18), see Figure. Among 72 hospitals in the lowest P/E quartile before accounting for DNR, 11 (15.3%) hospitals reclassified to a higher quartile after inclusion of DNR. Among 72 hospitals in the highest P/E quartile before accounting for DNR, 14 (19.4%) hospitals reclassified to a lower quartile after inclusion of DNR.
Conclusion: DNR on admission rates vary widely between hospitals and DNR predicts in-hospital mortality. Adjustment for DNR alters hospital risk-adjusted mortality classification and may be useful for risk-adjusted mortality measures.
Author Disclosures: J.T. Bruckel: None. A. Mehta: None. B.K. Nallamothu: Research Grant; Significant; NIH (1R01HL123980), VHA-HSRD (IIR 13-079). S. Thomas: None. S.M. Bradley: None. A.J. Walkey: None.
- © 2016 by American Heart Association, Inc.