Consider stroke severity when assessing hospital performance

Incorporating stroke severity into a 30-day mortality risk model based on claims data for Medicare beneficiaries improved prediction of risk for death and allowed more accurate assessment of hospital performance when evaluating outcomes for ischemic stroke patients, according to new data.

Gregg C. Fonarow, MD, of the University of California, Los Angeles, and colleagues analyzed data on 127,950 fee-for-service Medicare beneficiaries (median age, 80 years) with ischemic stroke from 782 hospitals participating in Get With The Guidelines-Stroke. Mean NIH Stroke Scale (NIHSS) documented between 2003 and 2009 was 8.23. The researchers evaluated performance of claims-based hospital mortality risk models with and without inclusion of NIHSS scores for 30-day mortality and compared hospital rankings from both models.

Within the first 30 days in the study cohort, 18,186 deaths occurred, including 7,430 deaths during the index hospitalization. Median hospital-level 30-day mortality was 14.5%.

Results revealed significantly better discrimination with the hospital mortality model, including NIHSS scores, when compared with the model without NIHSS scores (P<.001). The researchers found that 26.3% of hospitals ranked in the top 20% or bottom 20% of performers based on the model without NIHSS scores were ranked differently when evaluated according to the model, including NIHSS scores. Further, 57.7% of hospitals originally classified as having “worse than expected” mortality were reclassified to “as expected” after accounting for NIHSS scores.

Data also showed that net reclassification improvement and integrated discrimination improvement indexes demonstrated significant enhancement of model performance after addition of NIHSS scores.

In an accompanying editorial, Tobias Kurth, MD, ScD, of the University of Bordeaux, France, and Mitchell S.V. Elkind, MD, MS, of Columbia University, New York, said excluding information on stroke severity could lead to incorrect ranking of hospital performance and should be examined in a different context than other CVD.

“Ischemic stroke is a much more heterogeneous condition than ischemic heart disease and is characterized by multiple subtypes, etiologies and diverse outcomes,” they wrote. “The assumption that what is true of MI is also true of stroke, therefore, is flawed, as the present data underscore. The particular characteristics of stroke have to be taken into consideration by clinicians, insurance companies and policy makers when comparing disease-specific health outcomes.”

For more information:

Fonarow GC. JAMA. 2012;308:257-264.

Kurth T. JAMA. 2012;308:292-293.

Disclosure: The Get With The Guidelines-Stroke program is currently supported in part by a charitable contribution from Janssen Pharmaceuticals and funded in the past through support from Boehringer-Ingelheim, Merck and Bristol-Myers Squibb/Sanofi Pharmaceuticals Partnership. See the study for a full list of researchers’ disclosures. Dr. Elkind reported receiving research support from diaDexus, Bristol-Myers Squibb/Sanofi Pharmaceuticals Partnership; serving on an event adjudication committee for Jarvik Heart; having received compensation for participation in litigation on behalf of Novartis, Organon and GlaxoSmithKline; and receiving compensation from the American Academy of Neurology for service as Resident and Fellow Section Editor for Neurology. Dr. Kurth reported receiving investigator-initiated research funding from the French National Research Agency, the NIH, the Migraine Research Foundation and the Parkinson’s Disease Foundation, and having received honoraria from the BMJ for editorial services; from Allergan, the American Academy of Neurology and Merck for educational lectures; and from MAP Pharmaceutical for contributing to a scientific advisory panel.