Stroke Patients Not Worse Off With Generalist Care (Depending on Your Analysis)

Although observational studies suggest that specialty care leads to better inpatient outcomes for certain conditions—like ischemic stroke—than general care, these studies have been confounded by prognostic differences among admitted patients. In a recently published retrospective study, investigators at UCSF found that, when adjustments were made to avoid a selection bias for specialist-versus-generalist admission, the rate of in-hospital mortality after ischemic stroke was not significantly lower with primary neurologic care.

When assessing the outcomes of 26,925 inpatients with ischemic stroke in a database of 113 US academic hospitals (1997-1999), Gillum and Johnston attempted to avoid confounding by applying a grouped-treatment analysis, which is used to evaluate individual outcome differences associated with practice variability among hospitals. According to these investigators, this analysis has been previously applied to health-services research (McClellan M et al. JAMA. 1994;272:859-866).

In their assessment, the authors discovered 1760 (7%) in-hospital deaths and, in the univariate and multivariate analyses, a significantly lower risk of these deaths among patients admitted by neurologists. However, there was no difference in the mortality rates among hospitals admitting different proportions of patients with ischemic stroke to neurologists (ie, grouped-treatment analysis).

Analysis Type	Odds Ratio of In-Hospital Mortality With Neurologist Care	Significance
Univariate	0.32	P < .001
Multivariate	0.60	P < .001
Grouped-treatment	1.02	NS

The authors conclude that observed differences in ischemic-stroke outcomes between neurologists and generalists may be due to differences in initial prognosis, given that mortality was not lower at hospitals where patients were admitted more frequently to neurologists.

Image of reflex hammer from iStockPhoto.