Background and purpose: To investigate the 30-day and 12-month mortality risks among hospitalized stroke patients according to compliance with guideline-based process indicators.
Methods: We used data from the Second Stroke Audit and the Mortality Register of Catalonia (Spain). The audit retrospectively explored quality of stroke care based on compliance with indicators among patients discharged from all public hospitals in Catalonia in 2007; they were identified and selected through a pre-established sampling method. The magnitude of the independent association of each indicator with 30-day and 12-month mortality was assessed using logistic regression with generalized estimating equations to account for clustering of patients within hospitals. Generalized estimating equations modeling was initially restricted to patients alive >72 hours poststroke to control for confounding by severity. Analyses were also run in 3 other samples (all patients, patients alive >7 days, and patients alive >14 days).
Results: Of 1767 stroke admissions in the Second Stroke Audit, 1697 patients survived >72 hours poststroke. Within this sample, the adjusted 30-day mortality risk was negatively associated with nonadherence to different indicators, of which only antithrombotics at discharge (OR, 4.3; 95% CI, 1.72-10.78) remained significant in all data sets. At 12 months, the adjusted mortality risk was negatively associated with management of hypertension (OR, 1.87; 95% CI, 1.22-2.86) and antithrombotics at discharge (OR, 2.79; 95% CI, 1.41-5.54). Both remained unchanged across different samples.
Conclusions: Assessing the impact of quality of stroke care on mortality is complex and is hampered by residual confounding, particularly in the short-term. Nevertheless, this study suggests that at least a few indicators should be used to monitor quality of stroke services.