Readmissions after stroke: linked data from the Australian Stroke Clinical Registry and hospital databases.

OBJECTIVES

To assess the feasibility of linking a national clinical stroke registry with hospital admissions and emergency department data; and to determine factors associated with hospital readmission after stroke or transient ischaemic attack (TIA) in Australia.

DESIGN AND SETTING

Data from the Australian Stroke Clinical Registry (AuSCR) at a single Victorian hospital were linked to coded, routinely collected hospital datasets for admissions (Victorian Admitted Episodes Dataset) and emergency presentations (Victorian Emergency Minimum Dataset) in Victoria from 15 June 2009 to 31 December 2010, using stepwise deterministic data linkage techniques.

MAIN OUTCOME MEASURES

Association of patient characteristics, social circumstances, processes of care and discharge outcomes with all-cause readmissions within 1 year from time of hospital discharge after an index admission for stroke or TIA.

RESULTS

Of 788 patients registered in the AuSCR, 46% (359/781) were female, 83% (658/788) had a stroke, and the median age was 76 years. Data were successfully linked for 782 of these patients (99%). Within 1 year of their index stroke or TIA event, 42% of patients (291/685) were readmitted, with 12% (35/286) readmitted due to a stroke or TIA. Factors significantly associated with 1-year hospital readmission were two or more presentations to an emergency department before the index event (adjusted odds ratio [aOR], 1.57; 95% CI, 1.02-2.43), higher Charlson comorbidity index score (aOR, 1.19; 95% CI, 1.07-1.32) and diagnosis of TIA on the index admission (aOR, 2.15; 95% CI, 1.30-3.56).

CONCLUSIONS

Linking clinical registry data with routinely collected hospital data for stroke and TIA is feasible in Victoria. Using these linked data, we found that readmission to hospital is common in this patient group and is related to their comorbid conditions.