Stroke care in Maine: Measuring quality of care with all-payer claims data

Background: Ongoing analysis of stroke care is vital to improving quality and outcomes. However, few of Maine's 36 acute care hospitals currently collect or report such data. Objectives: Assess feasibility of using medical claims data to: measure quality of care for individuals hospitalized for stroke; Establish state-wide baselines; Inform ongoing quality improvement efforts. Data/Methods: An analytic file of inpatient stays (n=12,202)was constructed using all 2003-2006 Medical and pharmacy claims for cerebrovascular diagnoses (ICD-9 430-436) paid by private insurers and Medicare. Comparisons of quality metrics were made by diagnosis, year, age, gender, and location. Confidence intervals were calculated for proportions in frequency tables and chi-square p-values for 2x2 crosstabs. Metrics derived from claims were compared with national data including CDC data and American Stroke Association's registry-based Get With the Guidelines-Stroke (GWTG) databases. Results: Maine claims data were comparable to GWTG (shown below as Maine claims/GWTG) for: • Stroke diagnoses: ischemic 60%/66%; hemorrhagic 11%/7%; TIA 29%/28% • EMS: 53.9% /54% stroke admissions had corresponding EMS claim • Thrombolytic use for ischemic admissions: 2.9%/2.9% Northeast Cerebrovascular Consortium; • Length of stay: 5.21/5.3 days; • Mortality: 7.2%/6.6% Comparisons of medications on discharge, diagnostics, neurological visits and surgical procedures were considered unreliable due to discrepancies or lack of comparable measures between sources. Conclusions: Claims data can support studies of state-wide stroke diagnosis and treatment quality and be utilized to plan ongoing systems improvements efforts.