Resource costs, health outcomes and cost-effectiveness in stroke care: Evidence from the Oxford Vascular Study

Introduction: Cerebrovascular events are a major cause of mortality, disability and healthcare resource use. Despite this, there is a lack of reliable information on their costs and outcomes, particularly related to transient ischaemic attacks (TIA) and minor stroke. Such information is vital to inform decisions about local and national service provision, and to provide reliable estimates for use in cost-effectiveness analyses. Aims This thesis estimates the costs and outcomes of stroke and TIA using data from a population-based study undertaken in a population of over 91,000 individuals in Oxfordshire (the Oxford Vascular Study – OXVASC). In addition, the thesis aims to estimate the short-term cost-effectiveness of two secondary stroke prevention programmes evaluated in a study nested within OXVASC. Methods: Using multiple methods of case ascertainment, 1,282 patients were identified as having suffered a stroke or TIA, of which 1,199 (723 stroke and 476 TIA) patients consented to the study. Follow-up of patients took place at 1, 6, 12 and 24 months, with data collected on patients’ disability, medication usage, living arrangements, and quality of life. Healthcare resource use information was derived from hospital and primary care records, and priced using published unit costs. Findings: Stroke patients had higher case-fatality rates than TIA patients (15% vs. 1%; p<0.001), with 5-year life expectancy being one year longer for TIA patients. For stroke and TIA survivors, the risk of disability remained higher, at around 30% at each of the four follow-ups, than at baseline (17%; p<0.001 for all follow-ups). After standardising for age and gender, average quality of life for stroke and TIA patients combined was significantly lower than English population norms (p<0.001 for all follow-ups). However, when quality of life was compared to population norms by event type, quality of life differences between TIA patients and English population norms no longer remained statistically significant. Important predictors of quality of life included event severity, baseline disability and recurrent vascular events. Total costs were considerably higher 1-year after the initial stroke or TIA than for the year preceding it and, except for day cases, increases were observed for all resource-use categories. Five years after the index event, stroke patients incurred costs of £16,923 (95% CI: 15,149 to 18,858) per patient, significantly higher than those incurred by TIA patients, at £13,904 (95% CI: 11,488 to 16,657; p=0.019). In multivariate analyses, event severity was found to be a significant predictor of inpatient care resource use and costs, as were the presence of recurrent vascular events, especially stroke and coronary events. For non-hospitalised patients, results showed that urgent outpatient specialist assessment and treatment reduced the 90-day risk of fatal or disabling stroke (0.4% vs. 5%, p<0.001) compared with less urgent assessment and treatment. In terms of resource usage, patients who were assessed and treated urgently had lower recurrent stroke hospitalisation (2% vs. 8%; p=0.001), and reduced overall number of days in hospital (average reduction of 4 days; p=0.017). These reductions in hospital resource usage generated savings of £643 per patient assessed and treated urgently in an outpatient clinic (p=0.028). Conclusion: Despite the impact of stroke on death, disability and healthcare resource use, there is a lack of reliable information on costs and outcomes, especially for TIA and minor stroke. Through the use of a population-based study, the gold-standard study design when assessing the incidence and outcomes of TIA and stroke, this thesis provides healthcare decision makers and researchers with a wealth of data on the resource use patterns, costs and outcomes of TIA and stroke patients, and their main predictors