TY - JOUR T1 - OP24 Assessing the Cost-Effectiveness of Alternative Care Pathways: A Case Study Evaluating Early Transfer to Neuroscience Centres for Critically ill Patients with Acute Traumatic Brain Injury JF - Journal of Epidemiology and Community Health JO - J Epidemiol Community Health SP - A10 LP - A10 DO - 10.1136/jech-2012-201753.024 VL - 66 IS - Suppl 1 AU - R Grieve AU - D Harrison AU - MZ Sadique AU - M Gomes AU - D Menon AU - K Rowan Y1 - 2012/09/01 UR - http://jech.bmj.com/content/66/Suppl_1/A10.1.abstract N2 - Background Acute traumatic brain injury (TBI) is a major cause of death, disability and cost. For critically-ill adult TBI patients who present outside a neuroscience centre, and do not require neurosurgery, there is little evidence on whether early transfer to a neuroscience centre is worthwhile. NICE guidelines list the issue as a key research topic. We aimed to assess the relative cost-effectiveness of ‘early transfer’ (within 18 hours of hospital presentation) versus ‘no or late transfer’ (after 24 hours) for these patients. Methods The Risk Adjustment In Neurocritical care (RAIN) Study validated risk prediction models following TBI. The RAIN Study recruited admissions following acute TBI to 67 adult critical care units during 2009–11. Detailed information was collected on baseline prognostic factors, the time of transfer to neuroscience centres, and mortality. Resource use data were recorded for six months, and combined with unit costs to report total hospital and community health service costs. At the six month follow-up, data were collected on health-related quality of life (HRQoL), by the EQ–5D–3L. The lifetime cost-effectiveness analysis extrapolated from six month endpoints, informed by the literature. We report lifetime incremental cost per QALY of ‘early’ versus ‘no or late transfer’, overall and for subgroups (age <=70, vs >70; mild/moderate, vs severe TBI; major extracranial injury, vs none). To adjust for baseline differences we used the previously validated risk prediction models in regression analyses. In sensitivity analyses we considered alternative approaches for extrapolating from the six month endpoints and undertaking risk adjustment. Results There were 584 patients in the ‘early’ and 263 in the ‘no or late transfer’ group. After risk adjustment, early transfer was associated, at six months, with lower mortality (odds ratio 0.52, 95% CI 0.34 to 0.80), higher HRQoL for survivors (mean gain 0.13, 0.032 to 0.225), but positive incremental costs (£15,000, £11,123 to £18,880). The lifetime cost per QALY for ‘early transfer’ was £11,000. For patients older than 70, ‘early transfer’ was associated with higher mortality, and was unlikely to be cost-effective (probability 0.15 at £20,000 per QALY). For other subgroups, the corresponding probabilities that ‘early transfer’ is cost-effective were between 0.7 and 1. Conclusion For critically-ill patients with acute TBI aged 70 or less, early transfer to a neuroscience centre appears cost-effective. While this finding is robust to alternative methodological assumptions and choice of risk prediction model, further research is required to investigate the potential impact of unobserved confounding. ER -