Risk adjustment in coronary bypass grafting: how EuroSCORE is related to cost, health related quality of life, and cost-effectiveness.
Presenter: unto Häkkinen, stakes
The aim of this study was to evaluate how EuroSCORE (the European System for Cardiac Operative Risk Evaluation) predicts short and long term cost and outcomes of CABG patients. We analysed the predictive power of EuroSCORE on various cost and outcome measures and evaluated which factors - in addition to the original EuroSCORE - affected the measures. We evaluated how patients' risk scores affected QALYs gained and cost per QALY gained. We assessed also the bias in cost and QALY estimates that are due to the fact that HRQoL information is usually available for specific patients that do not represent the total patient population.
We studied prospectively first-time consecutive coronary bypass patients operated in Helsinki University Hospital between 09122000 -21122001. The patient level risk score data was collected preoperatively from almost every patient. HRQoL was measured with 15D. It is a generic, standardized, self administrated instrument that can be used both as a profile and as a single index score measure. The patient level cost data on surgery hospital admission was based on cost accounting system of Helsinki University Hospital, which is derived "bottom-up approach" and is as such very accurate. In addition, with unique identification numbers various register data were linked in the data base. This allows us to follow patients(cost and survival) for five years.
We evaluated the performance of the risk system using various multivariate analysis methods. Since our analysis is based before and after comparison many important assumptions should be made in order to evaluate the incremental cost effectiveness (CE) ratio (=Δ cost/ Δ QALY). Thus we calculated CE ratio by five cost and two QALY specifications.
According to the results EuroSCORE associated quite well with costs, various mortality indicators and life expectancy but not with HRQoL. In addition variables included in EuroSCORE, also pervious year cost and diabetes were significant additional "risk factors". We found that change in HRQoL was heavily dependent on preoperative HRQoL status. CE ratio was crucially depending on measurement of QALYs and specially on the assumptions on the effects of treatment on life expectancy. If the operation affects more on life expectancy of high risk patients than low risk patients, the cost per QALY difference between EuroSCORE groups will convergence. The cost per QALY figures derived from selected samples will overestimate positive results.
Authors: Unto Häkkinen, Tuula Kurki, Antti Vento, Mikko Peltola
Session: Risk Adjustment and Costs
Time: Tue 10 a.m.-11 a.m.