COST-EFFECTIVENESS ANALYSIS OF DIAGNOSTIC TESTS: AN EXAMPLE STUDY OF TRANSESOPHAGEAL ECHOCARDIOGRAPHY TO DIAGNOSE THE PRESENCE OF ASCENDING AORTIC ATHEROSCLEROSIS IN CARDIAC SURGERY PATIENTS

Monday, October 25, 2010
Sheraton Hall E/F (Sheraton Centre Toronto Hotel)
H. Koffijberg, PhD1, B. van Zaane, MD, PhD1, A.P. Nierich, MD, PhD2 and K.G.M. Moons, PhD1, (1)University Medical Center Utrecht, Utrecht, Netherlands, (2)Isala Clinics, Zwolle, Netherlands

Purpose: Most diagnostic tests are (only) evaluated on their discriminative or predictive accuracy. However, such (commonly cross-sectional) accuracy studies do not provide direct evidence on the tests ability to change patient outcomes, let alone on their cost-effectiveness. Recent guidelines suggest that a proper cost-effectiveness analysis of (novel) tests or biomarkers should be done before implementation in practice, using randomised diagnostic strategy studies. We present an example study how the cost-effectiveness of diagnostic tests/markers can be done without such randomised studies but making use of cross sectional accuracy study data and therapeutic intervention studies on patient outcome. The example study concerns post-operative stroke in on-pump cardiac surgery. This is often caused by emboli merging from the atherosclerotic ascending aorta (AA) after manipulation. A new diagnostic transesophageal echocardiography tool showed to accurately determine the presence (and extent) of atherosclerosis prior to sternotomy, such that it allows the surgeon to change surgical strategy to reduce/avoid emboli production. We assessed the tools cost-effectiveness when it would be applied in practice, as compared to care as usual (manual palpation for detection of AA atherosclerosis).

Method: A Markov decision-analytical model, using input from the cross sectional accuracy study and longitudinal patient outcome studies, was developed and used to assess differences in costs and health effects between the two strategies. The incremental cost-effectiveness ratio was calculated for various subgroups of patients. Conservative as well as observed prevalence rates of AA atherosclerosis were defined per subgroup. Probabilistic sensitivity analysis was used to determine the robustness of the model results.

Result: Using the new tool consistently resulted in more adapted procedures and, consequently, in a lower risk of stroke and a (slightly) higher number of life-years. The incremental costs decreased whereas incremental effects increased with patient age. The incremental costs-effectiveness ratio (ICER) ranged from €4937/QALY for 55-year-old men to €-6191/QALY for 75-year-old women.

Conclusion: The new tool reduces costs and increases health benefits in patients older than 65 years. In some subgroups the additional costs will likely be small compared with the additional health benefits. More general: cost-effectiveness estimations of diagnostic tests or biomarkers can be inferred without specifically designed randomised diagnostic studies. Finally, such estimations may guide more efficient designs of randomised diagnostic studies.