Presentation (CTI)

OBJECTIVES: Modelling cost effectiveness (CE) of test and treat strategies is complex and challenging, requiring the incorporation of all data inputs relevant to the characteristics of the diagnostic test such as cost, biomarker prevalence and diagnostic accuracy, together with assumptions on survival estimates for true and false positive (TP and FP) and true and false negative (TN and FN) patients. This research explores if model complexity can be reduced to a more simplified structure, minimising data requirements and additional assumptions.
METHODS: We analysed the equations that estimate the probability, costs and Quality-Adjusted Life Years (QALYs) of each outcome node of a typical decision tree (DT) model that estimates the CE of a ‘biomarker test’ strategy versus a ‘no test’ strategy. Through algebraic reduction we derive a simplified yet equivalent expression for the incremental cost effectiveness ratio (ICER). In a hypothetical example we compare the ICER obtained using this approach versus one using the more complex model.
RESULTS: Based on the inherent symmetry within the model structure, the equation to calculate the ICER of a test vs no test strategy reduces to a simple mathematical expression. Only two parameters of the diagnostic test are required: test positive rate and cost. In the hypothetical example, assuming the weighted survival outcome of TP+FP patients is the same as that observed in the study on which the CE analysis is based, the ICER was estimated to be £71k/QALY using both methods.
CONCLUSIONS: The data requirements to model the CE of a diagnostic test and treat strategy can be significantly reduced and the model structure simplified, further obviating the need to make additional assumptions around survival and costs for FP, TP, FN and TN patients. This may have implications for countries that develop guidelines for methods of health economic evaluation specifically for co-dependent technologies.