Introduction:Health state utilities measured by the generic multi-attribute utility instruments (MAUIs) differ. Empirical evidence suggests that some MAUIs are more sensitive than others in reflecting the quality of life (QoL) of patients in particular disease areas. Additionally, in order to estimate utilities based on cancer-specific health-related quality of life instruments (CSQoLs), a number of mapping functions have emerged. Although it is common practice to apply a CSQoL instead of a MAUI in clinical trials, CSQoL cannot be used to estimate utility values for economic evaluations. Mappings based on MAUIs that are not sensitive to changes in cancer patients’ QoL may result in misleading approximations of utilities that could affect allocation of resources. The study objective is to explore the validity and sensitivity of the major MAUIs to variation in the QoL measured by cancer-specific instruments. We aimed to investigate (i) the sensitivity of the general MAUIs scores to changes in the CSQoL, and (ii) whether particular dimensions of the general instrument are more sensitive.Methods:A two stage systematic literature review is conducted. First, an update of the review done by McTaggart-Cowan et al. (2013) on the mapping methods used to determine utilities from cancer-specific instrument. Second, an analysis of studies that measure the relationship between CSQoLs and general MAUIs.Results:The literature suggests that differences exist between MAUIs in their capacity to capture the QoL dimensions of the CSQoLs. Additionally, the main challenge to build an appropriate mapping function for deriving utilities values from CSQoL is the definition of an appropriate methodology that (i) responds to the distribution of the selected sample and (ii) can successfully be validated in additional samples.Conclusions:In the context of health technology assessment and cost effectiveness analysis, it is crucial to carefully select and report the CSQoL and MAUI involved in the estimation of the additional benefits. Policy makers need to be awarded of the sensitivity of the instruments to changes in QoL in relation to the CSQoL dimensions QoL.

Breast cancer (BC) is a heterogeneous disease representing a substantial economic burden. In order to develop policies that successfully decrease this burden, the factors affecting costs need to be fully understood. Evidence suggests that early detection in Stage I has a lower cost than late detection. We aim to provide conservative estimates of BC's stage-wise medical costs from German healthcare and the payer's perspective. To this end, we conducted a literature review of articles evaluating stage-wise costs of BC in Germany through PubMed, Web of Science, and Econ Lit databases supplemented by Google Scholar. We developed a decision tree model to estimate BC related medical costs in Germany using available treatment and cost information. The review generated seven studies; none estimated the stage-wise costs of BC. The studies were classified into two groups: (1) case scenarios (five studies) and two studies based on administrative data. The first sickness funds data study (Gruber, Stock, et al. 2012) used 1999 information to approach BC attributable cost; their results suggest a range between €3,929 and €11,787 depending on age. The second study (Kreis, Plöthner, et al. 2020) used 2011-2014 data and suggested an initial phase incremental cost of €21,499, an intermediate phase cost of €2,620, and a terminal phase cost of €34,513 per incident case. Our decision tree model based BC stage-wise cost estimates were €21,523 for Stage I, €25,679 for Stage II, €30,156 for Stage III, €42,086 for Stage IV. Alternatively, the modeled cost estimates are €20,284 for the initial phase of care, €851 for the intermediate phase of care, and €34,963 for the terminal phase of care. Our estimates for phases of care are consistent with recent German estimates provided by Kreis and Plöthner et al. Furthermore, the data collected by sickness funds are collected primarily for reimbursement purposes, where the German ICD-10 classification system defines a cancer diagnosis. As a result, claims data lack the clinical information necessary to understand stage-wise BC costs. Our model-based estimates fill the gap and inform future economic evaluations of BC interventions.