OBJECTIVES: Myelodysplastic syndromes (MDS) are clonal haematopoietic stem cell disorders characterised by ineffective haematopoiesis leading to blood cytopenias, with a high incidence of progression to acute myeloid leukaemia (AML). In those with lower-risk MDS, the risk of AML is reduced, and overall survival is longer. In these patients, the main objective of treatment is typically to manage cytopenias (predominantly anaemia), reduce the number of blood transfusions and improve quality of life. Previous cost-effectiveness studies have considered treatments for MDS or transfusion-related complications other than anaemia. Objectives of this study were to develop a de novo disease progression and cost-effectiveness model specifically for the treatment of anaemia in low-risk MDS, and to validate the underlying MDS model structure.

METHODS: Following a review of the published literature, a lifetime Markov state transition model was developed using Microsoft Excel, with health states reflecting MDS through multiple levels of transfusion dependency, AML and death, using a 28-day cycle length. Data to populate the developed model (rates, costs and utilities) was available from 8 out of the 19 studies identified by the literature review. Subsequently, the model was validated by matching inputs and comparing outcomes to these 8 published studies. Modelled outcomes were contrasted to reported costs, life years and quality-adjusted life years (QALYs).

RESULTS: In total, 73 endpoints were compared. After normalisation, the validation provided an overall R of 0.990 (non-normalised cost outcomes: 0.991; non-normalised QALY outcomes: 0.984; non-normalised life year outcomes: 0.997), a root mean square percentage error of 26.51% and a mean absolute percentage error of 16.60%.

CONCLUSIONS: This study is the first known publication to report the development of a cost-effectiveness model comparing treatments for anaemia in low-risk MDS. Importantly, outcomes associated with the underlying MDS structure exhibited a high degree of consistency with previously published models of low-risk and high-risk MDS cohorts.