Cost‐effectiveness of pathogen inactivation for platelet transfusions in the Netherlands

Abstract

The objective of this study is to estimate cost-effectiveness of pathogen inactivation for platelet transfusions in the Netherlands. We used decision tree analysis to evaluate the cost-effectiveness of the addition of pathogen inactivation of pooled platelets to standard procedures for platelet transfusion safety (such as, donor recruitment and screening). Data on transfusions were derived from the University Medical Centre Groningen (the Netherlands) for 1997. Characteristics of platelet recipients (patient group, age, gender and survival) and data/assumptions on viral and bacterial risks were linked to direct and indirect costs/benefits of pathogen inactivation. Post-transfusion survival was simulated with a Markov model. Standard methods for cost-effectiveness were used. Cost-effectiveness was expressed in net costs per life-year gained (LYG) and estimated in baseline- and sensitivity analysis. Sensitivity was analysed with respect to various assumptions including sepsis risk, reduction of the discard rate and discounting. Stochastic analysis to derive 90% simulation intervals (SIs) was performed on sepsis risk. Net costs per LYG for pathogen inactivation were estimated 554,000 euro in the baseline-weighted average over the three patient groups (90% SI: 354,000-1092,500 euro). Sensitivity analysis revealed that cost-effectiveness was insensitive to viral risks and indirect costing, but highly sensitive to the assumed excess transfusions required and discounting of LYG. Given relatively high net costs per LYG that are internationally accepted for blood transfusion safety interventions, our estimated cost-effectiveness figures for pathogen inactivation may reflect acceptable cost-effectiveness in this specific area. Two main assumptions of our model were that the pathogen inactivation was 100% effective in preventing transmission of the pathogens considered and was not associated with major and/or costly adverse reactions. Validation of several crucial parameters is required, in particular the Dutch risk for acquiring and dying of transfusion-related sepsis.