Estimating the True Cost of Weight-Based Chemotherapy

OBJECTIVES:

Drug acquisition costs are a key driver of cost-effectiveness of chemotherapy treatments. Costs are proportionate to the dose required, often calculated based on weight or body-surface-area (BSA). Combined with dose banding and wastage, these factors introduce uncertainty around the true cost of drug acquisition. This study aimed to estimate mean treatment costs using a range of methods identified from published NICE appraisals, and to assess the sensitivity of estimated costs to the method used.

METHODS: Treatments were identified from NICE appraisals from 2015 onwards for the four most prevalent cancers in the UK. Mean treatment cost was estimated using five different methods assuming the same dose, vial strength and patient characteristics: (1) mean BSA/weight without wastage or dose banding, (2) mean BSA/weight including waste or (3) dose banding individually, (4) including both waste and dose banding, and (5) including waste and dose banding applied to a representative distribution of BSA/weight. Estimated costs were compared to those used in published HTAs and evaluated against UK average BSA/weight distributions and trial data to assess the sensitivity of the estimated cost to the method used.

RESULTS: Ignoring the effects of waste, dose banding and variability in patient characteristics systematically underestimated chemotherapy costs. Consideration of wastage (2) increased estimated costs by 19.3% (range: 1.7%-46.2%) versus method (1). When considering dose banding (3), estimated costs decreased marginally by 2% (-0.3%-0.7%) versus method (1), with the combined impact of dose banding and wastage (4) increasing estimated costs by 14.3% (-2.5%-27.7%). Modelling the impact of variability in population BSA/weight in addition to wastage and dose banding (5), estimated costs increased further by an average of 19.3 (3.7%-38.6%) versus method (1).

CONCLUSIONS: Estimating the cost of chemotherapies should account for dose banding, wastage, and variability in patient characteristics. Ignoring any of these factors may bias cost-effectiveness estimates.