Assessing the Value of New Antimicrobials: Evaluations of Cefiderocol and Ceftazidime-Avibactam to Inform Delinked Payments by the NHS in England.

Abstract

Objectives: The UK has recently established subscription-payment agreements for two antimicrobials: cefiderocol and ceftazidime-avibactam. This article summarises the novel value assessments that informed this process and lessons learned for future pricing and funding decisions.

Methods: The evaluations used decision modelling to predict population incremental net health effects (INHEs), informed by systematic reviews, evidence syntheses, national surveillance data and structured expert elicitation.

Results: Significant challenges faced during the development of the evaluations led to profound uncertainty in the estimates of INHEs. The value assessment required definition of the population expected to receive the new antimicrobials; estimating value within this heterogenous population; assessing comparative efficacy using antimicrobial susceptibility data due to the absence of relevant clinical data; and predicting population-level benefits despite poor data on current numbers of drug-resistant infections and uncertainties around emerging resistance. Though both antimicrobials offer the potential to treat multi-drug resistant infections, the benefits estimated were modest due to the rarity of true pan-resistance, low life expectancy of the patient population and difficulty of identifying and quantifying additional sources of value.

Conclusions: Assessing the population INHEs of new antimicrobials was complex and resource intensive. Future evaluations should continue to assemble evidence relating to areas of expected usage, patient numbers over time and comparative effectiveness and safety. Projections of patient numbers could be greatly enhanced by the development of national level linked clinical, prescribing and laboratory data. A practical approach to synthesising these data would be to combine expert assessments of key parameters with a simple generic decision model.