Is it clinical benefit? Cost? Value? The availability of treatment alternatives? To answer this question, a paper by Wranik et al. (2019) conducted a discrete choice experiment DCE to determine HTA stakeholders stated preferences. The sample consisted of HTA stakeholders from 5 countries: Australia, Canada, Germany, Poland, and the United Kingdom. The stakeholders included not only HTA committee members but also clinical and economic experts and, patients.
The attributes evaluated included survival benefit, adverse events, incremental health system cost per patient, number of patients eligible for treatment, and the availability of other treatment options. The authors first used DCE and binary logistic model to estimate how treatment attributes would affect the likelihood of public reimbursement. The authors also used an estimation technique that combines a DCE approach with profile best-worst scaling (BWS). In the BWS–as the name indicates–respondents are asked to determine which treatment attribute is best and which is worst. The authors justify their approach as follows:
The combining of DCE with BWS has 2 advantages: (1) typically the BWS task is considered to be cognitively less strenuous, and (2) a hybrid DCE/BWS model allows for the elicitation of more preference information from a given sample while limiting the respondents’ cognitive burden
The DCE/BWS approach was estimated using a Bayesian approach with non-informative priors. The Bayesian model allowed for the estimation of the impacts of attribute levels on the preferred choice, the best characteristic and the worst characteristic.
The authors found that:
HTA stakeholders consistently focus on a strong clinical benefit as the most relevant characteristic of a cancer drug. Some concerns have been voiced by the public that HTA agencies focus on cost-effectiveness alone,62 but our respondents suggested otherwise. Cost attributes ranked below a strong clinical benefit, and recommendations based on cost attributes were made with less conviction.
The results are interesting but highlight the drawback of this approach. Decisionmakers perhaps value clinical information most in a hypothetical situation but in the real-world, value (i.e., cost-effectiveness) often dominates. For instance, a paper by Dakin et al. 2015 found that:
Cost‐effectiveness alone correctly predicted 82% of decisions; few other variables were significant and alternative model specifications had similar performance.
Cost‐effectiveness alone correctly predicted 82% of decisions; few other variables were significant and alternative model specifications had similar performance.
Sources:
- Wranik WD, Jakubczyk M, Drachal K. Ranking the Criteria Used in the Appraisal of Drugs for Reimbursement: A Stated Preferences Elicitation With Health Technology Assessment Stakeholders Across Jurisdictional Contexts. Value in Health. 2019 Dec 16.
- Dakin H, Devlin N, Feng Y, Rice N, O’Neill P, Parkin D. The influence of cost‐effectiveness and other factors on nice decisions. Health economics. 2015 Oct;24(10):1256-71.