Periodic benefit-risk assessment using Bayesian stochastic multi-criteria acceptability analysis

Guidance on risk-benefit assessment of foods

The EFSA Scientific Committee has updated its 2010 Guidance on risk-benefit assessment (RBA) of foods. The update addresses methodological developments and regulatory needs. While it retains the stepwise RBA approach, it provides additional methods for complex assessments, such as multiple chemical hazards and all relevant health effects impacting different population subgroups. The updated guidance includes approaches for systematic identification, prioritisation and selection of hazardous and beneficial food components. It also offers updates relevant to characterising adverse and beneficial effects, such as measures of effect size and dose-response modelling. The guidance expands options for characterising risks and benefits, incorporating variability, uncertainty, severity categorisation and ranking of different (beneficial or adverse) effects. The impact of different types of health effects is assessed qualitatively or quantitatively, depending on the problem formulation, scope of the RBA question and data availability. The integration of risks and benefits often involves value-based judgements and should ideally be performed with the risk-benefit manager. Metrics such as Disability-Adjusted Life Years (DALYs) and Quality-Adjusted Life Years (QALYs) can be used. Additional approaches are presented, such as probability of all relevant effects and/or effects of given severities and their integration using severity weight functions. The update includes practical guidance on reporting results, interpreting outcomes and communicating the outcome of an RBA, considering consumer perspectives and responses to advice.

Methodological guidelines and publications of benefit-risk assessment for health technology assessment: a scoping review

OBJECTIVES

To map the available methodological guidelines and documents for conducting and reporting benefit-risk assessment (BRA) during health technologies' life cycle; and to identify methodological guidelines for BRA that could serve as the basis for the development of a BRA guideline for the context of health technology assessment (HTA) in Brazil.

DESIGN

Scoping review.

METHODS

Searches were conducted in three main sources up to March 2023: (1) electronic databases; (2) grey literature (48 HTA and regulatory organisations) and (3) manual search and contacting experts. We included methodological guidelines or publications presenting methods for conducting or reporting BRA of any type of health technologies in any context of the technology's life cycle. Selection process and data charting were conducted by independent reviewers. We provided a structured narrative synthesis of the findings.

RESULTS

From the 83 eligible documents, six were produced in the HTA context, 30 in the regulatory and 35 involved guidance for BRA throughout the technology's life cycle. We identified 129 methodological approaches for BRA in the documents. The most commonly referred to descriptive frameworks were the Problem, Objectives, Alternatives, Consequences, Trade-offs, Uncertainty, Risk and Linked decisions and the Benefit-Risk Action Team. Multicriteria decision analysis was the most commonly cited quantitative framework. We also identified the most cited metric indices, estimation and utility survey techniques that could be used for BRA.

CONCLUSIONS

Methods for BRA in HTA are less established. The findings of this review, however, will support and inform the elaboration of the Brazilian methodological guideline on BRA for HTA.

TRIAL REGISTRATION NUMBER

https://doi.org/10.17605/OSF.IO/69T3V.

From Individual to Population Preferences: Comparison of Discrete Choice and Dirichlet Models for Treatment Benefit-Risk Tradeoffs

Introduction. The Dirichlet distribution has been proposed for representing preference heterogeneity, but there is limited evidence on its suitability for modeling population preferences on treatment benefits and risks. Methods. We conducted a simulation study to compare how the Dirichlet and standard discrete choice models (multinomial logit [MNL] and mixed logit [MXL]) differ in their convergence to stable estimates of population benefit-risk preferences. The source data consisted of individual-level tradeoffs from an existing 3-attribute patient preference study (N = 560). The Dirichlet population model was fit directly to the attribute weights in the source data. The MNL and MXL population models were fit to the outcomes of a simulated discrete choice experiment in the same sample of 560 patients. Convergence to the parameter values of the Dirichlet and MNL population models was assessed with sample sizes ranging from 20 to 500 (100 simulations per sample size). Model variability was also assessed with coefficient P values. Results. Population preference estimates of all models were very close to the sample mean, and the MNL and MXL models had good fit (McFadden's adjusted R² = 0.12 and 0.13). The Dirichlet model converged reliably to within 0.05 distance of the population preference estimates with a sample size of 100, where the MNL model required a sample size of 240 for this. The MNL model produced consistently significant coefficient estimates with sample sizes of 100 and higher. Conclusion. The Dirichlet model is likely to have smaller sample size requirements than standard discrete choice models in modeling population preferences for treatment benefit-risk tradeoffs and is a useful addition to health preference analyst's toolbox.

A Bayesian approach for individual-level drug benefit-risk assessment

In existing benefit-risk assessment (BRA) methods, benefit and risk criteria are usually identified and defined separately based on aggregated clinical data and therefore ignore the individual-level differences as well as the association among the criteria. We proposed a Bayesian multicriteria decision-making method for BRA of drugs using individual-level data. We used a multidimensional latent trait model to account for the heterogeneity of treatment effects with latent variables introducing the dependencies among outcomes. We then applied the stochastic multicriteria acceptability analysis approach for BRA incorporating imprecise and heterogeneous patient preference information. We adopted an efficient Markov chain Monte Carlo algorithm when implementing the proposed method. We applied our method to a case study to illustrate how individual-level benefit-risk profiles could inform decision-making.