Framework for Addressing Structural Uncertainty in Decision Models (Bojke et al.)
This method helps decision-makers account for uncertainty about how a model should be structured when evaluating health technologies. Instead of just testing different model versions informally, it provides a formal way to compare and combine different model structures using data or expert opinion.
At a glance
Use when
Evaluating health technologies with uncertain model structures, especially when standard sensitivity analyses are insufficient
Avoid when
Model structure is well-established and uncontested
Inputs
Decision analytic models with uncertain structural assumptions, either with or without relevant empirical data
Outputs
Weighted estimates of cost-effectiveness that account for both parameter and structural uncertainty
How it works
The framework formalizes structural uncertainty in decision models by expanding the model to include parameters that represent alternative structural choices. It distinguishes between parameter uncertainty (imprecise estimates from data) and structural uncertainty (lack of data or competing model forms). When data are available, model averaging is performed using weights based on predictive performance. When data are lacking, expert elicitation is used to inform structural parameters. The approach integrates probabilistic sensitivity analysis with model averaging to estimate expected costs and effects.
- HTA domains
- Clinical Effectiveness, Costs & Economic Evaluation, Organisational aspects
- Categories
- AppraisalHeterogeneityModel Validation
- Assumptions
- Structural choices can be parameterized; models can be compared based on predictive ability; expert judgments can be elicited as probability distributions
- Strengths
- Provides a systematic and formal approach to structural uncertainty; integrates with existing probabilistic methods; allows use of expert judgment when data are lacking
- Limitations
- Requires careful specification of alternative model structures; model averaging depends on quality of data or expert elicitation; may increase model complexity
- Also known as
- Structural Uncertainty Framework, Bojke Framework
Questions this answers
References & sources
Similar by meaning
- Model averaging for structural uncertainty in health economic decision models (Jackson et al.)
- Iterative Decision-Making Framework for Model-Based Decision Making
- Conceptualizing a Model (ISPOR-SMDM Modeling Good Research Practices Task Force-2)
- DARTH Decision-Analytic Modeling Coding Framework (A Need for Change!)
- Model Parameter Estimation and Uncertainty
Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.