The Application and Implications of Novel Deterministic Sensitivity Analysis Methods

Area of Science:

• Health Economics
• Decision Analysis
• Biostatistics

Background:

• Deterministic sensitivity analysis (DSA) is crucial for interpreting parameter uncertainty in cost-effectiveness analyses (CEAs).
• Classic DSA methods and tornado diagrams can provide misleading information regarding parameter effects, correlations, and non-linearities.
• Advances in DSA offer more reliable methods for assessing uncertainty's impact on CEA results.

Purpose of the Study:

• To discuss advancements in DSA methods beyond classic approaches.
• To evaluate the implications of these advanced DSA methods for decision-making in health economics.
• To compare stepwise, distributional, and probabilistic DSA techniques.

Main Methods:

• Stepwise DSA: Enhances interpretability of marginal effects and non-linearities.
• Distributional DSA: Utilizes parameter probability distributions to assess outcome likelihood.
• Probabilistic DSA: Incorporates parameter correlations for a comprehensive uncertainty assessment.

Main Results:

• Stepwise DSA improves understanding of parameter effects, especially with arbitrary ranges.
• Distributional DSA offers insights into the likelihood of model outcomes.
• Probabilistic DSA, particularly when expressed in incremental net benefits, is most effective for understanding individual parameter uncertainty's impact on cost-effectiveness.

Conclusions:

• Probabilistic DSA is the most appropriate advanced method for assessing individual parameter uncertainty in CEAs.
• Distributional DSA can be sufficient depending on model complexity and decision-making needs.
• Decision-makers should evaluate and potentially adopt advanced DSA methodologies, adjusting guidelines accordingly.