Pictures copyrighted

T. Tervonen - Computational methods for evidence-based decision support in pharmaceutical decision making

Clinical trials are the main indicator for efficacy and safety of medical treatments. Although pivotal importance in evidence-based medicine, the trials are often underpowered due to insufficient sample size. Evidence synthesis methods allow to pool results of multiple trials and consequently help to reach sample sizes sufficient for drawing conclusions about the treatments' benefits and risks. Although these methods enable comprehensive analyses of a large number of trials, the more advanced ones suffer from an overly complex model specification. Through model generation algorithms also the advanced methods can be implemented in decision support systems and be made usable for the mathematically less educated clinical researchers and health policy decision makers.

In my presentation, I will use the case of regulatory drug benefit-risk analysis to introduce pharmaceutical decision support and the advantages of applying computationally intensive methods. The following topics will be covered:

  • Regulatory drug benefit-risk analysis
  • Evidence synthesis methods for clinical trials
    • Meta-analysis
    • Network meta-analysis
  • Multi-attribute value theory for pharmaceutical decisions
  • The ADDIS software (