Study of Operating characteristic of Bayesian methods that borrow treatment effects

Study identification

• PURI:
  https://redirect.ema.europa.eu/resource/106787
• EU PAS number:
  EUPAS105257
• Study ID:
  106787
• Official title and acronym:
  Study of Operating characteristic of Bayesian methods that borrow treatment effects
• DARWIN EU® study:
  No
• Study countries:
  France

  United States
• Study description:
  While the appropriateness of Bayesian methods in the paediatric context is largely recognised, most existing methods refer to borrowing control arm data only. From a regulatory perspective, an important issue with Bayesian methods is that they potentially do not control the Type I Error. Many simulation studies have been performed within the last decade in order to evaluate the operating characteristics, e.g. regarding Type I Error inflation and possible gains with respect to power, associated with Bayesian methods. Moreover, these methods are known to be potentially sensitive to small changes in underlying parameters, such as a weight parameter which determines how much information is borrowed from the historical data, and could potentially leads to different conclusions. Considering the limitations associated with these methodologies, an extension towards borrowing treatment effect, instead of historical control data only, is considered. The primary aim of this Tender is to thoroughly evaluate the available Bayesian methods for borrowing information on treatment effects in the context of paediatric evaluations. More specifically, specific objectives of this project include: - Review and evaluate available and published methods for dynamic borrowing via a structured and critical review (not limited to paediatric extrapolation) - Provide a clear mathematical specification of each model identified, articulating the interpretation of underlying parameters, and define the relevant operating characteristics (OCs) and performance metrics to evaluate for a given context - Evaluate and compare (via simulations) performance of such methods along the pre-defined OCs in various common scenarios - Synthesize and disseminate overall conclusions and recommendations
• Study status:
  Ongoing

Research institutions and networks

Institutions

Quinten Health

Daniel Lee New York, USA

Study timelines

Date when funding contract was signed

Planned: 25/01/2023

Actual: 25/01/2023

Study start date

Planned: 25/01/2023

Actual: 25/01/2023

Date of final study report

Planned: 25/01/2025

Sources of funding

Other

More details on funding

European Medicines Agency

Regulatory

• Was the study required by a regulatory body?:
  Yes
• Is the study required by a Risk Management Plan (RMP)?:
  Not applicable