Loading

The ESPERANZA Project: new statistical methods to improve clinical trials

Retour

The ESPERANZA project, funded by the ANR, aims to develop new statistical methods and practical tools to better evaluate public health interventions. It brings together several national and international teams, including the BPH’s BIOSTAT team, with Virginie Rondeau acting as co-leader of two work packages.

 

 

 

 

 

 

 

Cluster-randomised trials (CRTs) are a type of clinical trial which, rather than examining the effects of an exposure on individuals, are based on groups of people called ‘clusters’.

 

 

These trials are increasingly being used to evaluate complex interventions, particularly in public health. They produce reliable, high-quality data to inform health policy and support the development of practices on a large scale.

 

 

 

 

Can we improve statistical methods by incorporating AI?

The methods for planning and analysing these trials are based on three types of judgement criteria for evaluating the results:

  • binary, relating to what is observed or not, for example ‘sick or not’
  • continuous, involving means, standard deviation, etc.
  • ‘time to event’ (TTE)

Whilst the first two are well established, methodological developments concerning the third remain more limited.

 

 

The ESPERANZA project, which stands for ‘Enhanced methodS and Practical tools for clustER rANdomiZed triAls’, aims to improve the statistical methods applicable to cluster-randomised trials where the outcome measures relate to the time to event and take competing risks into account.

In the long term, the integration of artificial intelligence into the calibration of the parameters required to calculate the number of subjects to be included could help to overcome certain limitations of pilot studies – which are often small in scale – and thus improve the planning of complex clinical trials.

 

 

The analyses will be based on several major existing databases, including cohorts and registries that collect data on the time to event among patients grouped within centres and/or followed up by doctors.

 

 

“The project will primarily have an impact on biostatistics and clinical trial methodology, through the development of new statistical methods and practical tools for designing and analysing cluster-randomised trials (CRTs) focusing on ‘time-to-event’ endpoints.”

— Virginie Rondeau

 

 

 

National and international collaboration

The ESPERANZA project, supported by the ANR, will begin in September 2027 and run for four and a half years, with a budget of 592,742 euros.

 

 

Led by Agnès Caille of the University of Tours, the project brings together two international teams from the University of Ottawa and University College London, as well as the BIOSTAT team at the BPH.
Among the BPH members involved in the project, Virginie Rondeau, a research director within the BIOSTAT team, is co-leading two work packages.
Since the start of her career, Virginie has applied her expertise in biostatistical development to a variety of public health topics, such as cancer, Alzheimer’s disease and the effects of air pollution.

 

 

The ESPERANZA consortium draws on the complementary expertise of biostatisticians and scientists specialising in computer science and epidemiology, as well as on the expertise of international collaborators in trial design and the modelling of correlated data.

 

 

This synergy is essential for the successful completion of a project that is at once methodological, computational and translational, the results of which may help to improve the design and analysis of clinical trials in public health.