In this blog we share a case study of work our strategic consulting team conducted supporting accurate event prediction for a cardiovascular outcomes research trial using the Major Adverse Cardiac Events ( MACE) composite endpoint.
Clinical Development Background
Our client was planning a Phase 3 randomized cardiovascular outcomes research trial. The trial used time to the first occurrence of the composite endpoint Major Adverse Cardiac Events (MACE) as the primary endpoint, and secondary endpoints included time to event analyses of components of this primary endpoint. The trial employed a group sequential design, and incorporated interim analyses.
In order that they could effectively plan the interim analysis and related activities, our client needed to develop accurate time-to-event predictions. The client also wanted to investigate a proposed amendment to the trial design that would potentially increase the power, but extend the trial duration. Obtaining an accurate prediction of the trial duration was critical to help evaluate this trade-off and make the decision about the potential amendment.
Cytel has extensive experience developing both custom and commercial software to support event prediction.
Cytel’s Strategic Consulting team customized a proprietary prediction model, and also used R programming to meet the specific requirements of the project.
Cytel prepared the raw datasets provided by the client, and created a time-to-event prediction report based on these.
As the project progressed, predictions for two more endpoints needed to be incorporated, and so the prediction model was extended.
A key challenge of the project was to reconcile the predictions over time based on accumulating data and an evolving and complex set of prediction models. To accommodate this, Cytel consultants developed a sophisticated validation procedure, which was embedded within the prediction model to track any shifts over time.
Accurate predictions enabled the client to effectively plan the interim analyses and Data Monitoring Committee meetings.
The project has led to the development of groundbreaking new prediction tools and validation techniques.
A strong relationship with the client team enabled Cytel to interpret the requirements and respond quickly with custom solutions.
Cytel’s unique blend of statistical and software expertise underpinned the solution.
Are you looking for statistical consulting support for your next trial? Click the button below to explore the various ways we help our clients.