Many trials tend to not make it to the planned Last Patient, First Treatment date set by trial managers. The client wanted a forecasting algorithm to predict a finishing date for a trial and how many patients they would end up with on the original planned Last Patient, First Treatment date.
A potential problem when running Clinical Trials is that trials might not make it to the planned Last Patient, First Treatment date set by the trial managers. An effective way of monitoring the expected progression of a running Trial would be by using a trial progression forecasting algorithm to accurately predict a Trial Recruitment finishing date as well as a projection of the number of patients they would end up with on the original planned Last Patient, First Treatment date.
Aside from these predictions the customer also needed the forecast model to be stratified by country to enable single country monitoring and country-to-country comparisons. Furthermore, the model was required to provide suggestions on reallocation of a patient from a struggling country to those countries who were ahead of schedule.
In order to accommodate the customer requirements several statistical models were evaluated along with different implementation routes using Qlik together with statistical and programmatical expansions like R and Python. When the best suited model was selected, it was adapted to the available data and subsequently implemented. Based on the statistical output generated by the model, several dynamic, interactive visualizations were made to illustrate the predictions and allow the end user to make on the fly changes.
After implementation of the Recruitment Forecasting app, the customer observed an increase in trial completion speed due to the ability to accurately reallocate patients where necessary. The Trial and Management team could also accurately predict and intervene in enrolment issues and address them in a timely manner. The solution also created the foundation for making data driven decisions based on relevant forecasting.
Selected Case Studies
How do you create a Central Clinical Monitoring Plan that meets the needs of all stakeholders?
Our solution enabled the client to fully automate their data extraction from source systems and reduce the manual interactions required.
We automated the process of pulling data reports from multiple systems, calculate various metrics, and create automated visualizations.