Joining the data analytics bandwagon, the pharmaceutical giant Pfizer has launched their first clinical trial predictive modelling system which is aimed at reducing study risk during protocol design and to better study execution phases. In a recent interview Jonathan Rowe, the Executive Director and Head of Clinical Development Quality Performance and Risk Management of Pfizer shed some light on these predictive modelling systems.
When asked in the interview about the purpose of their predictive model and what it is meant to achieve, Rowe responded as follows…
It is true that there are quite a few models in the realm of GCP quality performance which we have developed and continue to refine. A relatively straightforward one is the correlation model where we correlate our clinical trial process performance to select the results of the GCP as is defined in the ICH E6.
We mainly focus on the following factors when correlating the outcomes – are the rights of the patients being preserved, their consent, well-being and safety are being safeguarded or not etc.
The way these clinical trial systems work is that the GCP outcomes are compared against the clinical trial quality metrics that have been collected and correlation is built to observe if any of those metrics can predict whether we are going to have any issues while trying to achieve the GCP outcomes. We further fine-tune these systems by adding specific time scale into the models so, that the team studying it can predict a few months ahead of time whether a GCP issue will come up. Thus, these acts like an early warning system. We have a large developmental space for clinical trial assessments and compare their relationship statistically within as well as across a varied cohort of studies.
We then asked him the most obvious question that the whole data analytics community is all ears to know about the future of data analysis and predictive modelling in healthcare. And that is, how do those predictive models influence the decision-making process of the study team during protocol design. To this Jonathan’s response was as follows…
Yes, a very anticipated question; as we have a question bank we can correlate and compare how the questions are answered to meet our levels of quality standards and the study team can review their protocols and take necessary actions to mitigate and reduce risk. They do so by either changing the protocols or with effective mitigation planning. For instance, there may be some protocols which may possess some inherent quality risks for factors such as complex dosage and such dosing regimen must persist. With such an efficient study risk predictive modelling tool the question bank will let the team to be thoughtfully and proactively capable of mitigating risk and to be more cautious in high risk areas to reduce risks, mistakes and deviations.
These endeavours are the fruit of our constant drive to reduce quality compromising events. A short term goal for the study risk predictive model is to utilize it well for monitoring and support. With better understanding of the risk factors we will be able to better resource plan when working on protocol designs.
However, as a summarizing statement Jonathan Rowe was pretty clear that this initiative is not completely about protocol optimization although it apparently seems so. The model is more focussed at mitigating GCP quality risk. This is a standard process to identify which trials are risky and allows the study team to be prepared to answer against such risks by putting in place measures to ensure quality of the study. Thus, the applications of Big Data analytics and predictive modelling seem to be bringing on more advanced holistic approaches on the table for healthcare pharmaceuticals and drug-testing.
Which industry will Big Data delve into next to bring about a similarly big transformation by putting in place order in place of chaos?
For more such news, discussions and articles on the adventures of Big Data analytics and how it is changing the world around us, stick around Dexlab Analytics for regular updates on the same.
Interested in a career in Data Analyst?
To learn more about Data Analyst with SAS Course – click here.
To learn more about Data Analyst with R Course – click here.
To learn more about Big Data Course – click here.