From the first clinical description of polio in 1789 to the World Health Assembly’s resolution to eradicate it in 1988, progress in eliminating this debilitating disease has largely been fueled by successful clinical trials and mass vaccination campaigns. Yet, not all clinical trials are created equal. The informativeness of a trial — or its ability to guide clinical, policy, or research decisions — depends on several factors. But how can we know ahead of time if a trial will be informative?
Fast Data Science’s Clinical Trial Risk Tool incorporates machine learning to predict the informativeness of a trial from its protocol text, and here’s how it works.
What Makes a Trial Informative?
First, it’s essential to understand what makes clinical trials informative. Studies suggest that key factors include the number of participants, length of trial, number of uncontrolled variables, and the ability to control exposure, among others. Plus, the absence of a completed Statistical Analysis Plan (SAP) in a trial protocol is a red flag for higher risk.
Applying Machine Learning to Protocol Texts
Using the aforementioned parameters as an example, Fast Data Science’s Clinical Trial Risk Tool analyzes trial protocols and retroactively compares them with previous trials deemed informative or uninformed. Through this large-scale text analysis, the machine learning system forms algorithms to predict the potential informativeness of a trial based on its protocol.
And with data from clinical trials across various sectors, including previous Polio clinical trials, the tool is remarkably robust.
The Role of Clinical Trials in Polio Eradication
Due to their nature, clinical trials played an indispensable role in global Polio eradication efforts. The successes and setbacks inherent in Polio trials provided invaluable data that informed future studies and mass vaccination campaigns. WHO recognizes that the continued use of the Oral Polio Vaccine (OPV) carries a risk. Indeed, in places with low vaccination coverage, the weakened vaccine virus in OPV can circulate in undervaccinated communities.
Therefore, the informativeness of Polio clinical trials keeps gaining importance. To eliminate Polio, we must not only conduct clinical trials but ensure that they are of high quality and capable of convincingly guiding clinical and policy decisions.
By leveraging machine learning capabilities, Fast Data Science’s Clinical Trial Risk Tool assists significantly in realizing this goal. This forward-thinking approach enhances our predictive abilities and informs more accurate decision-making, ultimately aiding in the continued fight against Polio.
