Aaron Cohen serves as the Vice President of Data Strategy at Definitive Healthcare. Life sciences organizations consistently face difficulties in choosing sites and enrolling participants for their studies. According to an analysis by the Tufts Center for the Study of Drug Development, which examined nearly 10,120,000 investigative sites in 151 Phase II and III worldwide trials, it was discovered that around half of the chosen sites either failed to enroll a patient or had fewer patients than anticipated.
1 Decisions regarding site selection often rely on historical performance indicators such as previous trial involvement and past enrollment, usually giving preference to sites with a proven track record over new site possibilities. This generates a sampling issue: if the collection of possible sites mainly consists of those chosen in previous tests, examinations of that group will naturally lean towards characteristics linked to previous selection.
Sponsors tend to choose sites based on past selection choices, which can lead to the repeated selection of established locations, possibly affecting trial speed, diversity, and overall quality. AI can assist organizations in finding new clinical trial sites by considering a wider range of attributes to guide site selection, ensuring that suitable candidates are not overlooked.
By employing predictive algorithms to identify appropriate patient populations, investigators, and clinical capabilities, among other factors, sponsors can potentially find sites that are a perfect match, broadening their site selection strategy. This approach considers trial history when selecting sites.
Sponsors typically assess websites based on four main areas: potential eligible patient population, research expertise, available investigators, and past trial performance.
