In this era of “big data,” when we all have access to more data than ever before, it’s tempting to take a ‘DIY’ approach to clinical research. However, as search engines become increasingly sophisticated, many data sources commonly used to support clinical trial design lack such dynamism. For example, whereas many clinical trial planners rely on sources like clinicaltrials.gov, which have some value in terms of providing basic information about individual trials, these sources cannot process the rich types of data in an integrated manner that can inform a sophisticated trial design. Once companies realize these limitations, they often react by licensing more data from other sources, but none of these sources can provide a comprehensive picture of the trial landscape.
The increasing dynamism of big data require an AI-driven solution that can pull data from thousands of sources at once, one that can essentially sort through information-filled grains of sand. That is the strength of Phesi’s analytics platform, which applies predictive algorithms that funnel huge data sets from over 70,000 sources in real time – with over 6,000 records being added each day to the source data – to extract much-needed insights to facilitate and massively improve protocol design and site selection. Unlike conventional data analytic systems, which typically take many months to process the mountains of information required for insightful analysis, the Phesi approach takes just a few days, and is much more precise and accurate, typically increasing efficiency by up to 45%.
Phesi’s predictive analytics platform can simultaneously process several of the major steps in clinical trial planning, including:
- Accessing high-quality data in multiple forms and file types
- Employing algorithms to interpret dynamically changing data to drive protocol design, flag alignment with, or deviations from, inclusion/exclusion criteria for example
- Accurately forecasting enrollment cycle time and the required number of sites to optimize enrollment
The power of Phesi’s platform is illustrated in a recent analysis of 25 clinical trials (mostly Phase 2) in patients with relapsing-remitting multiple sclerosis (RRMS). In the following chart, each bubble represents a single trial; each trial randomized 200 to 374 patients. The size of each bubble represents the enrollment cycle time (ECT).
The analysis shows how RRMS trials with specific inclusion/exclusion criteria (e.g., disability scores, number of relapses) that deviate from modal values negatively impacts the number of patients per site per month, a phenomenon that increases ECT and therefore study duration and cost.
The hypothesis is borne out in reality when we look at one of the trials in detail below, that deviated from design modal values for the number of relapses, age (modal value is 18-55 years) and EDSS scores, that we predicted would impact trial execution and ECT. The trial sponsor amended its protocol several times which in turn led to prolonged ECT, increasing frustration and the burden for site staff and the client team, and, of course, significantly increasing financial costs if we assume each protocol amendment costs approximately $300,000.
The sponsor of this trial would have benefited from global predictive analytics, which would have accurately forecasted the optimal design and predict the optimal number of sites. Just as too many sites can unnecessarily prolong ECT and increase costs, so can too few sites. Arriving at the right number using the Phesi analytics platform requires complex and multi-dimensional analyses of large volumes of dynamic data, which is impossible to do well when performed manually or as segmented activities or with only a partial data set to hand.
At Phesi, all integrated analytics take place concurrently and in just a few days, avoiding the current trial and error approach taken today. We are guided by the principle that it’s better to execute the right plan than to have to rescue a trial. After all, why waste time and resources and risk trial failure, when patients are eagerly waiting for new and better therapies? Smarter trials equal faster cures.