This blog post is the third in a series on site performance KPIs based on an analysis CRIO performed for the Site Accreditation and Standards Institute (SASI), a nonprofit organization that serves as a quality accreditation organization for clinical research sites. Click here to learn more about how this analysis was performed.
From site selection through first patient in, sites have an extensive number of tasks to complete. Our system has metrics on some of the critical activities that can serve as bottlenecks, and reveals that sites with strong processes can greatly accelerate performance.
Turnaround times in the startup process vary widely
Regulatory turnaround: Sites often receive an IRB package (e.g., Form 1572) and must complete the package and return it to the sponsor. The median turnaround time for this is 7 days, but for top-quartile studies, we observe a turnaround time of only 4 days. The bottom-quartile metric is 15 days – over 2 weeks.
Budget turnaround: Sites receive a proposed budget from the sponsor or CRO and typically respond with a negotiated request for higher rates in certain areas. The median turnaround time for this is 9 days, but for top-quartile studies, we observe a turnaround time of 5 days. The bottom-quartile metric is 23 days – over 3 weeks.
Time from activation to enrollment varies as well
First patient enrollment: Sites receive formal activation or greenlight to enroll, and then proceed to screen patients for enrollment. The median time from activation to first patient screened is 20 days, but top-quartile performance is observed at just 8 days, or a little over a week. Bottom-quartile performance is 34 days, or just over a month.
Source template creation timelines are a hidden drag on startup, often invisible to sponsors
A process that is widely misunderstood and underestimated is the task of creating source templates. In CRIO, that’s done through the eSource template creation process. Using the creation of a study as the start date, and publication of the template as the completion date, we observe a median turnaround time of 11 days, with top-quartile performance at 5 days, and bottom-quartile at 26 days.
These figures are consistent with the data surfaced from a recent study performed by the Tufts Center for the Study of Drug Development on the protocol-driven source preparation process. That study showed that for high complexity studies, 80-93% (depending on TA) of sites report that source creation takes 21 or more hours to create. This observation on billable time for source creation certainly aligns with the 5-26 day elapsed time for source creation observed in our metrics.
These KPIs understate the true burden on sites. The Tufts study showed that sites rely on a variety of documents – over half described each of 5 documents as a “rate limiter” in the process – and that it therefore takes on average 6 weeks for the site to receive all the documents necessary for source completion. In some cases, the process can take up to 20 weeks.
Since sites generally begin the process of building source only after they have these documents in hand, the metrics observed in our system on elapsed time represent only the ending part of the entire process – i.e., when the site has the necessary materials at hand and sits down to devote the actual time to the source template construction.
This process is repeated across all the participating sites in a trial – a major source of redundant effort, enrollment delays and quality degradation.
CRIO’s Central eSource offering allows sponsors to streamline this process by enabling them to provide a centrally designed, but locally configurable, source template, built in tandem with the EDC guidelines and either harmonized with, or mapped for integration to, the final eCRF template. In our case studies (see this one for example), sites that participate in Central eSource are 30 or more days faster to first enrollment than sites that don’t. This faster time to enrollment may be surprising to sponsors at first – but it’s not surprising given the data observed from the Tufts study and our own system metrics.