Evidence-based site selection and study planning
The Pulse on Technology by Elisa Cascade
Any sponsor, CRO or third-party involved in site selection and study planning will tell you that data driven tools are the key to success. Thankfully, sponsors have a range of tools to choose from in today’s market that offer tables, charts, graphs, heatmaps, projections, scoring and in-depth reporting. The most important difference between these tools, however, is not the user interface, but rather the underlying data sources. So for sponsors, the question remains, “Are the tools responsible for successful study planning and site selection? Or should we worry about the data that feeds them instead?”
While clearly both the tool and the data offer value, most people in clinical operations would say “show me the data!” In fact, if you ask how much data is needed to support evidence-based study planning and site selection, most respond with, “All of it . . . or at least as much as my company can afford!”
However, too much data itself can create problems. We often hear from clinical operations colleagues who are overwhelmed with trying to integrate information from multiple data sources. Thus, companies usually end up either expending significant labor in manually matching data across sources or they compile sources individually and sequentially, resulting in sub-optimal study planning and site selection decisions.
From a technology perspective, solving the data integration issue requires a common data standard and a universal identifier for both people and facilities (unique identifiers for studies currently exist both at pharmaceutical companies and through ClinicalTrials.gov). However, not even standardized data, universal identifiers and technology will completely solve the issue. That’s because combining data sources for study planning and site identification also requires an alignment of various permissions for investigators to share their information. Unfortunately for the industry, obtaining permissions for each investigator has always been a tricky issue.
But this doesn’t mean that the industry has stood still. Several pharmaceutical companies have started down the path of obtaining permissions for data integration through the Investigator Databank and TransCelerate collaborations. Both initiatives have accepted a common data standard for CTMS and profiles data, and have also begun moving forward in capturing investigator permission to enable cross-company data sharing.
Although public data can be helpful in understanding a site’s capabilities for feasibility (e.g., the ever popular -70 degree freezer question), it is ultimately secondary in importance to actual site-level performance data from CTMS. Publicly available data sources such as ClinicalTrials.gov can also provide insight into experience at the site level; however, performance is estimated based on study-level data as opposed to actual site-level performance. As demonstrated by a recent publication1, use of estimated data at the study level often results in inaccurate planning and a need for rescue sites due to differences between average metrics from study-level data, and actual site-level median performance.
Trying to move beyond CTMS and profiles sources and into integration of EMR/EHR data has also been challenging due to the inability to obtain permissions. To date, we have been unable to identify a system that has the permissions to transfer the data for matching to a universal identifier, which is ultimately needed for integration with other study planning and site identification data sources. Our discussions with providers suggest that this is because of the legacy contracts that exist between the providers and users of EMR/EHR systems. Thus, clinical operations experts have needed to revert back to a manual, sequential approach for combining EMR/EHR with other study planning and site identification data sources.
Outside of the inefficiencies from the sponsor perspective, it is also important to evaluate the impact on sites. Sites who are forced to work with unrealistic sponsor expectations based on sub-optimal data are frustrated, which can lead to subject recruitment delays and potential investigator turnover.
So what can sites do? First, if invited to consent to have your data shared, consider saying “yes,” especially if the collaboration includes sponsors that you have either worked with previously or would like to work with in the future. Providing permission to have information shared will make your site visible to more study sponsors. Second, we recommend that sites know how to leverage their EMR/EHR to support identification of patients for clinical trials. This too makes sites more attractive to study sponsors while at the same time helping to facilitate more rapid subject recruitment once the study has gone live.
- Sears CE and Cascade E (2016). Using Public and Private Data for Clinical Operations. Applied Clinical Trials Volume 25, Issue 12. http://www.appliedclinicaltrialsonline.com/using-public-and-private-data-clinops