Using big data to overcome cumbersome study-start up processes

The process of initiating clinical trials is cumbersome, challenging and fraught with delays – it has the greatest impact on patient recruitment, trial duration and its associated costs, unfortunately, it is also the worst performing stage of any clinical trial. Study startup is the most resource- and labor-intensive period in a study’s life cycle after data cleaning and database lock activities. Any increase of duration or responsibilities during startup has an increased logistical and financial impact. According to the CenterWatch Survey of Investigative Sites in the U.S., nearly 70% of all trials experience enrollment delays¹, while nearly half complete later than originally planned² and one-fifth of sites never recruit a single subject³.

Protocol complexity and increased partnering in study startup is here to stay. According to a recently released report from Research and Markets,⁴ by 2020 close to three-fourths of all clinical trials will be performed by CROs. As outsourcing continues to increase (as well as the amount of data to be shared), sponsors and CROs must work on ways to collaborate by building on the strengths of each organization and utilizing information and resources in a more uniform, consistent manner. Close collaboration with CROs and taking a risk-based approach, which relies on Big Data analytics, is key to enabling rapid study startup in an increasingly complex clinical development environment. 

Problem and status quo
Typically, the site nomination and selection process is collaborative and completed utilizing internal and external data supplied from either the CRO data source and/or the sponsor sources. This data is critical to the success of site selection and includes such information as:

• Site capabilities and skills

• Past performance of sites

• Background on the Principal Investigator and the institution

Too often the process is inefficient due to the lack of data availability related to operational cycle times, site submission timelines, and other factors. Without this data, the sponsor and CRO are at risk of selecting non-active or non-enrolling (NANE) sites, which ultimately drives up the costs and wastes valuable time in study startup. Currently, 80% of trials fail to meet enrollment timelines⁵ and up to 50% of research trial sites enroll one or no patients.⁶

Today, most operational cycle time tracking is still conducted via Excel sheets and information cobbled together from CTMS and EDC databases, which are then manipulated to provide rudimentary views of site performance. Often this cycle-time tracking is incorrect or lacks data governance and sufficient detail to be of value for decision-making. In the future, successful sponsor-CRO partnerships will require the combination of both their institutional memories with regard to site capabilities, patient availability and performance to reduce the number of NANE sites. Further complicating the decision making process for sponsors is the use of multiple CROs, with disparate systems for collecting operational data, and the inability to collect and then combine operational data into their own system.

Big data enables risk-based SSU

Tracking this low level activity completion times and cycle time data involves thousands of data points. Like other industries, clinical operations teams are now realizing the challenges and opportunities that lie in effectively managing this big data. 

For organizations looking to increase transparency in the study startup process, the sheer volume of data and the silos in which they exist can be a daunting hurdle. New generation systems enable teams to capture, analyze, share and visualize study startup data in one system. Trial project managers can now know exactly if a study is on track and if not, make the required decisions to remove bottlenecks and eliminate or refine unnecessary activities to ensure that all information required for regulatory submissions are ready by the submission dates. Having access to the data trends in a central location drives decision-making, allowing study teams to focus on potential risks and the most critical data and processes necessary to achieve study objectives–a risk-based, data-driven approach to study startup. There is growing consensus that risk-based approaches to monitoring are more likely than routine visits to all clinical sites and 100% source data verification to ensure subject protection and overall study quality. Like global monitoring, taking a risk-based approach to study startup is a best practice for accelerating activities from site selection through to activation. Rather than relying on individual knowledge or select outdated data, taking a risk-based approach requires that you gain insight into the key bottlenecks and processes that are most likely to affect study startup performance.

How to implement risk-based SSU

Establish Data Transparency. The first pivotal step involves increasing transparency in study startup to enable identification of key trends and processes. Having data in multiple locations with multiple partners is an obvious bottleneck. Efficient site selection requires that organizations combine both internal and external data sources. Data sources may include both in-house repositories such as CTMS, investigator databases, feasibility surveys, quality/risk Assessment information, as well as third party sources such as epidemiology data, site performance data, and subject availability. This data warehousing enables study teams to compute selection and performance variables, which ultimately drives improved decision-making regarding site selection.

Promote Collaboration Across the Study Team. It is not enough to have central access to a comprehensive data repository. Globally dispersed study teams need a way to collaborate in real-time and track milestones, role assignments, site selection, and study startup progress. During site selection and into site activation, protocols may change and key processes may be impacted by external factors or general unforeseen challenges. The team needs to be able to watch these changes/trends and change course accordingly to move the study forward. That means being able understand risks in advance to be able to pivot team members onto critical task where risks are identified. Which can only be accomplished with transparency of the trends and the allocation of team members across multiple trials and activities.

Identify and Optimize Key Processes. With the data warehouse in hand, it is essential to track study activation and real-time cycle time metrics as the study proceeds. This provides visibility into the success of your site recruitment strategy, allowing you to identify risks and put mitigation plans into place ahead of time.

Monitor Progress though Routine Data Visualizations. Data visualization is an on-going activity, which should have dedicated resources to enable proactive risk management. Data visualization with predictive trend monitoring allows teams to see risks better than numbers in a columnar view like Excel. Performing data visualizations with historical and just in time data can help teams mitigate risk factors to recruitment and retention by finding the optimum alignment of top performing sites with high patient availability. Teams can then quickly assess which sites have performed best in past studies on a variety of performance categories, such as startup, throughput, retention and quality.

Don’t Forget End-to-End Lifecycle Optimization. End-to-end lifecycle optimization should be considered as part of any solution for optimizing study startup. To activate sites on-time and meet enrollment targets, you must take a holistic view of the process, looking at potential bottlenecks and how they may impact downstream activities. Being an informed team allows for proactive mitigation of bottlenecks for sites, which have great potential but are in need of support to be fully efficient in the executing of the study.

Complexity in study startup is an everyday reality that is here to stay, but a risk-based study startup approach, which relies on a centralized, data-driven approach integrating insights and processes within the sponsor team and in collaboration with CRO partners, can position clinical trials for ultimate success.

1. CenterWatch Survey of Investigative Sites in the U.S. (2009).

2. Gwinn, B. (2011, June). Metrics for Faster Clinical Trials. PharmaVOICE.

3. Enrollment Performance: Weighing the “Facts.” (2012, May). Applied Clinical Trials.

4. Research and Markets Report “Research and Markets: The New 2015 Trends of Global Clinical Development Outsourcing Market”
http://www.businesswire.com/news/home/20150130005621/en/Research-Markets-2015-Trends-Global-Clinical-Development#.VXHHFFxVikp

5. Kremidas, Jim. Recruitment Roles: Sponsors, CROs, and investigator sites must all work together for effective patient recruitment. Available at:
http://www.appliedclinicaltrialsonline.com/recruitment-roles

6. Miseta E., Bring Down the Cost of Clinical Trials With Improved Site Selection. Clinical Leader. Dec. 19, 2013. Available at:
http://www.clinicalleader.com/doc/bring-down-the-cost-of-clinical-trials-with-improved-site-selection-0001?atc~c=771+s=773+r=001+l=a