Data management is key to keeping trials on track and bringing drugs to market in a timely way.
Drug development is a grueling process with long timelines and excruciating costs. Throughout the trial process, sponsors often face obstacles that delay timelines and inevitably increase costs. Even more frustrating, the odds of success are stacked against them— approximately 90 percent of drugs that reach clinical development never make it to regulatory approval and marketization.
Clinical drug trials can fall off track for many scientific reasons, though sponsors often associate trial delays with slow patient recruitment (more than 80 percent of trials fail to enroll on time). However, many studies go haywire for reasons that can be controlled with better planning and due diligence when it comes to choosing the right vendors. When problems arise, sponsors often look to different vendors in an effort to “rescue the study.”
The common challenges they often face in data collection and reporting that lead to costly delays include:
- Poor trial design and inconsistency with endpoints
- Inconsistencies in trial data and failure to detect erroneous or fraudulent data
- Failure to understand regulatory requirements and regulatory feedback
- Dispersed data that can lead to inaccurate analyses or poor data quality
- Vendor problems with staff and project team turnover
While a “rescue study” might be necessary to save a clinical program, these studies pile on thousands or even millions of dollars of additional costs. According to a market study by IMS Health, a single organization running 100 clinical trials a year spends $26 million annually to “overcome avoidable protocol design flaws and patient recruitment difficulties.”
Even scaling that figure for smaller pharma and biotech companies results in inexcusable costs that could be avoided. Sponsors considering a rescue study have to account for tasks such as new database build, data migration and import, new CRF/eCRF design, new SOPs and of course a new project team.
How to Plan Effectively?
1. Invest in data quality from the beginning
Many sponsors fail to define a global clinical data strategy from the outset, which includes plans for the entire phase of development including post-market. Clinical data is the greatest asset, and therefore the company should invest in a clinical data team with the proper experience and know-how for efficient database build and management as well as query management.
2. Involve a biostatistician from the beginning and keep a consultant statistician on your team throughout the development cycle
The best time to involve a biostatistician in a trial is from the very beginning in order to understand the study design and make suggestions on hypothesis testing and analysis. The statistician plays a vital role in protocol development and design, data management, monitoring and reporting. Keeping a consultant statistician on your team throughout the study will help alleviate problems that arise with trial design and data analysis as well as reporting to regulatory feedback.
Statisticians can apply trial design methods—such as adaptive trial design—which can make significant changes in the study that reduce timelines including early study termination if necessary. Most importantly, statisticians can support in DSMB (data and safety monitoring board) and regulatory meetings and help make sense of clinical data.
3. Centralizing clinical data with a specialized vendor
When data is dispersed across vendors, organizations often face problems with data traceability, cross-product analysis, query management and data inaccuracies. Centralizing clinical data services—including biostatistics, data management and medical writing—saves time by creating a global library of databases, shorter learning curves between CRO and sponsor and easy access to study metrics. Quality is greatly improved through standardization, familiarity with customer processes, formats, templates, and communication.
4. Strategic Functional Service Model to address staff augmentation and turnover
The FSP model facilitates a scalable, expert team of resources for a particular function and results in improved quality, eradication of change orders, reduced training and greater efficiency. Using FSP, there can be savings on recruitment fees, training costs, and HR management time.
The CRO is responsible for producing the required resources and ensuring continuity of trained resources. In this case, the sponsor can be guaranteed that it will have the same resources dedicated to a project that understand both the CROs requirements and the sponsor’s requirements. Smaller companies can consider the “Micro FSP” model that is a scaled down version of FSP for companies requiring generally less than 10 resources.
5. Risk Management Metrics from data experts to enable Risk-Based Monitoring
Risk-Based Monitoring combines on-site monitoring along with centralized remote monitoring by coordinating centers. Based on risk assessments about how the clinical information is captured and protocol has been designed, risk-based monitoring activities can be proactively supported by the use of reporting tools.
One important component of RBM is the metrics that enable source data verification and triggering alerts when sites have inconsistent data patterns of problems. Centralizing your data with a specialized team of data managers and statisticians allows for accurate and timely metrics about site performance.
Following the steps outlined above can help to create more streamlined clinical trial workflows. Data management is key to keeping the trial on track for successful outcomes in order to bring drugs to market in a timely way. Pharmaceutical companies have a lot invested and keeping a keen eye on data for the ever-important clinical trial process, will help to save valuable time.
About the Author:
Paul Fardy, Vice President of Data Services for CROS NT, has more than 27 years of experience in biometrics team management. Most of Paul’s career was spent in large pharmaceutical companies managing functional teams in data management, statistics, statistical programming and medical writing.