As we reviewed in the first part of this series, data governance is critical to any organization building data products for monetization. In the clinical trials industry, it’s even more impactful due to the highly-regulated nature of dealing with patient data.
Simply put, if two similar drugs are submitted for approval, the one with the most accurate and complete data will be approved first. So if you consistently submit high-quality data to regulatory agencies, you are bound to shorten time to market because your data is complete, accurate, and of high quality, requiring no back and forth with regulators.
Because of the sheer amount of data needed for regulatory submission and longer-term post-approval commercial objectives, the need for data governance in clinical research is more critical than ever.
Pharma is drowning in data
The amount of data flowing into a clinical trial is continuing to increase, with the burgeoning influx of mHealth devices, wearables and other sources of continuous, real-time data. Real world evidence (RWE) data such as medical claims data and prescription usage data are also adding complexity to clinical trials, requiring data governance from a quality and trustworthiness perspective.
On top of this, during the course of a clinical trial, pharma relies on hundreds of vendors, providing a range of services from Interactive Response Technology (IRT), electronic Clinical Outcomes Assessment (eCOA), labs, imaging, cardiac safety, respiratory, and Electronic Data Capture (EDC) to analytics, reporting, data sciences and beyond. Most vendors are hired to work in a specific domain, lacking a long-term strategy to operate under the same data governance umbrella. There is also a lack of data standards and operability among these vendors, making data governance even more critical.
Then, when you factor in data variety including unstructured, binary and social data as part of a clinical trial, the integration challenges become compounded in the absence of effective data governance.
To remain competitive and achieve speed to market, sponsors and pharma companies need to take true control of their data and rapidly ingest, consume, integrate, analyze and act on large volumes of data from an extensive range of data sources across a variety of formats. Figure 1 depicts a holistic approach to data governance.
Best practices for effective data governance
In order to achieve effective data governance, one must first accept a methodical approach to decision-making using data. Pharmaceutical companies should consider:
- What the data means
- Where the data is coming from
- Where the data should be used
- How accurate the data is/needs to be
- What rules the data needs to follow
- Security of the data along the data pipeline
- Who can access what data and via which channel
- Data protection and privacy
- Data lineage from source to destination
- What the system of record is for each data element
- Building data stewards and data custodians – owners and keepers of master data
- Master data management and metadata management
- Establishing a data governance council, a cross-functional team consisting of business and technology team members
Data Governance is a particularly important component of mergers and acquisitions, business process management, legacy modernization, financial and regulatory compliance, AI and ML, business intelligence application, data warehouses, and data lakes.
Data governance is a journey and needs to be a part of what you do when you design data products, not an afterthought. Data is too important to the clinical trials industry to be left alone and ungoverned.
In clinical research, data governance is the glue that holds quality, consistency, and trustworthiness together.
Prakriteswar Santikary, PhD is the Vice President and Global Chief Data Officer at ERT.