– Written by Moulik Shah, Founder & CEO, MaxisIT Inc.
An efficient process from data analysis and reporting to regulatory submission is critical. Here’s how to streamline A&R for faster time to submission.
The Last Mile: How to Open the Data Pipeline for Clinical Data Analysis and Reporting
In shipping, last-mile delivery is the final and arguably most important step of the journey. Drug development has a last mile of its own: the journey from data analysis and reporting (A&R) to regulatory submission.
Both scenarios demand an efficient process to ensure on-time delivery. However, just how shippers get bogged down by traffic and disorganised delivery routes, so do drug developers get bogged down by data and incompatible systems.
The 2020 Tufts CSDD–IBM Watson Health benchmarking study found the average cycle time to convert raw data to analysis-ready data was 15.5 days for larger companies and 21.4 days for smaller companies. (1) That’s reasonable. But what about the analysis itself? If manual workarounds and IT issues are holding you back, consider a new approach.
Accurate clinical data A&R and on-time regulatory submission require a centralised, cloud-based statistical computing environment. In this environment, biostatisticians, clinical programmers and other analysts can travel the last mile faster and with fewer complications.
Challenges Impacting on Clinical Data Management and A&R
The factors that hold back clinical trial data operations and management impact on A&R even more acutely. Because the end goal of clinical trials is to obtain regulatory approval, improvements in A&R cycle times impact on the timing of regulatory submission and, ultimately, revenue.
Last mile roadblocks include one or more of the following:
1. Increasing Data Volume and Diversity
Clinical trials today generate vast amounts of data from a wide array of sources, including electronic health records (EHRs), medical images, apps and traditional site-collected data. The diversity of these data sources, each following different standards, complicates integration, quality management and review processes. This complexity can delay data readiness for analysis, making manual processes too time-consuming and prone to error.
2. Planned and/or Unplanned Mid-Study Changes
Protocol amendments and study database updates can significantly delay trials. While modern electronic data capture (EDC) systems have mitigated some technical delays by enabling real-time adjustments, the impact on A&R can still be substantial. These changes often necessitate additional data reprocessing and analysis, further extending timelines.
3. Siloed Systems
Clinical trial sponsors and CROs use an average of five different applications to manage clinical trials. About a third use six or more, according to a survey of clinical operations leaders. It’s no surprise that most (70%) list integrating these applications as their number one challenge. (2)
All these applications – combined with spreadsheets, which are still commonly used – lead to data and process silos. These silos hinder clinical trial operations and data management. Using disconnected legacy applications and spreadsheets for complex clinical trials negatively impacts on visibility between stakeholders and slows the process from study launch to A&R.
4. Legacy A&R Platforms
Many A&R platforms in use today are based on outdated technology and unable to scale to meet the demands of modern data requirements. These legacy systems often lack the flexibility and transparency needed for efficient statistical operations, making them ill-suited for today’s data requirements. (3)
Statistical operations require transparency through traceability of the statistical analysis to meet more stringent regulatory requirements. (4) That type of visibility is either impossible or difficult to achieve on rigid legacy applications.
How to Improve A&R Through a Data Pipeline
To overcome these challenges, a modern approach to A&R is essential. Timely access to high-quality data is crucial for biostatistics teams to develop regulatory submissions and other reports efficiently. Real or near real-time data access enables the creation of tables, figures and listings (TFLs) without the bottlenecks associated with manual data processing.
Automation and standardisation are key components of an efficient A&R process. Automation reduces the burden of repetitive tasks, while standardisation ensures consistency and compliance throughout the data pipeline. Integrated statistical computing environments (SCEs) can streamline A&R processes by bringing data closer to computing resources, thereby improving cycle times.
A ‘modern statistical computing environment (SCE)’ is designed around the three Cs: Collaboration, Compliance, and Computing horsepower. By integrating these elements, an SCE harmonises the delivery of milestone data, analytics and reports, while enhancing capabilities for statistical programmers.
A Few Key Features
1. Data Quality
Ensuring high-quality data is fundamental to any A&R process. A robust data pipeline should facilitate real-time updates so validated data can be delivered to the appropriate statistical programs without unnecessary delays.
2. Integration Capabilities
An effective A&R pipeline should be capable of integrating data from multiple sources and accommodating various data formats, sizes and types. Seamless integration minimises the risk of data silos and ensures a continuous flow of information.
3. Flexibility in Programming
Flexibility in statistical programming is vital for adapting to different study requirements. A modern SCE should support multiple programming languages and offer tools that enable statistical programmers to work efficiently while maintaining compliance.
4. Automation
Automation is essential to eliminating downtime and accelerating the A&R process. Automated refreshes of datasets, analysis-ready data, and TFLs reduce the risk of delays and enhance the accuracy of submissions.
5. Agility in Project Management
Agile project management practices, supported by appropriate tools, can improve the traceability and quality of deliverables. A streamlined approach to managing changes and validation requests helps keep the A&R process on track.
6. Exploratory Data Reviews
Facilitating exploratory data reviews through interactive visualisation tools allows for quicker safety assessments and cohort analyses. This capability reduces dependency on programming for every review request, speeding up the overall process.
Conclusion
The last mile is arguably the most critical in the journey from drug discovery to regulatory submission. To reach the destination as quickly as possible, while maintaining data quality and compliance, A&R demands a flexible and modern SCE that facilitates compliance, collaboration and computing at scale.
References
(2) https://www.veeva.com/resources/clinical-operations-survey-report-2020/
(4) https://journals.sagepub.com/doi/abs/10.1177/009286151004400104