The clinical research field has benefitted from an explosion of available data, but questions still remain about how to best collect, disseminate, and utilize that data. To learn more, Outsourcing-Pharma spoke with Srini Anandakumar, head of product with Saama Technologies.
OSP: Could you please talk about the evolution of data in clinical research?
SA: Data in clinical research has been increasing at a tremendous pace. In 2005, the average clinical study consisted of fewer than 500,000 data points, and today many clinical trials have upwards of 10m. We’ve seen outliers with more than 100m data points, and I believe one day that will be quite commonplace.
This data explosion is due in large part to a greater prevalence of decentralized clinical trials and an increasing number of data sources, such as IxRS, ePRO, eCOA, eCRF, specialty labs, wearables, sensors, medical imaging, omics, and biomarkers. More than 70 percent of the clinical data collected today comes from outside the standard EDC system.
OSP: What are some of the areas in clinical data collection/processing/analysis you feel there might be room for improvement?
SA: Sponsors have to do a better job managing the increasing volume and complexity of data. Having information is great, but if you can’t gain insights from it in a timely manner then it’s really not worth having. Today’s clinical data managers are often unable to influence cycle times that are dependent on external data sources and vendors, and programming teams need longer lead times to standardize content.
Data silos create the need for multiple manual processes that drive up costs unnecessarily and make it more difficult to get drugs into patients’ hands, which is a big problem for our industry. It’s interesting to note that, despite a small increase in 2020, the return on drug research and development has declined by more than 80% over the past decade.
OSP: How might automation of various data-related processes be helpful in clinical research? How would that impact costs, speed, effectiveness, and other aspects?
SA: Using automation, analytics, and artificial intelligence can have a significant impact on clinical research. Sponsors that implement these technologies can expect to achieve 10X efficiency gains, which means they can do a lot more with fewer resources.
Data managers can set up new studies faster, better manage ongoing data review, and achieve data lock much faster; Saama has helped clients get to data lock in a single day when it used to take 30 days or more. A great example of this is the work Saama did to support a 47,000+ person clinical trial for a major pharmaceutical company’s COVID-19 vaccine. Saama’s Smart Data Query (SDQ), a domain-centric, deep learning/AI system, helped to ensure analysis-ready data in 22 hours versus the industry standard of one month or more.
On the programming side, technology can make it much easier to manage global transformation metadata, generate SDTM datasets at the study level, and re-execute datasets due to protocol amendments.
OSP: Please talk about why the adoption of automation in clinical data processing might be going more slowly than is ideal.
SA: While life sciences organizations are generally risk-takers when it comes to pushing the boundaries of science, they are risk-averse to technologies that enable clinical pipelines and submissions, largely due to regulatory requirements.
Without an automated way of leveraging data assets, sponsors have gotten into the bad habit of throwing people at problems and taking a project-based approach to their clinical studies.
The other major issue is that automating clinical studies is far more difficult than automating something like navigation or product recommendations. Because clinical trials are experiments with many variables, a single change can challenge system pipelines in unexpected ways.
Thanks to data collection, analytics, and artificial intelligence technologies, the life sciences industry now has powerful tools available, such as Smart Data Quality (SDQ) and Smart Auto Mapper (SAM) from Saama, to handle this kind of complexity and regulatory bodies in fact encourage their use.
These innovations create a transformation synergy by applying the latest advances in machine learning and natural language processing and leveraging structured and unstructured assets from historical trials.
OSP: What advice might you offer sites and sponsors who are hesitant to adopt data automation?
SA: I’d have to say that the sooner pharma embraces automation, analytics, and AI, the easier it will be for them. The data explosion is unstoppable, and any further delays in ClinTech implementation will continue to lengthen lead times, increase headcount, create milestone delays, and put quality at risk.
OSP: Then, please talk about how advanced analytical tools like ML and AI can help improve such processes.
SA: At Saama, we’ve been making great strides at solving the inefficiencies of clinical data management and standardization with scalable automation, analytics, and artificial intelligence technologies that actually work in complex, experimental environments.
Saama’s Smart Auto Mapper (SAM) solution uses algorithms to come up with study rules, perform edit checks, and generate and answer queries using machine learning techniques that instantly identify patterns and discrepancies in massive datasets across multiple sources.
When algorithms do the most burdensome work within SAM, data managers are freed up to review anomalies and investigate issues that the algorithms can’t finalize on their own. Not only does this accelerate data management while increasing overall quality, having humans in the loop at every stage trains the AI models for even better performance over time.
SAM can also provide a single source of truth for standards and transformation metadata, which makes it much easier for standards and study teams to collaborate. Transformation metadata can be set up and updated much more efficiently, resulting in near-real-time standardized data for review, analysis, and reporting. Submission package generation is also improved, resulting in higher-quality standardized datasets and a reduction in the time required to respond to regulatory queries.
OSP: How can a company like Saama help?
SA: Saama has a tradition of partnering with sponsors to solve specific problems and making those solutions available to the entire industry. Our work in artificial intelligence, and machine learning in particular, is revolutionizing the life sciences industry in many ways and is currently making a huge impact on clinical data management and programming.
Sponsors that really want to benefit from the increasing volume of data from disparate sources–without getting bogged down by the complexity of it all–are smart to embrace AI-enabled clinical research solutions across therapeutic areas. Saama is here to make the technology implementation as smooth as possible today and to continue innovating far into the future.
OSP Do you have anything to add?
SA: AI technologies enable life sciences companies to get valuable medical treatments into patients’ hands much faster. Sponsors who don’t embrace AI now will only fall further behind.