Rethinking Clinical Trial Design
07/21/21
PwC and AWS collaborate to help pharmaceutical companies optimize clinical trial design decisions with AI-enabled insights and intelligent automation.
For pharmaceutical companies (sponsors), bringing a drug to market is a lengthy, costly and complex process. Studies estimate that it takes up to 9 years and costs $1.3B on average to bring a new drug to market.[1]
While sponsors face intensifying pressure to meet speed and efficiency goals, trials are also becoming more complex. This is partly due to an increasing focus on factors such as diversity and inclusion, rare diseases, decentralization to reduce patient burden and the use of real world evidence for synthetic control arms. The global COVID-19 pandemic has accelerated these trends, pushing sponsors to adapt and evolve at a rapid pace.
This increasing trial complexity can be seen in clinical trial data volumes and cycle times. For example, Phase 3 trials now collect three times more data compared to 10 years ago. Similarly, the number of procedures performed has increased by 44% since 2009.[2] And the average clinical phase for trials has increased by ~6.7 months, compared to trials conducted 10 years ago.[3]
Based on our experience, most of these issues can be addressed early in the clinical trial lifecycle - during the design phase.
What does clinical trial design look like today?
Pharmaceutical companies are struggling to design trials in a repeatable, scalable manner. Outside of healthcare, most industries are using large datasets and AI/ML to optimize business operations on a day-to-day basis. But in the pharmaceutical industry, once a clinical trial is completed and submitted, the data is rarely seen or utilized again.
An incredibly rich dataset on protocols and their amendments, operational trial data, patients, safety risks, and learnings from the trial lies in a server somewhere - inaccessible and forgotten. Instead of leveraging insights from historical protocols and trial data, clinical asset teams largely rely on institutional memory, medical expertise, and intuition to design clinical trials.
And it can be difficult to wrangle the data that is available. When teams seek to analyze historical data for insight generation at scale, it can become a Herculean task - given the many disparate data sources that have not been integrated or standardized. As a result, analytics and scenario modeling for trial design is conducted on an ad-hoc basis, generally reserved for failed or unexpected outcomes, or high stakes trials.
We believe intelligent automation and AI-enabled insights can change this.
What is the clinical trial design of tomorrow?
PwC and AWS are working together to bring AI-enabled analytics and automation to enable pharmaceutical companies to design clinical trials that can be executed faster, more efficiently, and with better patient outcomes. With access to insights from historical clinical trial data and real world evidence, pharmaceutical companies can make more informed decisions across the clinical trial design lifecycle.
This can be enabled by a foundational technology platform for intelligent automation built on AWS services:
The example use cases above are enabled by a foundational technology platform for intelligent automation built on AWS services:
- Data ingestion pipelines - Ingest historical clinical trial data, publicly available datasets (e.g., ClinicalTrials.gov, CTTI, PubMed), and real world data for analytics and insights
- Text extraction pipelines - Use Natural Language Processing (NLP) pipelines built on AWS Comprehend Medical, to extract key entities from clinical trial documents (e.g., protocols and amendments, statistical analysis plans).
- Data transformation & standardization models - Leverage pre-built models for data standardization, based on industry standards such as OMOP, SDTM, and ADaM.
- AI model deployment - Predict the impact of trial design decisions on downstream metrics, such as trial costs, cycle times, feasibility, and quality risk - using models built and deployed on AWS SageMaker.
- Intuitive user experience - Surface insights from historical trials and enable users to create, compare, and optimize trial design scenarios using human centered design.
How will these technology capabilities enable intelligent clinical trial design?
Let us illustrate with a user journey (see Figure 1).
With insights from historical data, real world data, and predictive models, sponsors can make informed decisions during clinical trial design and proactively plan for and mitigate potential downstream scientific, operational, and quality risks.
How will this impact sponsors' bottom line?
We believe that scenario modeling and optimization can be an integral step in designing every clinical trial. And our clients have realized significant value from intelligent automation solutions for clinical trial design:
- Lower Resource Constraint - Many clients have seen a 30% - 40% reduction in manual effort using intelligent automation.
- Reduced Patient Burden - Schedule of activities and eligibility criteria are optimized to reduce patient burden and reach more diverse, representative patient populations
- Shorter Trial Timelines - Trial designs are optimized to improve enrollment timelines and reduce the likelihood of protocol amendments.
- Reduced Overall Trial Costs - Schedule of activity design, country allocation, and site selection are optimized to reduce costs.
By making more informed decisions at the beginning of the clinical trial design and planning lifecycle, sponsors can proactively reduce the risk of costly delays and protocol amendments - ultimately accelerating time to market and bringing lifesaving treatments to patients faster.
Are you ready to rethink your clinical trial design process?
PwC and AWS are working together to help clients leverage intelligent automation and AI-enabled insights for Intelligent Clinical Trial Design.
If you’re ready to rethink your clinical trial design process, we are here to help.
[1] CB Insights: The Future of Clinical Trials
[2] Tufts Center for the Study of Drug Development: Rising protocol design complexity is driving rapid growth in clinical trial data volume
[3] Tufts Center for the Study of Drug Development: Faster New Drug Approval Times Are More Than Offset by Longer Clinical Times in U.S., According to Tufts Center for the Study of Drug Development
