In the past few years, the volume of patient data created through clinical trials and electronic medical records (EMR) has soared, creating huge data lakes. The ability to store, process and analyze this data has also expanded, with technological advances in AI and analytics, cloud utilization and computing power enabling more ambitious use of patient data. These advances are occurring at a time when pharmaceutical companies are still using highly time-intensive and costly means to obtain data in the clinical trial phase of drug development. A key driver of the cost and duration of clinical development is the number of participants needed for each trial. This is particularly an issue for indications with high unmet medical need, such as those within oncology, or in rare diseases, where the limited number of patients make it even more difficult to recruit patients without delay.
Today, a synthetic control arm can be generated from historical data. Patients who have received standard of care in the past have electronic medical records that track their outcomes. Data on trial protocols and results from completed clinical trials, observational studies and data from registries can all be incorporated into a robust synthetic control arm or used to reduce the size of the comparator arm. Using synthetic data for the control arm substantially reduces the demand for patient recruitment, saving time and resources.
In addition to minimizing the need for placebo patient enrollment, synthetic patient data can be used to model target patient populations and to define the boundaries of a trial. This can optimize clinical trial design and feasibility to positively impact operational success.
These four building blocks are crucial for pharma companies to make strides in the usage of synthetic data tools.
1. Draft your plan
Create a bold global vision for leveraging synthetic data in clinical trials, prioritize assets to pilot the new approach.
2. Take stock of your data & look outside for more
Supplement internal data with external data sources to achieve a robust validated data set.
3. Deploy smart algorithms for “what-if” analysis
Integrate and automate algorithms into analytics platforms to evaluate and predict potential outcomes.
4. Evolve your operating model
New governance, skills and processes are needed so predictive data analytics can inform clinical development plans.
The use of a synthetic control arm instead of a patient cohort receiving standard of care reduces the patient burden of participation in clinical trials.
Getting the approach to insilico clinical trials right will enable deployment at a meaningful scale, leading to a more reasoned process for clinical development, with lower patient burden and treatments that are much faster to market, at a lower cost. The potential of this opportunity is transformative, and companies that take the innovation leap now will find themselves in a much better position to benefit from the wealth of data that is out there.
Thank you to Jonathan Peachey, Dr. Gen Li, Dr. Paul Chew, Dan Manak and Dr. Michelangelo Barone for your contributions to this report.
Listen to Accenture’s Boris Bogdan speak on synthetic patient data for clinical trials: