The rapid development of Covid-19 vaccines, although under extraordinary circumstances and with emergency approvals from regulatory bodies, has set new expectations for drug development. We believe the time is right to leverage advances in technology and data to meet these expectations. Not only will this shorten timelines, but it will also decrease the overall development costs for new therapies and reduce patient burden.
This change towards a technology and big-data-driven approach to drug development will benefit patients and is a huge opportunity for life science companies. Our recently published whitepaper, co-authored with drug development AI and clinical data science specialist Phesi, and with contributions from Dr. Michelangelo Barone of Alfasigma, outlines how synthetic data in clinical trials can be a key accelerator. To be very clear, there is nothing synthetic about this kind of data; it is real patient data taken from previous clinical trials, clinical research and from electronic medical records (EMRs). The only reason we call it "synthetic” is because the data is not being gathered de novo as part of the current trial.
<<< Start >>>
<<< End >>>
(Re-)using existing data in clinical trials: Are we wasting potential?
The volume of data created through clinical trials, clinical research and EMRs has grown exponentially. Sponsors of clinical trials also have a lot of historic data from previous trials. Biopharma companies already use this data for study design and, more specifically, to model the target patient population and investigator sites to involve. But the opportunity is much bigger. More data relevant to each trial exists that may not be obvious or accessible to individual sponsors. And with the rapid evolution of integrated analytical capabilities, additional value can be achieved with the curation of existing data.
Here’s where synthetic data comes into play as a true game changer in clinical development. Modeled properly, it can simulate a synthetic patient profile or simulate views of the entire control patient population in a trial. This optimizes the design and feasibility of a clinical trial thereby increasing operational success, while simultaneously saving time and resources.
Furthermore, the control (placebo) arm can be significantly reduced or in the best case, fully replaced by synthetic data. That’s possible wherever we can define patient populations either across clinical trials or from EMRs with sufficient similarity. We know that for many indications, trials run by different biopharma companies often show statistically identical results in the control arm. So why should a sponsor run yet another control arm when the data already exists?
<<< Start >>>
Synthetic data helps patients benefit directly from new, innovative – and in some cases life-saving therapies – much earlier.
<<< End >>>
How trial sponsors and patients benefit
For patients: Many clinical trials come with ethical concerns where there is no established standard of care for the control arm; moreover some patients do not wish to participate in a trial when there is the possibility of not receiving the new treatment. Leveraging the data from very similar trials to reduce or, in some cases, fully replace the control arm, avoids exposing patients to unnecessary burden and risk. In addition, synthetic data helps to speed up the development of new drugs, or new indications for existing drugs. Patients benefit directly from new, innovative – and in some cases life-saving therapies – much earlier. The opposite is also true- patients can be prevented from being exposed to trials that are never likely to yield positive outcomes.
For sponsors: Recruiting patients for a trial is a time-consuming and costly process. With synthetic data, there is a unique opportunity to reduce the number of patients. In addition, using historic patient data as well as information from EMRs will optimize trial design and therefore increase operational success. In oncology, for example, it is common to conduct single arm trials in the early clinical development phase. Synthetic data can be used to better explain the results from those single arm trials, minimizing the risk of detecting false positive signals.
For health care professionals (HCPs): Additionally, HCPs benefit by gaining access to new treatments sooner, and having more treatment options for patients.
Ultimately, it benefits everyone if new innovative therapies reach patients faster.
Access to data and the view of regulators
We recognize the road to greater use of synthetic data is challenging. Access to data in great enough volumes and with significant enough details at a patient level is a key challenge.
In addition to the access challenge, we often hear arguments about regulatory authorities and health agencies being too conservative in their approach to synthetic data. In our eyes, that’s simply not true. There are a growing number of examples in the EU and the US where synthetic control arms based on historical patient records have been approved by the regulators as full replacements for a real control arm. In the future, we foresee regulators actively demanding the use of synthetic data in clinical trials to avoid exposing patients to a placebo or inadequate standard-of-care treatment. This ethical argument will become more important once viable alternatives, such as robust synthetic control arms based on relevant data, are more widely implemented.
Let’s seize the synthetic data opportunity
The Covid-19 pandemic has proved that speed matters when it comes to delivering innovative treatments against serious diseases. In a post-Covid-19 world, even more emphasis will be put on reducing both the time and costs needed to run a trial without compromising today’s high quality and safety standards. We believe that properly chosen synthetic data is a key part of the future drug development and commercialization operating model; if used to reduce the demand for patient recruitment, maintain quality, and to optimize design and feasibility, it will speed up trials while delivering robust results.
If you’re interested in further insights and examples on the use of synthetic data in clinical trials, download our latest paper on this topic.
The authors wish to acknowledge the contributions of Dr. Sandra Dietschy-Kuenzle, Phesi and Dr. Michelangelo Barone of Alfasigma for their continued support of our shared perspective.