Virtual cohorts: an innovation in clinical trials.
ALIA SANTé
Find out all about virtual cohorts.
Reading time: 5 minutes
Clinical trials play a central role in the development of new medical treatments. Historically, they have involved the physical participation of volunteers, often resulting in high costs, long delays and limited access to diverse samples. Today, with the advent of virtual cohorts, a new era is beginning in this field.
In this article, we will explore what virtual cohorts are and understand the significant advantages they offer to modern clinical studies.
What are virtual cohorts?
Virtual cohorts are created from synthetic data, generated by advanced artificial intelligence algorithms. These algorithms are trained on real databases to produce virtual databases that faithfully reflect the original population. Synthetic data generation models retain the statistical properties of the population under study, while preserving the integrity of medical information and the realism of patient profiles. So, although these “new” patients are virtual, they remain consistent and representative, reliably reproducing trends observed in real patients.
This makes it possible to enrich clinical studies by creating virtual cohorts capable of complementing physical samples, particularly when it is difficult to obtain a sufficient number of participants. These virtual patients, while generated from a real sample, increase the robustness and diversity of clinical trials, while optimizing the time and costs involved.
How do you validate a virtual cohort?
Virtual cohort validation is a crucial process for ensuring the reliability and validity of synthetic data used in clinical trials. The validation of a virtual cohort is based on two essential steps: the quality of the real database and the validation of the synthetic data generated.
- Actual database quality
The first crucial step is to ensure that the actual database used to train the AI models is medically relevant and representative. A poor-quality database would introduce biases that would be reflected in the virtual cohort. To guarantee the validity of this database, the intervention of medical experts is necessary. They validate that the data accurately reflect the target population, both demographically and clinically. A non-representative or poorly structured database could lead to biases in the virtual cohort, compromising the validity of studies carried out with the latter.
- Virtual cohort validation
Once the synthetic data have been generated, it is essential to validate that the virtual cohort is realistic and representative of medical reality. To this end, a statistical quality report is drawn up, verifying that the virtual data faithfully reproduce the clinical and demographic characteristics of the population studied. This report compares the distributions, correlations and results observed in the virtual cohort with those of the real database. At this stage, human intervention is crucial to ensure overall validity. Domain experts, including clinicians, biostatisticians and medical data specialists, play a key role in the analysis of these results.
These experts intervene to assess the clinical relevance of the virtual patients created. They check, for example, that the health trajectories of simulated patients are realistic, that the comorbidities observed are consistent with current medical knowledge, and that responses to treatment follow established patterns of clinical practice. Their expertise ensures that the virtual cohort remains anchored in a reliable medical reality, preventing the introduction of bias or erroneous conclusions.
This human validation provides an additional layer of security, ensuring that the results simulated by the virtual cohort are credible and usable for clinical studies. Their critical judgement ensures that the virtual cohorts represent not only a statistically correct population, but also a medically valid population in line with the realities of the clinical field.
Finally, the final validation consists of comparing the results obtained from the virtual cohort with those of a previous study carried out on a real population. If the results of the two studies converge significantly, this confirms the reliability of the virtual cohort and, by extension, of the technology used. This comparison ensures that virtual cohorts can be used to complement or even replace certain parts of traditional clinical studies.
Advantages of virtual cohorts in clinical studies
Cost and time reduction :
Traditional clinical trials are often very costly, and can take over a decade to complete the development of new drugs. The use of virtual cohorts, thanks to synthetic data, can significantly reduce these costs and delays. By reducing the need to recruit a large number of participants for the control group, financial and logistical resources are optimized.
Adaptability to rare diseases :
Clinical trials on rare diseases frequently encounter recruitment difficulties due to the small number of patients available. Synthetic data provides a solution to this challenge, enabling the creation of virtual cohorts that faithfully reproduce real patient profiles. These virtual cohorts make it possible to broaden the sample, while remaining representative of rare disease patient pathways. This facilitates the evaluation of new treatments and speeds up clinical trials, despite recruitment limitations.
Solving recruitment challenges :
Recruiting participants for traditional clinical trials is often complicated by strict criteria and a limited number of available patients, particularly in the case of rare diseases or when the logistics of accessing study centers are difficult. Virtual cohorts offer a solution to this problem. By enabling the creation of synthetic patients, they complement studies where participants are lacking. This is particularly useful when the number of real volunteers is insufficient to achieve a statistically significant sample size. Virtual cohorts can then enrich trials by increasing the number of patients in a reliable and representative way.
Increased trial efficiency :
Virtual cohorts play a key role in improving the efficiency of clinical trials. Even before launching a traditional clinical trial, virtual cohorts can be used upstream to simulate different scenarios, thus “derisking” the study. By testing and adjusting hypotheses or protocols in a virtual environment, researchers can identify and correct any flaws before committing significant resources to real patients. This not only improves study preparation, but also reduces the risk of failure.
Challenges and limits of virtual cohorts
Although virtual cohorts offer considerable advantages, a number of challenges still need to be overcome, particularly with regard to their acceptability. Scientists and regulators have yet to adapt to this new technology, and tests are currently being carried out to prove the reliability of synthetic data. These studies aim to demonstrate that virtual cohorts can reproduce results comparable to those of real patients, which would reinforce their legitimacy in clinical trials. The regulations governing virtual cohorts are still being drawn up, but are evolving rapidly. This testing and adaptation phase is essential to ensure that synthetic data can be used with confidence in the future, offering a promising new avenue for clinical studies.
In conclusion
Virtual cohorts, based on synthetic data, represent a major advance in the field of clinical trials. By overcoming recruitment challenges, reducing costs and delays, improving trial efficiency and opening up new opportunities for rare diseases, they pave the way for faster, more flexible and inclusive medical research. Although challenges remain, particularly in terms of acceptability and regulation, ongoing trials and evolving regulations are boosting confidence in this technology. As virtual cohorts gain in recognition and validity, they are positioning themselves as an essential tool for transforming the future of clinical trials, making research more accessible and efficient while meeting the highest scientific and ethical standards.