AI and effective data management have the potential to revolutionize clinical studies.
By addressing administrative challenges, such as digitizing and automating study data, AI improves study design, compliance, and drug approval processes. AI's ability to analyze complex data, combined with real-world evidence, enables regulatory authorities to monitor drug safety and effectiveness more closely. AI facilitates patient selection, recruitment, and site identification by screening large datasets and predicting patient response and site performance.With AI-enabled data management, data analysis becomes quicker, decentralized, and interoperable, leading to faster decision-making and increased study flexibility. The combination of AI and effective data management holds great promise for a more efficient and personalized approach to clinical studies and drug discovery.

‍

In a previous article, we presented the main challenges with respect to conducting clinical studies, categorized as administrative, human and data challenges.
Here, we will explain how AI-enabled technologies and effective data management can tackle these challenges and improve the way clinical research is currently conducted.

AI has the potential to transform many key steps in clinical trials, from protocol design to study execution, thereby improving trial success rates and reducing development costs and time¹.The implementation of innovative AI technologies can speed up drug discovery and preclinical stages and improve the agility of the research process by a factor of 15².

AI’s main contribution is the ability to organize and analyze the increasing amount of complex, siloed data collected in clinical studies. To understand how this data management ability can tackle the different clinical studies challenges, we are now going to outline the contribution AI brings to each of the three categories of challenges previously mentioned.

‍‍

AI vs. ADMINISTRATIVE REASONS

‍‍

Digitization, organization and automation of study data can, in a first instance, help in drug approval, by enhancing the preparation of the study protocol and by making easier the gathering of necessary documents such as case report forms (CRFs) and patient informed consents. Thanks to a better study design, compliance risks, approval delays and subsequent costs can be optimized.

With the aid of AI-digitized, organized, and automated study data, coupled with real-world evidence (RWE), regulatory authorities can now closely scrutinize and monitor the safety and efficacy of drugs in a wider population, extending their oversight beyond mere drug approval.
This is an important field of contribution considering that ⅘ of trials fail because of inability to demonstrate efficacy and/or safety³. Given the ever increasing cost of bringing a new drug to the market, also a small percentage of improvement in the accuracy of predictions on the efficacy and safety of drugs, could save billions of dollars spent on drug development.

‍

AI vs. HUMAN REASONS

‍

When it comes to patient selection and recruitment, AI can effectively screen large patient datasets and execute the three clinical trials enrichment strategies defined by the FDA^4,5:

• Reduce population heterogeneity by harmonizing and mining data with different formats and levels of accuracy, coming from EHRs and other unstructured data sources, and automatically extracting meaningful insights to identify participants with matching criteria for a specific clinical study. 
• Select high risk patients/ with higher probability of having a measurable disease-related endpoint, namely prognostic enrichment.
• Select patients more likely to respond to a treatment than other patients with the condition, namely predictive enrichment.

AI can also enhance the identification of sites with the highest patient recruitment potential, by drawing on historical operational data from previous trials, including the failed ones. This allows not only faster decision-making regarding the choice of the site but also potentially to predict site performance in the future.

By collecting data from wearables, apps and sensors, AI can provide real-time patient monitoring, detect non-adherence and predict the risk of dropout for an individual patient. As patients are informed and supported in real time with these connected solutions, patient retention can also be strengthened, together with fostering patient empowerment.

As the main role of AI technologies is to support researchers and other clinical stakeholders and not to replace them, their contribution in the field of study team coordination is related to opportunities of easier data accessibility and shareability, soon explained.

‍

AI vs. DATA REASONS 

‍

AI-enabled data management systems have the unique capability to analyze semi-structured and unstructured data at unprecedented volume and velocity.
With AI, it is possible to aggregate complex and fragmented data collected from different locations in clinical studies, including real word data, into a single, unified data environment, making data management a quicker, dynamic and seamless process.

Once the study team has access to a single unified data environment, which also includes data from previous trials to improve future designs, data analysis is decentralized across study stakeholders and the analysis time can be reduced from years to weeks or even days.

With AI data management systems loading data coming from different countries and sources in one location but ensuring they are accessible from anywhere, interoperability is realized and collaborative research is fostered.  Consequently, faster GO/NO GO decisions are enabled and a clinical study’s flexibility is increased, allowing it to shift to other pipeline opportunities earlier in the drug development timeline.

Even if data can be shared between stakeholders, the data privacy remains protected as data is generally stored in cloud-enabled platforms, where cybersecurity policies and standards are automatically implemented, in order to adhere to regulatory requirements and patient privacy.

AI and effective data management are fundamental assets for the future of clinical studies and drug discovery: pushing towards a more time and cost-efficient process, they open the way for detailed insights to precision medicine and more personalized, targeted treatments.

‍

‍

References

¹ Deloitte. Intelligent Clinical Trials. (2020). https://www2.deloitte.com/content/dam/insights/us/articles/22934_intelligent-clinical-trials/DI_Intelligent-clinical-trials.pdf

² Deloitte. Intelligent drug discovery. (2019). https://www2.deloitte.com/content/dam/Deloitte/ch/Documents/life-sciences-health-care/deloitte-ch-en-intelligent-drug-discovery.pdf 

³ Dowden, H. and Munro, J. (2019). Trends in clinical success rates and therapeutic focus. Nature Reviews Drug Discovery, [online] 18 (7), pp.495–496. https://www.nature.com/articles/d41573-019-00074-z .

⁴ Bhatt, A. (2021). Artificial intelligence in managing clinical trial design and conduct: Man and machine still on the learning curve? Perspectives in Clinical Research, 12(1), p.1.

⁵ U.S. Food and Drug Administration (FDA). (2019). Enrichment Strategies for Clinical Trials to Support Determination of Effectiveness of Human Drugs and Biological Products Guidance for Industry. https://www.fda.gov/media/121320/download 

‍