As discussed in Part One of this review on the potential of artificial intelligence (AI) and machine learning (ML) in the early stages of drug development, there is clearly a long way to go before they will genuinely make drug development cheaper, faster and more efficacious.
That said, steps are being taken in the right direction with an increasing amount of investment and funding being placed on AI and ML start-ups focusing on different stages of the clinical trial process. The hope is that despite the healthcare sector being inherently conservative and risk averse (most probably as a direct result of the Hippocratic Oath), by applying AI and ML to multiple stages of both drug development and clinical testing, improvements relating to costs, speed and efficiency will be seen.
Unsurprisingly the same technological systems being developed by drug development focused AI and ML, with predictive analytical tools relying on complex algorithms and big data sets, are being used in respect of clinical trials.
The hunt for the right clinical trial
Numerous drugs are delayed or prevented from ever making it to market due to failures to meet enrolment timelines or enrolment difficulties.
AI companies seeking to address such issues have proposed GPs and/or hospitals agreeing to AI software scanning and extracting key information from individual patient records and comparing such information against live studies, suggesting matches – essentially bringing the studies directly to patients that have genuine chances of qualifying on the study.
This is an interesting approach. However, before it can barely step off the start line, concerns are being raised in relation to data protection, data privacy and wider regulatory concerns in its application, and possible abuses.
Keeping interest and focus on point
An alternative focus for AI and ML developers focusing on improvements to clinical trials has been in respect of natural language processing and deep learning to better manage patient engagement. This is through the identification of key markers flagging patient disengagement from clinical trials and human intervention to reengage individual patients.
The impact of these technologies is unclear and highly questioned, with some arguing that the technologies, expert knowledge and funding would be better spent in other areas of AI and ML in clinical trials, notably drug development rather than clinical testing.
AI and ML provide a wide array of opportunities and potential benefits to clinical trials - from matching patients to studies, study enrolment, patient monitoring and adherence, managing workflows to data analysis. However, as highlighted regarding drug development, it is vital that appropriate consents and data protection security features are considered and applied throughout the development and application of these AI and ML technologies.
Further, as with drug development, companies using AI algorithms in connection with clinical trials must carefully assess the potential application of patent law (both in relation to protection of the AI technology used, but also in the invention derived using the AI technology).
What does the future hold?
The use and application of AI and ML in both drug development and clinical trials unfortunately remains behind other sectors, including transportation, finance and education. Until clearer guidance is provided by the relevant data protection regulatory bodies on how companies can push technological boundaries whilst ensuring patient data and consent is appropriately obtained and managed, both technological and medical advancements will be hindered and delayed.