Unlocking the potential of autonomous clinical research

Artificial intelligence (AI) technology has revolutionised how we approach life, work, and social interaction. In the healthcare industry, AI and machine learning (ML) have caused an unprecedented shift in clinical developments. The increasing use of automated technology, such as orchestration, digital workflows, and the capturing of data in digital forms have already started to make leeway into what may come to be known as the self-driving clinical trial.

As advanced tools continue to offer greater personalisation of treatment for patients, while making trials run more efficiently than ever before, we’ve seen remarkable results for pharmaceutical companies, medical organisations, and, most importantly, patients. Let's take a closer look at the implications these advancements are having on the clinical trial landscape.

Comprehensive study design: The key element for a successful trial

Effective clinical research starts with thoughtful study design. Establishing strong protocols is essential to ensure that a trial is cost-effective, successful, and yields accurate results. AI and ML enable the production of an ideal and productive design which can be implemented with little to no manual input, making it possible to set up studies quickly and precisely, while avoiding mistakes. Once that is done, trial organisers can assess the results of the validation process and make any necessary changes before the study is ready to run.

Advanced technologies can also help analyse different research options to determine which is best for stakeholders, including regulatory bodies, insurers, and patients. AI can further evaluate which countries and research sites are most suitable for the study and how to optimise participant recruitment and study launch processes, based on past data using generative AI to ‘Build your own Clinical Trial Set-up’.

Once the research design has been established, the focus shifts to the organisation of the trial. AI and ML can lighten the load for coordinators by aiding in the following:

• Site & patient identification: A major challenge in clinical trials is finding sites with adequate amounts of suitable participants that both satisfy the set criteria for inclusion and exclusion from the study. As trials become more specific, recruitment goals can become harder to reach, leading costs to increase, timelines to extend, and the likelihood of success to diminish.

  However, with AI and ML, organisers can minimise the risk of unsuccessful recruitment by employing mapping technologies to identify patient populations and proactively target potential study sites with the highest probability of recruiting the ideal subjects. Such models can be tailored to take into account different requirements, such as patient safety, compliance, and data quality.

  On top of that, data collected from previous studies can be applied to the model, enabling it to make informed predictions on the effectiveness of the sites for the new project, allowing sponsors to open fewer sites, expedite the process of recruitment, and reduce the likelihood of an insufficient participant pool.

• Pharmacovigilance (PV): In PV (the science and activities related to detecting, assessing, understanding, and preventing adverse effects associated with drugs), large volumes of structured and unstructured data need to be combined and examined to ensure quality control. Technologies such as optical character recognition (OCR), natural language processing (NLP), and deep neural networks aid organisers in managing data.

  Additionally, AI and ML automate time-consuming processes, such as converting and digitising safety case processing and adverse drug reaction (ADR) documents for risk and study performance evaluations. These tools can spot side effects by tracking conversations on social media and other sites, helping project leads optimise patient safety, while efficiently executing tasks.

• Monitoring risk: Clinical monitoring is an important part of conducting research; it serves to identify and mitigate risks associated with patient safety, data accuracy, and protocol adherence. However, it is labour intensive and can take up a lot of manual effort to assess risks associated with multiple sites. By applying AI and ML, the burden can be reduced considerably, given predictive analytics take over analysing the risk environment to generate more useful insights - such as which sites may experience recruitment and performance issues, or which patients may be more likely to experience adverse events.

  Utilising cutting-edge technology has enabled the creation of algorithms that can process medical data, such as symptoms and treatments, to recognise symptoms earlier on. This technology can quickly analyse a patient's information, suggest necessary steps, and promptly alert the doctor for additional evaluation. This type of early signal detection promises quicker and more proactive care for patients, as well as more comprehensive research into diseases like Alzheimer’s, where diagnosis usually occurs only after the condition has already advanced significantly.

• Real-time quality review: The more clinical trials can run with digital content the more that can be checked in real-time against expectations. Examples include the collection of all data sources, including documents and signed content, and even video and voice – all can be converted into a digital content data lake that allows natural language processing and generative AI tools to run through it, identify the content, and interpret it into insights and actions. This may include checking that content flowing into the trial is correct, signed-off, version controlled, and readable. It can also lead to ensuring the ‘chain of evidence’ where its feasible to ensure investigators have read the latest protocol, the protocol feeds downstream into building out the trial, and results from the trial link back to the protocol, and forwards into the subsequent e-CTD filing - all tracked and QA reviewable through the content data lake.

Harnessing technology to optimise clinical research

How we conduct research, drug discovery, diagnosis, and patient treatment have been significantly transformed for the better due to the implementation of advanced technologies. Decisions are now more precise and accurate, with positive and specific outcomes as a result. To reap the benefits of these technologies, their use must be equipped with global access, robust security, and reliability that complies with the governing regulations.

As we move forwards with the advancement of automated clinical trials, AI needs to be objective, unbiased, and trustworthy. Only then will the prospects of faster and more efficient research be attained.