Out With the Old and in With AI

Despite being vital to new drug development, the clinical trial process is broken. Patient recruitment and retention provide such significant challenges that their impact is felt industry wide. More than three-quarters of all clinical trials experience delays and with the daily bill for this somewhere between $600,000 and $8 million, the desire to find a more efficient method sits right at the top of the agenda. In addition to the astronomical costs, the stark reality is that these delays often slow down the entry of new medicines into the market, preventing patients from accessing new and innovative treatments. Even when patients are willing to participate, the hurdles are still significant; not least because the whole business of clinical trials is shrouded in mystery with very few people understanding or acknowledging the role they could play by participating.  

As a highly regulated industry, the pharma market is often understandably cautious of change and, as a result, behaves relatively conservatively. AI and other innovative technology applications have the potential to disrupt every stage of the clinical trial process from matching eligible patients to monitoring and data collection. Whereas, the traditional clinical trial process is inefficient, time-consuming and incredibly costly, with drug development taking an average of a decade and, often, costing billions of dollars. It is time we made changes.

A great deal of time (and budget) has been spent on the development and implementation of Electronic Medical Record (EMR) solutions. Clinicians are able to “code” the diagnoses relevant to the patient’s care and this information can then be shared with other healthcare professionals. However, every diagnosis starts with a story – and this is why doctors write notes. Those notes are rich in detail around symptoms, relevant previous history, impressions and ideas about what might be wrong and plans to investigate and treat. This documentation process is an essential step for the physician as s/he organizes their thinking and decides what to do next. An enormous amount of useful insight lives within these notes but, from an EMR search perspective, they are unstructured and therefore impossible to automatically “read” and interpret. Today, that work is done manually by human researchers.

By searching this narrative data using AI, clinical trial recruitment, for example, can be significantly enhanced.  AI-driven software can dramatically reduce the time and investment currently required to find and enrol patients into clinical trials. By quickly and efficiently processing large volumes of existing unstructured patient data, sites are able to identify eligible subjects against trial-specific inclusion and exclusion criteria more quickly than reviewing patient data manually.

There has been a great deal of interest and enthusiasm for these types of technologies but as with any disruptive platform, it takes time for people to adjust and fully realize the benefits. However, even in the somewhat conservative pharma industry, there is a general consensus that change is on the horizon, so the adoption of such technologies is accelerating and outpacing the usual rate of change.

While AI is unlikely, at least in the short term, to replace conventional clinical trials processes in their entirety, it can remove some of the more repetitive or tedious tasks experienced by physicians and research nurses, promoting improved patient participation. There are many different creative streams in AI currently being explored and it is an incredibly exciting area to be involved in. There is no doubt that AI offers enormous opportunity, not only in clinical trials but also in clinical practice.

This in turn will expedite drug development and also change the shape of what we understand about disease and how we diagnose it. All to the benefit of the patient.