How can AI transform sluggish clinical trials for the better?
Clinical trials are a vital part of the development of new treatments and drugs for life threatening conditions. This year, 10,000 clinical trials will take place in the US alone. The problem is that the $40bn market is notoriously slow and, as a result, expensive. Clinical trials take place in three phases, with each stage testing more patients. Unfortunately, outdated data techniques, confusion over eligibility, and high patient drop out rates can make the process last, on average, almost 10 years. This has a negative impact not only on the clinics trying to test treatments, but on the doctors and consumers who prescribe and use them too. Enter AI. How can the transformative technology bring clinical trials into the digital age?
1) Matching patients with trials
Enrolling in a clinical trial is far from easy. In many cases, patients don’t know if they are eligible, and even when they are the process can require considerable time and effort. This is partly why so few people – less than five per cent of cancer patients, for example – sign up for clinical trials. Patients often drop out part way through the trial because they are unable or unwilling to continue. Through AI, patients can be matched with trials via their health records. These pairings could be based on location, lifestyle, symptoms, previous medical results, and the severity of their illness. In theory, fewer people will drop out of trials that are deemed more likely to help their condition.
2) Tracking medication application
Traditionally, patients who take part in clinical trials are expected to record their own medication adherence. This comes with obvious problems. Instead of relying on patients themselves, Pfizer and Novartis are working on ingestible, AI enabled sensors that can recognise when a patient has administered their medication. If it is not ingested or taken at the right time, then the sensor issues an alert. Clinics and companies can therefore identify which patients have followed the advised administration of medication and which have not. This could help to explain anomalies and encourage accuracy. Catalia Health has released an AI healthcare companion and coach to influence behavioural change in patients via tailored interactions.
3) Feedback from biosensors
Making sure that patients actually take their medication and participate in the correct way is only the first step. The next hurdle is making sure that results are recorded accurately. Relying purely on patient observations and sporadic medical examinations to track symptoms is prone to error. However, biosensors like those developed by ContinUse Biometrics can monitor heart rate, glucose levels, and blood pressure to offer a real time, continuous profile of patients. Earlier this year, the company raised $20m in Series B funding. An embedded AI recognises anomalies, and flags abnormal readings, reducing the pressure on patients and clinics by providing them with more in depth information.
4) Understanding Electronic Health Records (EHRs)
Many clinical trial records are still written in paper diaries, by hand. Not only is this confusing to human clinicians but also to artificially intelligent technology. Digitising these handwritten notes will be crucial if AI is to reach its full potential. Once information has been standardised and made digitally available, it can be easily analysed and understood by intelligent systems. This will lead to the faster analysis of results, more accurate readings, and the elimination of human administrative errors.
5) Predicting drug effectiveness
By working through patient data, artificially intelligent technology can flag up anomalies, identify patterns and predict how effective a drug is likely to be. This depends on the patient, the disease, and their specific symptoms. While doctors are able to do this well, they cannot do it at scale. This is where AI comes in. Through computerised reasoning, artificial intelligence can take a huge set of EHRs and anticipate how a patient might respond to a treatment based on their medical history – and it can do it instantly.
AI will be hugely important in shaping the clinical trials market, as well as various other aspects of future healthcare. However, before AI can truly begin to address the challenges of the industry, the legacy processes and systems currently used in trials need to undergo digitisation. Following this, AI will be able to help rather than replace clinicians and healthcare professionals by providing insights at scale. Through adopting AI, clinical trials could go from a traditionally costly and time consuming endeavour to a precise, streamlined process that eases the burden on all parties. While the initial advantages are largely associated with the development of drugs, AI’s data crunching skills could equally be applied to any form of clinical trial. Reducing cost, time, and administrative effort will hopefully encourage more patients to apply for trials, and make it easier for them to do so at the same time.
Learn more about AI in healthcare by signing up for our free weekly newsletter.