2 ways AI can improve clinical trial efficiencies

Chris McCann Christopher McCann | 22nd January 2019

The potential benefits of artificial intelligence (AI) may be overhyped, but in medicine, there’s a consensus that AI could dramatically speed up the process of developing potentially life-saving drugs and therapies.

Producing and bringing new drugs to market is both time-consuming and extraordinarily expensive -- which pushes up the costs of U.S. healthcare dramatically. On average, the cost of new drug creation is approximately $2.55 billion, taking up to 10 years in some cases, according to the Tufts Center for the Study of Drug Development.

AI promises to positively disrupt every stage of the drug development process in particular: identification of molecular targets, improved matching of patients to clinical trials, predictive stratification of adverse events, and processing the volumes of data that these trials produce. The payoffs could be huge.

The following are two of the biggest areas where AI can improve the efficiencies and success rates of clinical trials. (I’ll post another blog later on two areas that are “under-hyped” opportunities for AI in clinical trials.)

1. Measuring and improving compliance

As soon as a participant leaves the hospital, clinic, or clinical research facility, they enter a black box. Researchers’ understanding of patient health after this point at an individual level is almost non-existent. Unless participants are required to remain within the research facility throughout the trial, researchers have no idea what a participant is doing once they leave or whether they are really following the pre-specified protocol.

This creates significant challenges. Researchers, pharmaceutical companies and CROs often don’t know whether their experimental drug is working as expected. For example, patient compliance to the intervention is difficult to account for, and this can create a mismatch between drug efficacy at different stages of clinical trials and in the real world. Do we really know whether the patient is taking the medication as prescribed? Are they doing so at the correct time? Are they making the correct lifestyle changes? Do they have an undetected risk of rare adverse events? We currently have limited objective data to assess these questions. This leaves limited context for critical decisions in the drug development process.

Combining AI with remote health monitoring, like the platform developed by Current, enables real-time, objective tracking of patient health at home. Researchers are provided the data to understand exactly what is happening with the patient, 24/7. We provide pharmaceutical companies far greater context. By adding medication reminders and a structured Q&A, we add transparency to these opaque situations - both for the trial participant and the researcher. Participants have an easier time knowing what to do and clinicians can better understand if they are adhering.

2. Precision, personalized medicine: a deeper phenotype

Population health is a major focus within healthcare technology, but due to the absence of the data discussed above, we lack deeper insights at the subgroup-level and patient-level. This obscures our understanding of the potential risks and benefits of a given intervention. In many cases, it is difficult to determine whether an adverse event was truly related to the intervention or some other root cause. This is a frequent question that only data can answer. Since trial patients are expensive, we want to get the most information out of each patient.

And yet, if researchers can identify which phenotypes have the greatest risk or benefit for a particular intervention, it enables them to better tailor them to specific patient subgroups. They can move ahead in a trial with greater confidence or a better understanding of the risks.

By collecting real-time, objective and subjective data from participants throughout the trial, we can understand a patient’s response to the intervention. But this is a huge volume of data, and it’s frequently unstructured. Current is helping take the first steps by having this key information de-siloed and well-structured so that the insights can begin. It’s these insights that will derive huge value. I’ve been amazed how quickly we’ve been able to take a single patient’s vital sign information and turn it into personalized insights. This crucial information would otherwise be lost in a population-based approach -- but first, you need to collect it.

Our platform enables large scale pattern matching within collected data to identify subgroups who most benefit and are most at risk to interventions. Through this, drugs can be tailored to the individual level - the true goal of personalized and precision medicine - and the efficacy of that drug at that individual level be better understood.

Toward digital therapeutics

While above we discussed the use of AI and remote health monitoring in the context of drug and therapeutic development, these tools can be just as valuable once a drug has been approved. Many pharmaceutical companies are now developing digital therapeutic and intelligent pharmaceutical teams. Every week it seems you hear another story of a big pharma company teaming up with a digital partner to solve big challenges.

By combining deeper phenotypical understanding of those who will benefit most from a drug, we can use remote health monitoring and AI to recommend the best drug intervention at the time. This moves drugs from a place of empirical prescription to objective prescription, based on that patient’s health and phenotype. This moves drug prescription and dosage into the data-driven domain. It provides context to help clinicians confirm a hunch or adapt their strategy of care and can better understand whether a patient is forgetting to report key symptoms or episodes.

At Current, we’re building our platform to do just this. We can support the progress and tracking of new drugs by deepening understanding of patient-level health, identifying patients who will most benefit or are at greatest risk, and enable more appropriate delivery of these medications at the right time.

Remote patient monitoring has the ability to transform the clinical trials and drug delivery process, help to potentially bring therapeutics to market faster, and enable earlier, proactive prescription once the drug reaches market. Interested in seeing how remote patient monitoring can help your clinical trial? Book an online demo today.

Chris McCann
Christopher McCann
Chris is the CEO & Co-Founder of Current. A Computer Science graduate, Chris founded Current as a medical student.

Lets talk about how Current can help you reduce readmissions and free staff time

Book a Demo