CWWeekly’s semi-monthly company profile feature, Insider Insights, interviews executives of companies and organizations in the clinical trials space. Writer Ronald Rosenberg sat down with Adam Hanina, CEO of AiCure.
Q: AiCure uses novel facial recognition to confirm medication adherence on smart phones and tablets. Which therapeutic areas are best suited for this type of monitoring in clinical research?
A: Any clinical trial will fail if participants don’t take the study drug. The quality hinges on getting accurate information about whether the trial participants are adherent with their medication. And that information also affects the cost and speed of completion of the trial. Accurate adherence monitoring is crucial. So crucial that, in phase I trials, they use the costly and somewhat intrusive directly observed treatment, or DOT system, to guarantee that patients across all disease categories really are taking the medication.
AiCure offers the same level of monitoring, which can be used beyond phase I trials. Our novel facial recognition technology emulates the process of directly observed treatment, visually confirming that the right patient is taking the right medication at the right time. There is no change needed in the medication manufacturing process.
We have designed our patented technology specifically to be very easy to use with standard hardware devices such as smart phones and tablets. This means it can be used by participants in the privacy of their own home, while providing real-world verification that they ingested the medication.
All disease categories are affected by sub-optimal adherence rates. Our technology platform was funded by the National Institutes of Health (NIH) to standardize medication adherence monitoring across all clinical trials and disease states. The platform currently is being deployed in phase II and phase III trials and NIH studies covering schizophrenia, stroke, bipolar disorder, depression and substance abuse.
Q: What have you learned about dosing behavior: Do patients change the time they take their medication or skip doses? Do they try to cheat? Do they swap their pills?
A: We’ve learned that there are different categories of patient behavior. Patients miss or skip doses. They take doses at different times. And sometimes they don’t take their medications for periods of time. According to the data we presented at the American Society of Clinical Psycho-pharmacology meeting in June 2014, patients fall into these specific categories of adherence behavior. What was revealed was that early behavior predicts future behavior. So if patients are erratic in their dosing in the first few weeks, they will remain erratic throughout the trial. You need the accuracy of visual confirmation to be sure of proper administration.
If a patient is intent on not taking the medication, then he or she will try to cheat, regardless of the monitoring system, technology or person. That’s why the NIH awarded AiCure $3.4 million in funding to address this very problem.
We’ve developed sophisticated algorithms to measure malicious intent. The platform uses continuous video analysis and computer vision software to monitor the entire motion of patients as they take their medication and to confirm their identity. It then confirms that the medication is correct and observes the patient’s actual process of ingestion. For oral medications, such as tablets and capsules, we monitor if the patient has placed the medication in his/her mouth and has drunk a glass of water, and finally we check if his/her mouth is empty.
For sublinguals, for example, we confirm m the medication has dissolved under the tongue. This is exactly how directly observed therapy is performed in phase I clinical trials.
Any movement out of the field of view, suspicious behavior or tampering with the medication immediately is flagged in real time. We monitor the entire process of medication administration and that includes duplicate enrollment and spitting out the pill. This gives us a complete view of what’s happening and whether someone is intending to cheat the system.
Q: How will your company either differentiate itself from or work with other technologies such as “smart pills” in the market?
A: We all are striving for similar things, namely highly accurate adherence monitoring and the option of intervening quickly with non-adherent participants. But there are differences. Our system is easily implemented and works across all clinical trials, whatever the disease category and route of administration.
We want to avoid the need for changing the medication manufacturing process. We’re sensitive about overburdening patients by making them wear a patch or ingest a microchip. We chose an easy-to-use system that is easily adopted by participants and relies on smartphones and tablets that they’re already familiar with. In addition, clinical trial costs have skyrocketed. We want to make sure the technology is affordable and can be deployed on a global scale. So we’re willing to work with any hardware-based monitoring company where the synergy would be advantageous for sponsors.
Q: What are some of the financial, logistical and other challenges faced by sponsors, CROs and sites regarding monitoring adherence, and how have you addressed them?>/p>
A: They are worried about high costs, recruitment and patient retention issues s, along with regulatory challenges due to increasing pressure for data transparency. All of these issues are affecting their decisions about how to monitor medication adherence.
Until now, sponsors haven’t had much choice but to rely on blister packs, which are routinely returned 100% empty, or expensive blood and urine tests, which are done only intermittently and provide only a snapshot of adherence. Beyond those two approaches, electronic patient diaries and text messaging have been used to assess medication adherence, but these have the obvious shortcomings that come with self-reporting.
While each of those systems has its merits, our design addresses these challenges by being accurate, affordable, easy to deploy and accepted by the patients themselves. The platform enables real-time intervention with non-adherent patients, reducing costly patient dropout and loss to follow-up.
Q: What have you learned about patients who are unintentionally or intentionally non-adherent about taking their clinical trial medication?
A: You need a monitoring approach that is, first and foremost, accurate. You can’t rely on self-reporting. You need to do visual monitoring in real time. Once that’s in place, you can start to predict how individual patients will behave and correlate that with interventions and outcomes.
With our platform, a CRO or site gets adherence data in real time. Built-in algorithms then trigger appropriate responses, ranging from an automated message in the case of a single missed dose all the way to a study coordinator beingalerted to follow up with a patient.
With an estimated 4.5 billion smart phones in circulation over the next few years, software approaches to patient monitoring and intervention offer exciting possibilities for clinical trials of the future.