Correction: An earlier version of this article was unclear about differences between Abilify MyCite and the Proteus Digital Medicine Platform.
Earlier this week, Proteus Digital Health and Otsuka Pharmaceuticals announced that Abilify MyCite, a new, sensor-enabled version of Otsuka’s drug for schizophrenia, had received FDA approval for mental health conditions including schizophrenia and bipolar disorder.
Otsuka Pharmaceuticals received approval for Abilify MyCite, which utilizes much of the same technology as Proteus's FDA-cleared medication management and adherence system: both include sensor-enabled pills, a peel-and-stick biometric sensor patch worn on the body, and a companion smartphone app. For the newly-approved MyCite, the sensor, and a new bespoke version of the patch, are supplied by Proteus, while the app is designed by Otsuka.
It's important to note that Abilify MyCite is approved by the FDA only for monitoring ingestion -- not for improving compliance or adherence. This could possibly change in the future. Notably, Proteus's platform is labelled for improving adherence as of 2015.
A historic approval
For those who have been aware of Proteus for a long time, the importance of this latest regulatory win might not be obvious. It’s far from the first FDA nod for the technology; Proteus saw its patch first cleared in 2010 and its pill first cleared in 2012. Proteus has also been using its technology in small hospital deployments for some time.
“They had it so that you can go to take the pill, add a chip, and put it in a capsule, so it’s just now more integrated but it’s the same concept,” Dr. Eric Topol, founder and director of the Scripps Translational Science Institute, told MobiHealthNews. “I don’t know that that’s so transformative, really.”
Nonetheless, this is something new and potentially game changing. For one thing, previous clearances were for Proteus's technology. They went through the FDA’s medical device pathway. By contrast, this is the first time a sensor/drug combination has been approved as a pharmaceutical, and it will allow the system to be prescribed like any other drug.
“I think it’s a milestone in the industry because it’s the first time an adherence technology has made its way into the supply chain,” Dr. Joseph Kvedar, VP of connected health at Partners HealthCare, told MobiHealthNews. “One of the biggest challenges around adherence in general, and there are a few, one is that these technologies are not something that I as a practitioner can pull out and write a script for. So this brings it to the point where that could happen and I think that’s an important thing for us.”
It’s also a historic move for the FDA, which has never before approved a drug and software combination like Abilify MyCite.
“We are for the first time ever incorporating silicon and software into the definition of a pharmaceutical product,” Proteus CEO Andrew Thompson told MobiHealthNews. “This is extremely important because if you’re looking you might have noticed silicon and software are the technologies that are transforming pretty much any industry that’s out there. We have an opportunity now for the pharmaceutical industry to embrace those kinds of tools, to build them into their drug products and create a whole new window of opportunity for innovation that leverages the entire library of drugs that already have been discovered. And that is probably one of the most important opportunities in 100 years.”
Informa Pharma Insight analyst David Filmore noted that the move also demonstrates an ongoing change in the FDA.
“The approval reflects the FDA’s increasing comfort with and support for clinical-grade digital technologies,” he wrote in an email. “It was a challenging approval process, held up by questions about how the targeted patient population, those with mental illness, would safely and effectively employ the system, which includes a wearable patch and an app that monitors treatment adherence."
If Abilify MyCite is eventually labeled to improve medication adherence, it will represent a big leap forward in adherence monitoring, the current state of which leaves a lot to be desired.
“Blue Cross is right now tracking my adherence to my statin which I take nightly,” Kvedar said. “They track that through my refill rate. It’s from CVS and that’s a very crude indicator of anything. If you are on any medications, the foolishness you get from your insurer or your pharmacy will remind you how stupid that system is.”
And while other attempts have been made to measure adherence through technology, none are quite as accurate or hard-to-fool as a connected, ingestible sensor. These include smart pill bottles, pillboxes, and blister packs, and computer vision systems like AiCure. But Thompson doesn’t see those technologies as being in the same league.
“None of these other things measure adherence,” he said. “They just measure whether the bottle beeps or whether the blister went off or whether the patient says they did or didn’t take it. This is the first time there’s been a medical product that makes an actual claim that we measure it, and we actually do measure it. It’s not a hope or a claim or a hand wave, it’s actual real fact and if you can think of a way to get around the system, I’d like to hear about it, because we can’t figure it out. You have to ingest the drug to trigger the system.”
Finally, while much news coverage has focused on the pill, the approval also included Otsuka’s app, which was mostly responsible for the FDA turning down the tech the first time. Proteus and Otsuka see the app as much more than just a data readout for the pill.
“The real difference [from other adherence tools] is not just that we measure, but that we are making use of the tool they most prefer to use to manage their daily life — their mobile phone,” Thompson said. “That’s critical because it makes it completely automatic. So it’s a very easy, very low-burden way in which people gain insight into how they’re using medicines and how their body is responding.”
Andrew Wright, VP of digital medicine at Otsuka, echoed that sentiment, noting that better patient engagement is the real goal.
“I think what this will do is usher in an era of collaboration both amongst the physician, the patient, their family, and their care team,” he said. “We think this will lead to a more informed discussion rather than the assumption that the patient is taking their medication, and that can lead to a better understanding of the reasons for not taking it. And let’s face it, non-adherence is a normal behavior. So it just requires a collaborative discussion to try and understand the barriers and then work together to find a way forward. So it really is about greater engagement.”
From a narrative standpoint, the association of a tiny sensor in a pill and a mental health condition associated with paranoid delusions is an unfortunate irony — and one that hasn’t gone unnoticed. Stephen Colbert made the joke back in 2012, but Dr. Paul Appelbaum, director of law, ethics and psychiatry at Columbia University’s psychiatry department, asked the question more seriously in his interview with the New York Times about the latest news, saying that “drugs for almost any other condition would be a better place to start than a drug for schizophrenia.”
And there are other reasons it’s an odd choice. Kvedar pointed out that, as a first test case, Abilify may not return the clearest data.
“I think it’d be fair to say they had a lot of different irons in the fire for which would be the first tablet to have this technology, so I can only guess there were some business reasons Abilify floated to the top,” he said, “I know they were trying for some some hypertension generics and things like that. Again, I’m only guessing, but my guess is they would have been much more excited to get … one of those widespread drugs that you can quickly map to a cost benefit. The cost benefit of something like Abilify is a little more elusive.”
But Thompson and Wright told MobiHealthNews that schizophrenia is as good a place to start as any — and better than some, since mental health conditions have a much higher than average nonadherence rate.
“What we learned before we were working with Otsuka is that the mental health community was one of the most enthusiastic about embracing the technology and understood very quickly from a patient perspective, a family perspective, and a care team perspective, how valuable these tools can be,” Thompson said.
Wright, at Otsuka, finds the way mental health conditions are being discussed in relation to the drug unfortunate.
“In a way it plays to the stigma, which is disappointing,” he said. “I think the important thing is this is actually a discussion between the physician and the patient and ultimately, the physician decides if this is the right treatment for them.”
“What’s really important for people to understand is, not withstanding the stigma and the speculation people have about the nature of this community, they are incredibly motivated to help themselves manage this condition and they do it really well,” Thompson added.
Wright also said Otsuka plans to roll out Abilify MyCite gradually and carefully, as befits a new category of therapy.
“[For] our initial rollout, the launch is going to be conducted in close collaboration with only a select number of these health plans and providers,” he said. “We want to limit the availability because having fewer people using the system means we can really focus on learning from the patient experiences. And then through that ongoing feedback from using the system every day, we can further improve that experience for prospective users moving forward.”
But Proteus has plenty more use cases in the pipeline, with other pharma companies as well as Otsuka. Thompson said that the company’s hospital pilots will continue and will serve, as they always have, as a test bed to see how different patient populations react to the system.
“You can view it, effectively, as a learning lab that enables us to know and understand where digital medicines can have the most impact, and it allows us then to make the decisions about which medicines should most likely then come next in the fully integrated path,” he said.
A few examples of where the company is looking next: hepatitis-C and cancer. Thompson said that hepatitis-C drugs are so expensive that patients are often denied access because demographic factors make them too high a risk for nonadherence. Cancer patients face adherence challenges because of other stressors in their lives as well as complicated dosing regimens.
“Those are just a couple of areas where you have very expensive drugs,” Thompson said. “In one example, you have patients being denied access because they are adherence risks, and then in another area, you see really significant adherence problems that lead to significantly worsened outcomes.”
The privacy elephant in the room
Most sources MobiHealthNews spoke to for this article agreed that this was likely the first of many sensor-enabled pharmaceuticals. There was less consensus on whether or not this is a good thing. From general uneasiness about “biomedical Big Brother” (also from the Times) to more specific worries about future outcomes, many concerns about this new technological direction are worries about privacy and data security.
“I think that this type of technology could be really helpful for providers and caregivers who want to support a person taking regular medication,” Lucia Savage, chief privacy and regulatory officer at Omada Health and former privacy chief at ONC, told MobiHealthNews in an email. “However, it is a digital signal of a new type. So, the provider who wants to prescribe this, and the individual and their caregivers, should have a frank and honest discussion about who will have access to the signal information, how that information will be secured, and whether given all that, it is the right medication for this individual. That type of straight-forward discussion is fundamental to the physician/patient relationship.”
Topol articulated some specific worries, including that the data, despite everyone’s best efforts, could be hacked. Additionally, there’s an open question about whether the public will accept the system.
“I think that this Big Brother concept is a real issue, and people want to function autonomously and they don’t necessarily want anybody looking over their medicine when they take the medicine, if they take the medicine,” he said. “It is important for treatment of certain conditions — like for example tuberculosis, you can’t get a cure without a very tight adherence. But for many conditions, lack of full adherence might not have a big impact. So privacy and security is a big deal, but there’s this tradeoff here, which is [potentially] counterbalanced by much better results.”
Proteus and Otsuka have done the work on privacy and security, according to Wright.
“From the Otsuka standpoint, we believe the data is owned by the patients and we are custodians,” Wright said. “And we need to take a big responsibility in protecting their privacy and their security. As well as that, we want to be very explicit and make sure they understand what we’re using the data for and who we’re sharing it with.”
Wright noted that the system is HIPAA compliant and completely opt in, with the patient taking full control of who sees what data. He also said Otsuka has been working with a panel of leading bio-ethicists.
“The way the system works is it’s all by consent,” Wright said. “So the patient has to consent to want to be on the system, then they determine who they share the information with. What they will be required to share from the beginning is the ingestion data with their physician. Beyond that, they can decide if they want to share other things like rest, movement, activity, etc. That gives them the option to share with up to five healthcare professionals, and, if they choose, they can also share whatever information or no information with the caregivers.”
But Kirsten Thomson, chair of the medical device and diagnostics practice group at McDonnell, Boehnen, Hulbert, and Berghoff, is worried not just about Proteus and Abilify now, but about the future. She thinks that when this technology becomes more widespread, insurance companies might start to mandate its use in order to collect adherence data. And as of now, there’s no legal protection on what could be done with that data.
“[Insurance companies] were able, for years, decades, to get away with denying coverage to individuals with pre-existing conditions — conditions that were outside their control,” Thomson said. “Well now, this is a situation where taking a pill, especially with reminders, is completely within an individual’s control. So you would certainly, from a policy standpoint, have an argument that ‘We’re not going to pay for any hospitalizations or treatments that are directly derived from your failure to take your medications, and we have this very clear, scientific data showing that you consistently failed to take your pill’. So I think they would have a better argument than they had with pre-existing conditions, for example.”
Thomson thinks that the high accuracy of the Proteus data could lead to that scenario despite, as Wright pointed out to MobiHealthNews, insurance companies already having access to the proxy adherence data within a member’s claims data.
“The only reason I feel a mandate could eventually come about is because of the business incentives and the drivers for insurance companies,” she said. “The numbers I’m seeing are ranging between 100 and 250 billion dollars in terms of medical treatments caused by failure to take these meds in a timely fashion. So I think insurance companies are highly incentivized to get to the point where there would be some sort of mandate that would then allow them to manage those back-end costs better, wholly within the control of the patient. I think a mandate is something we should absolutely have our eye on because I can’t think of a reason why we wouldn’t get to that point. Two hundred and fifty billion dollars is a big incentive.”
Kvedar thinks Thomson’s scenario is unlikely, among other reasons because it’s hard to explicitly prove the link between nonadherence and a treatment event in a specific case. But he does think the payers will want the data, and those who don’t want to be tracked could end up paying more.
“Just like in Facebook when you trade information for something free, in healthcare we are starting to see this potential trend where you can trade something about yourself for something of value. It is not out of the realm of possibility. It wouldn’t show up, I believe, the way you described it; it would show up as ‘Your premium cost is higher because you are not adherent or not willing to share.’”
Thompson (the Proteus CEO) told MobiHealthNews that he considers the whole conversation “disappointing”.
“If you have a credit card, if you own a phone, if you use Google, all of these questions apply," he said. “Frankly, why aren’t you asking the question about Google selling payers information about how any times a week you go to McDonald’s? It’s a very interesting sort of red herring that people throw out. What we’ve created is a very powerful solution to help patients and families who are struggling to manage medication do a better job. And it’s a very secure, very private system in which we put the patient totally at the center and nothing can be done without their consent.”
A long road behind, a long road ahead
While it may be prudent to worry about the implications of a world where every pill has a sensor, it’s also good to remember that as of now, we live in a world where only one pill has a sensor. And though the second FDA approval in this category will assuredly come easier than the first, it still could be a while before we see another digital pill from Proteus, much less from a competitor.
“If you just look at the investments we’ve made in intellectual property, with 440-odd issued patents and 300 more under prosecution, and the investments we’ve made in technology to create the ability to integrate semiconductors into pharmaceuticals with high-speed manufacturing, and the investments we’ve made in human system design so that we can create an enterprise that embraces silicon pharmaceuticals, … these are very large and complex investments and I think it’s extremely unlikely that there will be enterprises who can figure all of that out in a short period of time,” Thompson said.
It also remains to be seen how the pill will perform in the market, and whether the cost and benefit will prove worthwhile to payers and providers.
“The next question is what actually happens with this product on the market and that’s an open question, I think,” Filmore said “It’s going to be a slow rollout, but the plan is to try to build a routine practice with providers, patients, helathcare, payers. How that works out is going to depend on how much of an opportunity this is as a differentiator for products.”
Thompson told MobiHealthNews last year that Proteus could be compared to Tesla, “showing the world what drugs with digital drive trains look like, just like Tesla is showing the world what a car with an electric drive train looks like.” He said the company’s innovation goes beyond the technology itself to a new way of thinking about adherence.
“One of the reasons why this is such a big problem is that the dialogue in big pharma, in big medicine, in big insurance, in big device, is always about why the consumer is the problem and why the consumer needs to change their behavior,” he said. “As opposed to saying, the problem is we’ve got incomplete products that haven’t been designed appropriately to meet the needs of our consumers and enable them to use our product effectively in their daily lives. If you approach it that way, which is what we’ve done, you end up with a very different answer.”