The Patient Equation

by Glen de Vries

  Thinking about all of these alternatives simply as tools for health care leads to an important way of thinking. Whether or not a life sciences company is successful can be defined by one simple idea: are they providing better tools to health care providers than the tools they currently have? Full stop. If a pharmaceutical company makes a drug that is more effective or safer (and ideally it's both) than what's on the market today, they will be successful. The drug will provide more value to patients than the current standard of care, providers will want to prescribe it, payers will see value in reimbursing for it, and the pharma company's revenue will cover the cost of R&D, allow the company to invest in new therapies, and return value to shareholders.

  Everyone wins when better tools are created. If the same theoretical drug isn't a better tool for treating patients, then no one will want to take it out of the toolbox. And room will be made for the next life sciences company to try to improve on the current toolset.

  Once we abstract everything to this idea of tools—molecules, medical devices, wearables, and apps—the competitive landscape becomes clear. Pharma companies may see digital therapeutics potentially taking away market share from them. But if you look at companies like Proteus, partnering with Otsuka to deliver drugs on digital and device platforms, or at the use of drug‐eluting stents (medical devices used to hold arteries open while releasing drugs into nearby tissue—which began to be deployed almost 20 years ago), some of the most ingenious and valuable developments have come not from one category of therapeutic, but from the combination of multiple modalities.

  The pharma and biotech industries have every incentive to find ways to track patients who are actually taking their drugs, make sure they are taking them optimally, and leverage patient engagement and behavior change to create the greatest possible biological effect. If you're a pharma executive, you should want to know that your patients are actually taking the medication they've been prescribed. You should want to have patients tuned to the right dosage. You should want them to exercise, eat, and act in ways that increase the efficacy and safety of your therapeutic. If you aren't doing that, it doesn't mean that your drug isn't spectacularly valuable on its own. But you do run the risk that a competitor will create a better tool—in combination with digital technologies—than your drug alone.

  You should also want to use these technologies to get patients who are not responding off the drugs as quickly as possible. Whether related to reimbursement models (which we'll talk about more in Chapter 14) or simply to show that your drugs are working for patients and providing them with the best possible treatment, you should want to measure whether therapies are working and, at the same time, if they are, to properly motivate and incent patients to use them. There are huge costs to drugs not being effective—certainly in a value‐based care scenario but also simply in marketing and word‐of‐mouth effects. It is not good for anyone to have drugs out in the world that are not helping patients.

  As one example, one of the biggest problems for pharmaceutical companies is that patients don't finish the full course of their medication—and outcomes suffer. If apps can help address this problem, that is potentially huge. There is useful information to be discovered here. Are patients stopping the medication because of side effects? Because of an inconvenient dosing schedule? Because symptoms have resolved and they just don't finish the course? The answers to these questions can shape future development, without a doubt.

  Digital from the Beginning

  For the life sciences companies who do leverage digital to supplement the measurable biological outcomes their products deliver, I believe the next biggest mistake is waiting until the end of the traditional drug development process to think about them.

  Consider the relatively simple problem (though without a simple solution) of medication adherence. A Proteus‐style digestible to measure whether a pill has been swallowed could certainly be part of a solution, but it's not the only option. Other engagement apps are possible, or wearables that measure effects, something that reminds the patient to take the pill, or alerts a provider that it's not being taken. Imagine, as in Figure 12.1, that the pharmaceutical company—forward‐thinking in today's market, at least—starts to think about a digital engagement and compliance measurement platform around the time that their drug is approved by the FDA.

  There is some inherent value in the drug itself, marked by the horizontal line A. The development process can obviously increase that value, through the kinds of precision targeting we've already discussed—but for this example, we'll assume the value stays constant. Let's also assume that the company is successful in being able to engage patients, increase their adherence, and change their behaviors in meaningful ways for the disease being treated. (We will attack that assumption shortly, though!)

  As the company develops and delivers this digital companion to the drug, it increases the therapeutic value (see line B). The better that engagement strategy gets—and we know that more and more data can improve engagement, as we saw with examples like OneDrop—there is additional patient value generated.

  Figure 12.1 Pre‐ versus post‐regulatory approval engagement strategies

  But, if the company began to develop that engagement strategy during the drug development process rather than waiting until release (line C), we can see an even greater amount of value generated over time (the total area between A and C). Patients benefit more as individuals, and more of them benefit.

  Moreover, the total patient value—literally the sum of all positive biological effect for the drug and the engagement platform—is greater at the time of drug approval than it would be without having plugged in that engagement strategy in advance. This may or may not be relevant to a regulatory agency. Perhaps there are no competing products for the indication. But if there are, anything that can be proven to increase safety, efficacy, or patient value is relevant. That additional value might make the difference.

  Plus, the pricing of the drug is going to be coincident—or at least will track with, potentially up to a year later in some parts of Europe—the regulatory approval. The more demonstrable benefit that is shown, the easier it will be to justify a higher price.

  Of course, we can challenge the underlying assumption that this theoretical company will be successful in creating a worthwhile digital strategy. Apropos to our discussion of digital snake oil, not every app, and not every engagement strategy—just like not every drug—will successfully create a positive effect. What better time for testing could there be, if a company wants to have a disciplined, scientific approach to ensuring that their engagement strategy will be successful, than when they are already working in the disciplined, scientific framework of drug development?

  Patients will be instrumented far more during development than during the commercial life of the drug, and there is the opportunity to control variables—to tune the dial on engagement the same way we might tune the dial to find the right dosage of the molecule.

  Digital strategies are at their most valuable when they are not just add‐ons, but integral pieces of the development process—not just a go‐to‐market strategy. Virtually every pharmaceutical company should be thinking about a digital companion to the medicines they are developing. And if they aren't thinking about it now, it may be too late.

  The development of the digital companion to the drug could even make its way to more customized drug labels. Just as unexpected biochemical or genetic biomarkers may have an effect on efficacy or safety, there may be cognitive or behavioral responses—measured explicitly, or the byproducts of digital strategy research and development—that can be triggered by some aspect of the labeling.

  But It's Not That Easy

  Laying out this idea—“digital from the beginning”—makes it seem like it's a no‐brainer, and that every drug company should have figured this out already. But there's a reason they haven't. Building an effective digital product is hard. For every Twitter and Facebook—undeniably sticky technologica
l applications—a thousand social media companies died along the way, unable to engage people enough to make a viable business. It is naïve to think that creating these apps is easy. In the end, it's not so different from evaluating a hundred compounds to try to find the one drug that works.

  Stan Kachnowski, professor at Columbia Business School and chair of HITLAB—the Healthcare Innovation and Technology Lab—sees data in action every day, and what it takes to get new technologies into the marketplace. In talking to him, he characterizes one of the biggest challenges as diffusion—can a particular platform make its way into the population, become a trusted part of a patient's everyday life, and actually be used to effect change?12

  On the one hand, Kachnowski sees a lot of tools rapidly expanding into health care and changing what we do as life sciences companies, as doctors, and as patients—but on the other hand, he also sees a lot of mistakes, and a lot of assumptions about diffusion that are simply incorrect. Diffusion—more than sales, more than profitability, more than press coverage, more than even results—is what matters, Kachnowski insists. The diffusion of apps into the patient population, or of digital systems by clinical trial sponsors, investigators, and subjects, has been considerably slower and more fraught than may have been expected a decade ago.

  “It is difficult to predict diffusion,” Kachnowski says. “We try to do it at HITLAB by doing very early testing, but when something is on the cusp, figuring out which way it's going to tip is hard.” He has seen things like online consultations and physicians being paid to write emails come into the world and not get nearly the traction he predicted. “I thought for sure that would be widely adopted…and it's not. I get it wrong a lot of the time.”

  Kachnowski has seen failures like a Harvard attempt to integrate payers and providers, which required the support of too many large stakeholders to truly take off. He hasn't seen an app for health that has picked up more than 10% diffusion into the population, not a single health care app with more than 3–4% engaged users over the course of a year, millions of dollars being put into these apps with very little to show for it.

  Pharma has such powerful incentives to get people to use apps—the kinds of engagement, tracking, and subsetting that we've already discussed—and the money to develop and distribute anything they want to put their energy behind. But Kachnowski doesn't see it happening. Instead, he sees regulation getting in the way—pharma companies legally can't see much of the data, they can't collect identifiable patient data, they can't really collect anything right now that they can use to help their adherence models. I'm less skeptical than he is—I think there are ways to work around this challenge, to stay within the regulations and still learn plenty from the data. But he is undeniably right that the amount of effort that the biggest pharmaceutical companies have put around initiatives like these is limited.

  In Kachnowski's mind, it's payers who will have to lead the digital revolution. Yet they are also bewildered by how little adoption they have seen with the apps out there. He has been involved in the development of CardioNet, a company providing ambulatory ECG monitoring to detect and treat arrhythmias, and he believes it has saved a very significant number of lives—but patients were never going to pay out of pocket for it, and it had to go through the payers. It has disrupted the industry—hospitals have lost tons of money from patients who used to be admitted for a week to monitor them in‐house and now they just get sent home with a CardioNet device—but it took the payers insisting that this was going to be the new standard of care in order to gain any traction in the market.

  Kachnowski sees the same thing happening with sleep. Home‐based sleep kits are hard to sell directly to patients, but a substantial fraction of payers now require a home sleep test before they will reimburse for a stay at an in‐person sleep center. So the number of patients being directed to sleep centers has fallen precipitously, and it's almost becoming a business of the past. Kachnowski says that Columbia's sleep wing used to be 5,000 square feet, is now 500 square feet, and is inching closer to no longer being a viable investment of resources.

  On the pharma end, he sees the money being spent on the wrong technologies. Not that many people are going to sign into standalone apps and do anything with them consistently. The value has to be obvious—and especially with many patients not wanting to face the fact that they have a particular disease, and not wanting to engage with their health any more than they have to, it is a huge challenge. He calculates that $15–$20 billion has been lost in the development of beyond‐the‐pill apps over the past decade, a ton of energy and zero return because patients simply won't engage consistently enough to produce useful data. And that even in the world of wearables, too many consumers don't want to wear them, they don't like the way they look, they don't understand them, they don't fit, and there's a social stigma. “What 85‐year‐old woman with thin wrists, and maybe not a great understanding of the value, is going to wear an activity tracker?” he asks.

  There are two saving graces Kachnowski points to. One is that, more and more, we have devices to collect data passively, which don't require people to log in, enter data, or do anything that might stand in the way of diffusion. The other is that when apps or wearables are deployed in clinical trials, compliance is virtually 100%—and it becomes a game changer. “When the physician goes to the subject and says ‘you have to wear this, it is vital for the study,’ they do it,” Kachnowski says. “They will wear it 24/7 because their physician is the one person on the planet they trust.”

  But Kachnowski has confidence that more companies will figure it out and that more will solve the diffusion problem and help carry digital health to the next level. The economic incentives for the successful development of digital engagement strategies and companion digital therapeutics are undeniable. It is a huge area of opportunity for life sciences companies, big enough for them to want to continue to try to tackle the problem, even after so many failures.

  Thinking about that future—the competitive landscape of pharma companies not transforming themselves into digital therapeutics companies, but incorporating digital into the fabric of the molecular products they development—I am optimistic. There will be winners and there will be losers, but progress will be made.

  Where That Leaves Us

  If there's one lesson I want readers to take from this chapter, it's that the life sciences industry going forward isn't and can't just be about drugs and physical devices. It is also about digital diagnostics, companion apps, and platforms that we use to manage diseases that include drugs—not merely drugs that have simple around‐the‐pill apps that don't create biologically‐based and measurable outcomes. This is something everyone in life sciences needs to be thinking about—even though these digital solutions are expensive to develop and successes have been hard to come by as yet. As MedCityNews writes, “diagnostics have been a tough sell. While they're necessary to fulfill precision medicine's promise, no one really wants to pay for them.”13

  Yet, diagnostics are what will elevate us from the cusp of the data revolution into a world of fully‐developed patient equations. They are going to tell us who is going to best benefit from the hugely‐valuable drugs being developed. Ultimately, apps and diagnostics need to end up being a part of the standard workflow for treating patients. Instead of a patient presenting with a particular indication and being prescribed a drug, we need built‐in intelligence to tell us whether that particular patient should get drug x or drug y, and then to track their response so that if it's not working, we can switch them as quickly as possible to something that will. An around‐the‐pill platform that incorporates this from the start—that subsets patients from the very beginning so that only the right ones even start down this course of therapy—is going to find itself hugely valuable.

  We need to think about the right platforms for managing dosing—for antibiotics, for infusion therapies, for everything. We can predict the curves for a particular patient as far as blood percentage concentr
ation over time, and just like the OneDrop app, alert patients for their next dose.

  And, finally, we need to make sure sensible regulations are in place. Not just to address Stan Kachnowski's point that pharma needs to have permission to collect the data they need in order to sensibly expect they will make the investments needed to wade into this area in a serious way, but also to make sure that patients are protected from apps that don't actually do what they claim to do, or what they intend. We have regulations for things we ingest or implant, but we don't think about them in the same way when we're talking about a digital device of some kind, or particularly just a smartphone app. But if a cardiac monitor is connected to an app that is producing recommendations about heart medication, or there is a system telling you how much of a drug to take and when, these need to be accurate, or they're dangerous. We need disciplined research to connect the inputs to the outputs. We can't have an app telling us that if we eat one more cheeseburger, we're going to have a heart attack—unless we've somehow validated that the prediction is likely to be correct.

  The endgame is pretty clear to me, even though it's obviously at least a generation away. Someday these systems—apps, wearable sensors, ingestibles, and implantables—are all going to work together to create an amazing synergy of health, lifestyle, and behavior. We'll be prescribing games for patients to play in order to determine whether they need a particular drug, when they need it, and how much they need. Or the patient's performance on the game will change the drug that comes with it, and then a sensor in the body will send data into the cloud that will change the game to better help us the next time it's played. We'll be printing pills at home that will give us the right dosage of everything we need for that day, based on our stool sample, our activity level, all the measurements being passively collected from our bodies. Or maybe we won't actually print the pill—it'll all be blended into a cookie, delivered via Amazon to our door, as part of our dinner. And when we open the package, it will have a sensor inside measuring our grip strength, which will serve as one more input into tomorrow's cookie.