The Patient Equation

by Glen de Vries

  The Motivation to Comply

  Pascal Koenig worked for years in digital health before co‐founding Ava, and has found that it has been much easier to get people to comply with wearing a device when the goal is something they're personally excited about—conception—and they're not just being told to do it by their doctor. “I had much more struggle with getting congestive heart failure patients to wear devices or fill out surveys,” he explains. “But with women trying to conceive, they're very motivated to follow the protocol.”

  Even better is when the device makes it easy. Koenig admits that solely from a medical perspective it's not necessarily the case that the wrist is the absolute best place to put the wearable—but in testing, they found that it was the most successful because it was the least intrusive in the users' lives. In improving the device itself over time, the company has tried to remove pain points, making the wristband easier to close, less likely to accidentally open during the night, and easier to charge—it recharges when it syncs with the user's phone instead of requiring another step. “With fitness trackers, people complain about having to recharge on a daily basis…but with Ava,” Koenig explains, “we wanted to make it an easy daily routine; you just charge and sync.”

  It's also the case that tradeoffs have been made to balance most‐accurate data with creating a reasonable experience that users will enjoy and actually comply with. “Urine or saliva may improve the accuracy of the product…but it's even more critical for people to use it. If you talk to a hundred doctors around the world, they will all say you need to use blood results, vaginal temperature…but even if from a medical standpoint it's a little better, the challenge to get people to do it is a big one.” Instead, the goal is good enough—and at 89% accuracy, it is already far beyond what single‐variable measures can even hope to attain, so the balance, for now, has worked.

  A Man Who Just Can't Ovulate

  I actually bought an Ava bracelet before I met Koenig, because I really wanted to understand the device. I have a habit of wanting to try every latest‐and‐greatest activity‐tracking gadget, but a fertility band for women was a fascinating leap beyond my usual interests. I wore it for a few weeks, and it's absolutely the case that some of the measurements recorded were interesting to see—its measurements of my resting heart rate, perfusion, and stress level—and as relevant to a man in some applications as they are to a woman trying to conceive. Fortunately, it did not tell me I was ovulating. And Koenig speaks to the need to keep focus on the problem you are actually trying to address and figuring out your target market instead of trying to reach all people with all conditions. “We saw a clear opportunity in women's health, and we want to be the best in the world there. We're not competing with Fitbit. We want to be the leaders for women, throughout their reproductive lives.”

  It's an important point to stress—a successful device for one audience is not necessarily going to look the same as a successful device for another. It's not just about the technology. It's about tailoring the device to the problem on the table, finding a question that users want answered and then building a system that is best equipped to answer it, through whatever inputs make sense, with whatever hardware ultimately fits the issue best.

  Finding a Niche in a Crowded Field

  Ava is not the only data‐driven technology player in the fertility space. But it is the case that a number of different companies have found success aiming their products at slightly different places on the patient journey. Bloomlife is a rental device that records contractions for women in the final stages of their pregnancy…though, like Ava, they plan to build beyond their initial entry point into the marketplace.6 The company's CEO told the website MobiHealthNews, “[Our] sensor tracks a lot more than [contractions]. We can monitor fetal movement, fetal heart rate, various aspects of maternal health, all through the same sensors, essentially upgrading software and algorithms.”7 They're starting with contractions, but clearly the goal is data support for the entire pregnancy, and perhaps beyond.

  There are also players on the men's side, including YO's at‐home semen analysis by smartphone,8 and Trak, a similar at‐home testing kit with accompanying app.9 These are entry points for potential patients, empowering them to do something on their own before seeing a doctor, and giving them the tools to gather their own personal medical information.

  With conditions like infertility, the outputs enabled by thoughtful, clinically‐validated patient equations are fairly straightforward. People want information, and, in the case of well‐designed products like Ava, actionable information that can change their lives. Things get more complicated when we look at chronic diseases—where the stakes are potentially higher, and the need for help is ongoing rather than limited to a defined point in time in the user's life. In the next chapter, we'll look at two issues where technology has started to make real differences in the lives of patients: asthma and diabetes. The challenge is greater in many ways than it is for a relatively simple system like Ava's, but the potential rewards are bigger as well.

  Notes

  1. “What's the Cervical Mucus Method of FAMs?,” Planned Parenthood, 2019, https://www.plannedparenthood.org/learn/birth-control/fertility-awareness/whats-cervical-mucus-method-fams.

  2. Janet Morrissey, “Women Struggling to Get Pregnant Turn to Fertility Apps,” New York Times, August 27, 2018, https://www.nytimes.com/2018/08/27/business/women-fertility-apps-pregnancy.html.

  3. Dave Muoio, “Ava Announces New Feature for Cycle‐Tracking Bracelet, Clinical Study,” MobiHealthNews, January 31, 2018, https://www.mobihealthnews.com/content/ava-announces-new-feature-cycle-tracking-bracelet-clinical-study.

  4. Pascal Koenig, interview for The Patient Equation, interview by Glen de Vries and Jeremy Blachman, February 23, 2017.

  5. “Ava's Research—Science‐Backed Technology | Ava,” Ava, 2015, https://www.avawomen.com/how-ava-works/healthcare/research/.

  6. “Bloomlife,” https://www.facebook.com/Bloomlife, Bloomlife, 2019, https://bloomlife.com.

  7. Jonah Comstock, “Bloomlife Gets $4M for Wearable Pregnancy Tracker,” MobiHealthNews, August 15, 2016, http://mobihealthnews.com/content/bloomlife-gets-4m-wearable-pregnancy-tracker.

  8. “YO Sperm Test | A Male Fertility Test You Can Use At Home,” YoSperm Test, 2017, https://www.yospermtest.com/.

  9. “How the Trak Male Fertility Test Works,” Trak Fertility, 2018, https://trakfertility.com/how-trak-works/.

  5

  One Breath, One Drop—Asthma and Diabetes, Chronic Conditions Being Conquered with Technology

  On the face of it, asthma and diabetes don't look like they have much in common. And the deeper you dig, the more you realize—they don't. Which is why talking about both of them here offers a lesson: patient equations aren't just about solving one kind of health care problem, or about one template we can apply over and over again. The game is really different when we compare asthma to diabetes. In one (asthma) we're trying to keep people out of a dangerous situation. Our inputs are all about the combination of internal and external variables. There's how the patient is breathing—and then there's also the environment outside and what triggers may be present. In the other (diabetes) it's all a single internal process we need to try to keep in balance—the regulation of sugar level and insulin.

  What they both do have in common is the need for real‐time intervention and adjustment. This isn't like cancer, where we're looking for signifiers about a larger process, indicators that might be able to tell us sooner than traditional measures. This isn't something where we can wait to sync with our phones, check an app, study the data. This is now, in that moment—we need to leave for a better air quality environment (asthma), or we need to dose insulin right away (diabetes). There is no margin for error if patients are going to trust these devices or the algorithms behind them. There is no time for human analysis. The equations have to be right, every time. They have to work.

  Out of the Danger Zone

  Dr. Veena Misra
is a Distinguished Professor of Electrical and Computer Engineering at North Carolina State University, and the director of the university's ASSIST program—the Nanosystems Engineering Research Center for Advanced Self‐Powered Systems of Integrated Sensors and Technologies—whose work centers around building health‐monitoring wearables and the sensors that power them.

  Dr. Misra's special interest—and one of the key differentiators as ASSIST works to bring its products to market—is in developing devices with ultra‐low battery needs. The goal, she explained in an interview with me, is to make devices that can run entirely off of the exceedingly small electrical charge generated by our own bodies.1 This means never having to remove the device in order to charge it, and never having to think about it once it's been placed on or in the body. Charging one's device is a huge roadblock right now in subject adherence—as evidenced by Ava's thinking in combining charging and syncing—and Dr. Misra and her team hope to make this problem a worry of the past. “We have picked our goal, which is ultra‐low power,” she told NC State's engineering magazine in an interview in 2017, “and in that space ASSIST stands out uniquely.”2

  Among the platforms that Dr. Misra's team has developed is an asthma wearable aimed at eliminating attacks. Through a wearable wristband and patch, and an integrated spirometer into which patients breathe several times a day, the system can combine user data and environmental information—heart rate, the amount of oxygen in the blood, hydration levels, air quality, ozone, carbon monoxide, nitrogen dioxide, humidity, temperature, and other factors—and create predictions as to whether an asthma attack is on its way. If so, it can warn the subject to head elsewhere and take steps to hopefully prevent the attack from occurring.

  Preliminary test results, presented in May of 2019, showed that this kind of monitoring can detect physiological changes in lung function and is useful for preventing asthma exacerbations.3 Dr. Misra's team ultimately hopes to partner with companies that can produce the wearables for widespread use and collect data along the way to provide new insights about the disease.4 Her team is also working on applications for cardiac health, pre‐diabetes, and wound care, using a range of different low‐power sensors.

  Lowering Barriers to Zero

  Ava is trying to make charging its device an easy part of a daily routine, but Dr. Misra wants to see charging go away entirely. She explains that her goal is to see fully‐connected patients, without any logistical hurdles to get them there—“just like you put on your clothing, you put on these devices and they become your companion, warning you about your health or the environment, and enabling you to live a better life,” she says. She envisions healthy people choosing to wear devices like hers—to catch warning signs across a range of conditions, whether it's environmental exposures in the workplace, or heart arrhythmias, or glycemic index management to avoid diabetes. We can collect clues that over time can lead to better and better predictions about when we should go to the doctor and get something worked up by a physician.

  “We don't want you to have to go to the doctor for a periodic measure of what's in your sweat, what your sugar level is, and more,” Dr. Misra explains. “These technologies can allow people to keep an eye on what is going on with their health, what changes if they change their diet, what is affecting their respiration…. And as we get more data, we can see what aspects are correlated, and what is most useful in predicting the future.”

  In a TEDx Talk in Raleigh, North Carolina in 2016, titled “Powering Your Own Wellness,” Dr. Misra painted a picture of a future of self‐powered wearables—engaging us, alerting us, empowering us all, rich and poor, healthy and sick—without maintenance of any kind. We are the batteries that power our sensors, and, in addition to the four times a year the average person sees a doctor, we can be monitored for the 8,700‐plus hours a year we are currently unwatched.5

  Maybe it's too much. Maybe we don't all want to be watched all the time, especially if we're healthy (or we think we are). Maybe we don't always want to know that we're putting ourselves in danger when we get off the plane in a smog‐filled city on vacation, or if we eat that second, third, or fourth doughnut. But if we have a chronic condition, like asthma, that can become life‐threatening at a moment's notice, then we probably do. And there's perhaps no common condition that requires more constant, vigilant monitoring than diabetes—which is why it has become an area of huge excitement and activity in terms of wearables.

  A Perfectly Artificial Pancreas

  Diabetes is the ultimate biological incarnation of the smart thermostat I talked about earlier in the book, a true closed‐loop system: you measure your blood sugar, you dose your insulin, repeat as needed. Which is why “solving” diabetes with an artificial pancreas‐type solution has seemed attainable ever since wearable technology began to emerge. It has been a race to bring to market the perfect machine to make active management of one's diabetes a thing of the past. Kady Helme, a type 1 diabetic, had the opportunity back in 2014 to test within a clinical trial a device that combined a continuous glucose monitor with an insulin pump, and an algorithm that learned her body's rhythms. She gave a talk at a Forbes Healthcare Summit that summarized her experience:

  “I could do everything right, and my sugar levels are still a total roller coaster. [But] this artificial pancreas trial basically took my management right off the table…. [The] pump would read [my glucose level] every five minutes, and decide what to do…like a normal person's pancreas would.”6

  Helme talked about how she was able to eat “crazy things” that she would have previously felt guilty about: pasta, alcohol, dessert. “I really gave it a test run,” she said, “and I had to put my trust in that system.” She didn't realize until the trial was over that she'd really fallen in love with the device, and the quality of life it allowed her to have.

  This is the dream for diabetics. The reality, however, has been more complex. The FDA has approved only one commercial artificial pancreas‐style system in the United States (for children above age 7 and adults with type 1 diabetes), the Medtronic Minimed 670G, although many other companies have been developing their own versions, for years at this point.7 The problem—and it's certainly not unique to the Medtronic device—is that these machines are not perfect. One recent study found that nearly 40% of users of the Medtronic device stopped using it over the 15‐month study period.8 The device alarmed too often, had trouble adhering to the skin, misread sensor input, and often switched out of automatic mode into manual delivery, making it no better from the patient's perspective than a traditional glucose monitor.

  The problem, explains writer Clara Rodriguez Fernandez in Labiotech.eu, a digital media site covering the European biotech industry, is that “the system can only make predictions when the person is not eating or exercising, and if the sugar levels go too high or too low, it will switch back to manual mode.”9 The device isn't fully automated: it's not a closed loop. Fernandez calls it a “hybrid loop” instead.

  The challenges to a more perfect system, Fernandez goes on to explain, are that, first, current forms of insulin need to be faster‐acting when facing big changes in blood sugar; second, insulin alone is not sufficient to regulate blood sugar as well as the body can in nondiabetics; and, third, the algorithms need to be smarter, and better customized to how each patient's body reacts to the insulin dose.

  Indeed, when it comes to the second of these challenges, a recent report in Science Translational Medicine backs up the need for a multi‐hormone closed loop system instead of just relying on insulin.10 Dual‐hormone systems have been able to achieve more glucose control and keep patients out of a hypoglycemic state with more success. But the innovations needed—more stable forms of other useful hormones, faster‐acting insulin, and smarter algorithms behind the scenes—are still part of the future rather than the present.

  Hacking One's Own Device

  There has been a movement—the Open Artificial Pancreas System project (#OpenAPS)—to take the building of art
ificial pancreas systems out of the life sciences industry and put it into the hands of type 1 diabetics themselves.11 It is estimated that between one and two thousand people have attempted to build their own artificial pancreas devices out of available parts—combining a glucose monitor, an insulin pump, and a computer to communicate between the two. Predictably, this has not been an unmitigated success. In May 2019, the FDA issued a warning after learning that one diabetic had accidentally overdosed using one of these hacked systems.12

  This is a serious problem, of course, but it speaks to the desperate desire of patients to have these new data‐powered tools to manage their chronic illnesses. It doesn't take much interpretation to hear Kady Helme's account of her time with her artificial pancreas and realize how significantly something like this can change a patient's life. Diabetes is a uniquely interesting area of development because patients are already using devices to manage their condition—the new generation of machines just makes it easier for them, and optimizes a situation that they have had to manually manage for their entire lives. The real‐time actionability of the information makes the condition feel conquerable in a way that cancer and other conditions don't. There is good work being done in this area by lots of companies, and the game now is improving these devices enough to say that we have mastered diabetes.