Clayton Christensen, the famous economist who popularized ideas of innovation and disruption, showed up at the recent Connected Health conference in Boston. Although billed as a panelist, he turned up without warning as a guest in a keynote and posed the same question I asked in an article back in July: “The overarching question: whither technology?”
Like my article, Christensen distinguished between incremental improvements that don’t challenge current power structures or sources of revenue, and disruptive change that offers new solutions to old, intractable problems. Disruptive change, as summarized by Christensen, changes processes as well as products.
A similar question came up in a panel, where the moderator asked his guests whether device and app manufacturers were thinking just about the health care field as it is now, or the field as it will be in four or five years. It’s telling that his question went unanswered. I believe that most health care developers are seeking a niche in the current ecology, although hoping to be able to make the evolutionary leap when that ecology changes.
So what could provoke the kind of disruption in health care that we’ve seen in retail sales, journalism, and taxi service? And crucially, what would the results be? It’s natural to explore such questions just after the Connected Health Conference, where technologists and medical experts come together to highlight the most advanced solutions in health care reform. I’ll draw on some sessions of this conference, as well as thoughts of my own, to offer speculations about health care disruption–as well as the reasons it might not change as much as we’d hope.
Don’t be too clever
One panel at the conference included Dr. Joseph Kvedar, who has run the conference for over a decade, runs one of the leading research facilities in connected health, and wrote two books on the subject (one of which I reviewed). Kvedar compared changing the course of health care to an ocean liner, and pointed out the caution with which clinicians approached new technologies because of the danger of harming patients. Megan Beck of Open Matters suggested that some useful innovations may not currently be legal.
Universally, business planners at this conference (and several others that I have attended) advised developers to use a top-down approach to entering the health care field. They are just citing a well-established design principle: consult the doctors or other users at the start and have them design goals for your product in a structured setting. But what about the serendipitous applications that can come to light when a new technology is dropped into an environment and given to creative users? This apparently doesn’t work in health care, thanks to regulations and culture. Ernie Rapoza of Medtronic, in one panel, wished for more experimentation with technology.
What all this comes down to is that individuals cannot execute innovation in health care. An intrepid doctor or administrator may pull together funds for an experiment, but she’ll have to inspire a whole organization to follow her. And any individual who wants to try a wild idea in his own home will be violating his care plan. Innovation will happen, but more slowly than in most fields. But of course, health care is not the only one with constraints–I wouldn’t want a lot of amateur tinkering on the airplanes I fly in, for instance.
A simple solution is often put forward as a kind of instant mix to trigger both health care reform and disruption: paying for value rather than for individual services. This is certainly a factor in connected health, but pay-for-value depends on prerequisites of its own. As I have pointed out elsewhere, quality is elusive and hard to measure. And someone must still be the impartial judge who balances the need to deliver care against the risk of overtreatment. So we must continue to address the specific needs of the health care field.
The long tail is the long haul
Chris Anderson’s well-known concept of the long tail raises the possibility that businesses will move from cookie-cutter mass production to customized products that meet the unique needs of each individual: clothing that fits perfectly without a tailor, audio streams directed to the lover of avant-garde jazz. No field needs customization more than health care, because my surgical recovery has somewhat different characteristics from your surgical recovery. Even more pertinent is that my motivations for doing physical therapy and dragging myself out of the house will be different from yours.
The Connected Health conference was all about this kind of personalized medicine. Much more important than the precision medicine that aims to find our destiny in our genes, connected health ideally creates conditions that nudge ourselves, our families, and our communities to healthier behavior. (Precision medicine will eventually pay off too, I trust, especially for cancers and orphan diseases with genetic bases, but that approach will take a long time and will be less useful for the chronic noncommunicable conditions that make up the bulk of our health needs.)
During the keynote where he surprised the audience by inviting Christensen on stage, Jonathan Ballon of Intel’s Internet of Things Group pushed remote monitoring as a key part of improved health care. He suggested that it become the “standard of care,” which in legal terms means that a clinician could be punished for failing to do it.
The deep intertwining of patients, clinicians, and technology in connected health puts to shame recent efforts at “patient engagement,” a term I have criticized for some time. There are certainly good things being done under this rubric, but the term suggests some kind of failing on the part of the patients. Why don’t they love our hard-to-navigate web sites, our antiseptic waiting rooms, our complex workflows? If patient engagement causes clinicians to improve those things, it will make interactions more pleasant, but it still distracts our focus from bringing health options to people in their homes, businesses, schools, houses of worship, and community centers.
Knowledge is power
Patient engagement turns here into patient empowerment, a term much preferred by activists and health care reformers. And central to this empowerment is patient control over data.
Health care institutions don’t want to give up control over patient records. The data silos have numerous deleterious effects, as basic as errors made in the care of patients sent to rehab facilities and the risks of unnecessary repeated tests. Regarding patient empowerment, millions of people are collecting valuable data on their exercise habits and vital signs through fitness devices, but rarely can they add it to medical records. Here, though, the conference struck a surprisingly positive note. Most respondents to a survey of health care institutions said they would consider prescribing wearable devices, particularly if the data sent back from the devices was integrated into the hospital’s electronic health record. This shows a wider acceptance than is commonly understood, because anecdotally we hear that doctors are afraid of information overload or of being sued for missing some indication of illness in the fitness data.
Connected health will inevitably involve personal health records held by patients and controlled by patients. This was the theme of a keynote by Dr. Samir Damani of MintHealth. There is no excuse for putting barriers between a patient and her caregivers, simply because another party (such as a hospital) holds the key to her records.
What about the very devices that people wear next to their skin, and even under it? The manufacturers are even more brazen than the doctors in keeping all the data to themselves. Many do present dashboards to patients or clinicians, but will not offer API access for analytics. Some devices present more information or longitudinal tracking to doctors than to patients. This smacks of infantilization and contrasts with the movement toward empowering patients, reflected in campaigns by Hugo Campos and Karen Sandler.
Recent FDA guidance recommends that device manufacturers offer data to patients, suggesting that “sharing ‘patient-specific information’ with patients upon their request may assist them in being more engaged with their healthcare providers in making sound medical decisions.” But the FDA acted with timidity, failing to require data sharing. They also gave manufacturers a pass if (just by happenstance, of course) the device was designed to make data sharing difficult.
Why do clinicians hoard patient data? The explanation I’ve heard repeatedly (when we get past the fundamental incompatibilities of electronic records, which is more a symptom than a cause) is their desire to hold on to patients and keep them from moving to other health care providers. Ironically, the linchpin of health care reform, payment for value, could exacerbate clinicians’ possessiveness, if interoperability isn’t handled right.
Consider that payment for value penalizes a doctor for if the patient’s condition worsens some time past the initial treatment. Now the doctor’s hold on the patient counts for more than deriving more revenue from future visits: the doctor wants to keep treating the patient in order to ensure a better outcome. The incentive to keep patients from leaving the practice is feverishly intensified. And this, I believe, is why so many hospitals are forming Accountable Care Organizations by buying up every clinical practice they can get their hands on around their region. The strategy is not to tear down the walls but to bring more services inside.
It’s still not clear whether patient health records will be adopted on their merits, or will require a political civil rights movement on the part of patients. The technology is already available. Dr. Damani promoted blockchains as a way to enforce contracts between patients and providers–with the patients owning the data–but we could achieve the same goals with more traditional encryption and certificates as well.
And once patients feel assured of their privacy, connected health in the home can really take off. Sensors can track us to alert caregivers of dangerous indications in our behavior. Voice interfaces, as I discussed in a recent posting, can simplify reports of behavior and automated responses from the health care providers. Devices can check whether we’re falling into depression, predict falls, and more.
Doctors on call?
Beck, mentioned earlier in this article, explicitly described the upcoming health care field as the “gig economy,” and the term even turned up in the title to a keynote. This analogy strikes me as misguided. It’s one thing to order a car and have a relationship with the driver for the fifteen minutes during which he takes you to your destination. In health care, the journey is a life time, seemingly calling for a long-lasting relationship with caretakers.
The other panelists were more willing to invoke ready-at-hand reassurances around technology, saying that artificial intelligence will not eliminate jobs, but will make them more rewarding and meaningful. Although IBM Watson was not mentioned during the panel, it was clearly on everyone’s minds as they suggested that AI will increasingly take care of diagnostics and leave more time for the physician to discuss implications with the patient, define a care plan ideally tailored to the patient’s life situation, and provide encouragement.
But we’re already seeing strategies to stretch doctors’ thinner (nurse practitioners, telehealth, urgent care clinics), and these will go further. Beck and other panelists suggested that local clinics could be staffed by social workers who use smart devices that allow non-specialists to take vital signs and perform other parts of an exam. The results can be shared with clinicians over the Internet, securely, to get diagnoses. This is particularly valuable for poor and rural areas that are chronically short of trained clinicians.
An example of the kind of device being developed for telehealth is the Tyto Care kit. It’s compact, robust, and powerful, able to measure vital signs and take pictures of a throat or ear. In some ways it’s better than a face-to-face exam, because the doctor can blow the picture up on a large screen and see it in glorious high resolution.
Telehealth, however, is not inherently a disruptive technology. At one health IT event, I remember a student snidely dismissing the term “telehealth” (which at least is better than “telemedicine” because it encompasses wellness). He said the term reminded him of hoary technologies that young people avoided: telegraph, telephone, television. Extending the reach of doctors in sophisticated urban centers to far-away places is wonderful, but here again, the basic roles of the patient and physician haven’t changed. Connected health, as it’s defined by Dr. Kvedar and the participants in his conference, is very different from the telehealth vision of connected doctors and patients.
An example of combining digital monitoring with the human touch was provided at the conference by Good Measures, a company dedicated to innovation around nutrition. Like many companies showcased at the conference, they track key elements of a person’s life style, such as diet and exercise. Good Measures offers several programs (including a Diabetes Prevention Program offered in collaboration with the American Diabetes Association) that combine personalized registered dietitian coaching with an app to help people better manage their health, prevent disease, and feel better. But person-to-person contact is also critical: Good Measures offers participants their own personal clinical coach to help them learn to eat better based on their unique health conditions, food preferences, social determinants of health, etc. Another comparable service is provided by the well-established company Livongo.
One stunning advance in testing may also serve as a test of disruption. At the Connected Health conference, I talked to a company called Group K Diagnostics that can perform blood tests through a paper strip. Instead of requiring a specially trained technician to take blood through a needle, the test requires only a finger prick and returns an answer within 20 minutes. The patient can then learn the progress of her condition during her visit to the clinic, and can discuss treatment with a doctor on the spot. This will save both patient and doctor hours of time collecting results and making phone calls. (CEO Brianna Wronko cites estimates that doctors spend two hours per day on logistics around tests.)
Of course, the most disruptive impact of Group K will be independent labs, because Wronko expects their advance in testing to move 70 to 80 per cent of tests into the clinic. Furthermore, they are planning another stage of FDA approval that will allow them to offer tests directly to consumers. The tests will be marketed as panels containing 2 to 5 tests for $8 to $10, so they’re cheap enough to use regularly, such as by pregnant women checking thyroid function or hypertension patients checking cholesterol levels.
The finger prick, done daily by many with diabetes, is easy to learn. Some tests, such as one for HIV, can return actual diagnoses, speeding medical intervention. Other tests don’t return diagnoses but help the patient track her condition in consultation with a doctor. As the artificial intelligence evolves, the vision laid out by conference panelists earlier in this article may emerge, with the doctor gradually being replaced by machines for diagnosis and spending more time on treatment and support.
A maze of twisty little passages
Just as famous among computer technologists as the quote in the title of this section (which comes from a classic adventure game) is the mantra of the “minimum viable product” that allows a tech start-up to gain some success with a small amount of funding and ramp up. Dr. Calum MacRae, in his Connected Health keynote, threw cold water on the idea of finding a minimum viable product in the field of medicine.
I believe MacRae was trying to direct developers to the holistic in health care. An effective intervention in an individual’s behavior encompasses her entire life: her interests, how she spends her time, the economic and social pressures constraining her, what her family does, and more. But things get even harder: to sell your product, you have to understand the goals of the clinical setting, their workflows, what they look for and consider important, etc.
The best overview of the health care field for developers I’ve seen comes from a presentation given by medical tech expert Shahid Shah at another conference. In the first dozen slides he points out the roles played by patients, providers, providers’ institutions, insurers or other payers, and other funding institutions such as employers. He showed that the developer must consider who uses the technology, who pays for it, who influences decisions, and who benefits–those may all be different people.
Anyone hoping to develop a new treatment–a drug, a device, a behavior project–also has to determine how to bathe in the rivers of data coming our way. Will individuals who collect their own health data share it with researchers? As pointed out in a conference panel, getting patient data requires a high level of trust. And Panelists realized that achieving such trust was an uphill ride for device manufacturers. As I pointed out four years ago, many wantonly collect and even sell patient data without comprehensible notice and consent.
But the benefits of data sharing also came through at that panel. Getting patient data could streamline corporate research, with results that prevent harm as well as saving everyone money. Device manufacturers could find out quickly whether a device is broken. Pharma companies could get real-time feedback during clinical trials, and shorten them by months. Furthermore, Justin Williams of Noteworth said that giving patients quick feedback on what they’re doing makes them feel more in control of their behavior, letting them both adjust what they’re doing and feel a sense of ownership for their own health.
Two more crucial keys to unlocking innovation in health care are open data and free software. Open data has been widely embraced by governments (although open access through APIs is lagging) and has found some adoption by companies too, including pharma. But patient data is so sensitive that some form of licensing may need to be preserved.
As for free and open source software, about which I have published an overview for heatlh professionals, no field is more ripe for it than health care. Open source software is easy to justify when institutions rely on sources other than the licensing of software for their income. And that precisely characterizes the health care industry, where the product is better health and software is just an enabler. Recognizing this, a couple hundred hospitals could each contribute a trivial amount of resources to the creation and maintenance of software perfectly tailored to each institution’s needs. And yet no field is further behind in the adoption of open source software than health care. Conservatism has led to irrational decision-making.
Although important footholds have been established by open source health IT projects, many failures also plague the field. The Department of Defense has repeatedly rejected calls for it to adopt the Department of Veterans Affairs’ classic VistA software, and now the VA seems to be abandoning it for a proprietary solution too. I have explored some of the factors that lead to success or failure for open source in healthcare. It’s not an easy path. But hopefully, the sheer financial burden of proprietary software, combined with a determination by employers and payers after so many decades to control costs, will open the eyes of the providers to the benefits of free and open source software.
A final note: public health must be addressed
We’ve looked at the possible drivers and targets of disruption in this article, from doctors to patients to tech developers and researchers. But the biggest advances in health, whether from vaccines or running water, are public health endeavors. Medicine cannot succeed if it ignores climate change, pollution, food options, and stress.
The biggest determinant of health may be the workplace. Before instituting employee wellness programs that place blame on workers and try to compel behavior changes, employers should cast an unbiased eye on the emotional and ergonomic environments they set up.
Consciousness of the values of connected health can help us direct attention to the big picture that determines who lives and who dies. A few social service agencies are already looking at patient environments and challenging the status quo; this must become a part of mainstream medical practice, no matter how much tension it generates in established institutions and the larger society.