Not long ago, many services such as tax accounting were delivered episodically and in-person, as most health care still is today. Periodically, a client and accountant would meet, review financial materials and status and, at the end of the encounter, make an appointment for the next meeting. Increasingly, in-person accountant visits have been replaced by phone or web meetings and do-it-yourself software like TurboTax. There is still a need for accountants and face-to-face meetings, but typically accountants now require such visits for only the more complicated cases that can’t be managed with software or a call.
Health care has proved resistant to a similar transition, although everyone would benefit. While some aspects of care clearly require doctor and patient to be in the same place at the same time, many demonstrably don’t. Nonetheless, even those parts of care that could be freed from the doctor’s office remain tied to it, with schedules optimized for doctors’ productivity rather than what’s best for the patient.
In this article, we identify the barriers slowing the transition of episodic, in-person primary care to innovative models that separate care from location and that empower patients to take on more of their own care. And we describe steps needed to overcome these obstacles, lowering costs and improving quality.
With the rapid expansion of information connectivity, health care requires much less hands-on work than it used to. Every now and then an examination or a procedure requires physical contact between provider and patient, but many of the important elements of health care have always occurred through oral communication alone and some others — like getting your blood pressure measured or your blood tested — can now just as easily and accurately happen at home or elsewhere outside of a doctor’s office. Much of what used to be restricted to clinical settings can now, from a purely technical perspective, be handled remotely.
But let’s imagine that you are ready to video or phone chat with your doctor, recognizing the convenience of never having to leave your house or wait in a waiting room. Unfortunately, your doctor may not be that eager to chat with you. Three forces keep telemedicine from achieving its potential to transform care delivery: the excess workload created for the care provider, the likely change in patients’ health-care-use behavior, and the economics of reimbursement.
Provider workload. Much of the current thinking about new service-delivery models is based on the notion that cost and efficiency of care delivery can be enhanced by directing more patients to primary care. Enthusiasm for this approach largely derives from the appealing and perhaps nostalgic notion that more comprehensive and coordinated primary care could provide better outcomes at lower costs. Indeed, many efforts at health care transformation are built on tying patients more closely to primary care providers. The primary care community has largely embraced these ideas, perhaps because they reinforce the community’s value and centrality to the health care enterprise. The ideas have also been embraced by public policy makers, health plans, and the general public.
Primary care physicians are used to being responsible for all aspects of patients’ care. They are also among the lowest-cost physicians available. For perhaps these same reasons, there aren’t a lot of them. As a result, they are already overworked. In contrast to cardiologists or orthopedists who can limit their responsibility to heart disease or joint- and bone-related issues and triage all other issues back to primary care providers, those providers have no safe harbor. The lower reimbursement and need to provide 24/7 coverage that makes a primary-care-based system so appealing to policy makers and others has long been shown to be a source of burnout and challenges in getting medical students to choose careers in primary care. The last thing that this group of providers desires is more responsibility. Just imagine the operational demand if primary care providers also had to actively monitor their patients’ health data between office visits.
From this perspective, the enthusiasm of policy makers and health plans to simply increase access to primary care as a way to address delivery challenges is a cop-out. Rather than do the conceptually hard and convention-challenging reorganizations of systems and finances required to truly transform care, it is easier to suggest tweaking the system and encouraging patients to use more primary care — and let primary care take it from there. That cannot be the solution if, in fact, primary care providers simply can’t take it from there — either because there aren’t enough of them or because they don’t have the support to do one more thing beyond what they are already doing.
Overconsumption. The workload problem is further increased by patients’ likely response to enhanced availability, such as what tax accounting and retail customers have enjoyed through digital channels.
The old care model creates friction between the practice and the patient: the difficulties and delays in scheduling visits, the hassle of traveling to the practice and waiting in the waiting room, and the penalty of copays and coinsurance make patients think twice before seeking in-person care.
In our recent studies of how patients responded to the introduction of a portal allowing them to e-mail health concerns to their care team, we found that the e-mail system that was expected to substitute for face-to-face visits actually increased them. Once patients began using the portal, many started sharing health updates and personal news with their care teams. These new channels did not keep the patients out of the office; rather, they encouraged patients to visit the office more frequently. What doctor would dismiss a patient’s e-mail that mentioned chest pain without calling the patient for an in-person visit? Everyone has ups and downs in how they feel. As those downs become easier to report, physicians are alerted to more symptoms. The disease burden is unchanged — but now the aches and pains are given greater voice.
Without the barriers to in-person care, every skin blemish that can be photographed and sent electronically risks turning from something that patients used to ignore into a demand for medical attention. That’s good if what gets brought to attention is serious and would otherwise have been overlooked. But there is an adage in clinical medicine: “The great secret, known to internists…but still hidden from the general public is that most things get better by themselves. Most things, in fact, are better in the morning.”
Third-party reimbursement. A central challenge in designing new delivery models that rely less on face-to-face encounters is that so much of the payment system that funds existing health-care-delivery models is built around such encounters — often requiring in-person contact as a condition of payment.
Health insurers have long believed that making face-to-face visits a requirement for reimbursement could protect them from overuse. If you make health care too easy to get — and it would certainly be easier to get if it were just a phone call or an app away — then you’d have to develop tighter controls to ensure that the care that was delivered was actually necessary. Otherwise, the fear goes, every minor ailment shifts from something that would likely resolve on its own into an encounter that creates a reimbursable insurance claim. Historically, health insurers did not have tools necessary to identify and act on the difference between care that is worth it and care that isn’t. Instead, they relied on past inefficiencies that made all health care harder to get — for example, by requiring in-person visits even when the goals of care might be more easily achieved by a phone call. This approach has had the secondary consequence of limiting innovation in delivery methods.
Requiring such contact not only reduces access and adds transactional friction, it also creates some perverse incentives that actually discourage using the most efficient routes to improved health. For example, we recently reviewed the payments to be received by ophthalmologists for performing eye exams for patients with diabetes. These are recommended annually to prevent diabetes-related blindness. Patients don’t particularly like these examinations because they require a separate visit and having their pupils dilated. The dilation itself takes some time but worse, it leads to a half day of bright lights and blurry vision. Adherence rates are low — no doubt contributing to blindness that could have been prevented if more people got screened.
Now, new cameras with special lenses can provide a nearly equivalent view of the retina without dilating the pupils. Patients don’t have to visit the ophthalmologist. They just put their chin on the device at a convenient location and in a few minutes high-quality images can be transmitted to the very same ophthalmologists for review. With the new “visit,” there are few waits and no blurry vision. It’s no wonder that screening rates are substantially higher using the new method compared to the old. Should the payments for these two “visits” be the same or different? And if they are different, which one should be higher?
In our review, a commercial insurer would typically reimburse about $254 for an in-office examination involving retinal photographs — about $26 of that for the photographs themselves, plus some facility and professional service fees. The same insurer would reimburse a total of only $16 (for the photographs) if the service were performed remotely, with no payment allowed for the interpretation of the images. Thus, the current third-party-payment model actually discourages physicians from providing convenient, high-quality, low-cost care. This is a paradoxical result of the payment scheme, since the remote process increases screening rates, preserves vision, and insurance companies are rated on their ability to serve this need. It doesn’t serve any stakeholder well to maintain that reimbursement differential between in-office and remote examinations.
Our view is that early conceptions of connected health and the patient-centered medical home (PCMH) traveled a flawed path. These models were largely premised on the idea that gains could be achieved by increasing the connections between patient and provider — through frequent engagement between patient and clinical staff in a PCMH or similar initiatives. However, this approach piled more responsibility on an already-taxed primary care system and was inefficient because it didn’t reflect the do-it-yourself trends observed in retail, travel, and finance. A better model might have aimed to decrease connections between patient and provider, substituting other forms of support to assure equivalent or enhanced outcomes. We propose that health care systems need to reimagine their relationship with their patients, which includes thoughtfully triaging the increasing amounts of data they produce and engaging them in their own monitoring and care. This means using technology to handle the simpler needs, escalating to the doctor only what can’t be delegated.
Getting the right data to the right recipient. The cycle of patient care involves feedback loops: A change in blood sugar, weight, blood pressure, or some other sign is detected, and that information triggers an action. Consider the power of mobile devices in such loops. A showcase for this has been the Parkinson App developed by researchers at the University of Rochester. Using the sensors built into smart phones, patient movements can be tracked 24/7. Slowed movement or gait asymmetries can be relayed to the care team, providing indications of care needs far more easily, and perhaps more systematically, than those based on semi-annual physical evaluations.
Technologic innovations such as these are exciting, but is the care team really the right recipient for all of the information they provide? And just because the information is collected at all hours, does it really have to be analyzed in (close to) real time? Imagine the work of a primary care physician receiving data feeds from all of his or her 2,300 patients.
Mightn’t we delegate some of the screening work to patients themselves? Empowering customers with easy-to-use tools transformed the tax reporting and travel industries. While we don’t expect patients to select what blood-pressure medications to be on, we probably can offload considerable amounts of the monitoring and perhaps even some of the treatment adjustment to them. Diabetes has long been managed this way, using forms of self-care that have advanced as self-monitoring technology has improved.
Some of the common thinking about new monitoring technologies assumes that the doctors need the same level of involvement with the new data streams as they’ve had with patient information all along — as if the goal of remote monitoring is to direct more and more information to the clinician rather than use it to empower others to take on activities that previously only clinicians could do. Under the old system, the patient’s gait was assessed in the office, and the doctor decided whether treatment adjustments were indicated based on that. Now that gait can be assessed elsewhere, we should consider whether the information still has to go right to the doctor. Probably only some of it does.
Detecting a change in gait rarely needs to trigger a 911 call. It is sufficient if that information is integrated into a report that is periodically made available to the patient and the care team. Patient monitoring and provider work hence can happen asynchronously, dramatically simplifying the work-flow.
Similarly, when an otherwise-healthy person shows a three-pound weight gain from one day to the next, it might reflect some combination of measurement differences and extra eating. However, if a patient with heart failure shows the same signs, the clinical response should be different. Twenty-four/seven surveillance of a chronically sick patient who already requires high vigilance from the care team makes a lot more sense than doing the same for a patient in good health.
Which incoming data streams can be batched into a six-month report and which data warrant an emergency visit with the provider shouldn’t be decided according to how technically easy it is to port the information to the doctor, but rather by clinical needs. The biggest opportunities come from applying clinical, not technical, sensibilities to these challenges: Future primary care providers need protocols and rules that automate ongoing triage and interpretation of incoming data. They also need to define workflows in which baseline care is first handled by the patient, perhaps with automated support; simple exceptions are then handled by technicians and physician extenders such as PAs and nurses; and more complex exceptions are escalated to the doctor. That doesn’t mean that all doctors are working at the top of their license at all times, but it does mean that the default recipient for information is not the doctor and the default for the timing is not right now.
Engaging patients in their own care. One reason patients’ involvement in their own care has not been made more central to health care is that simply making medical information easy to understand and act on is often insufficient. For example, we recently completed a study among patients with poorly controlled diabetes in our clinics and found that even those with diabetes for many years could not correctly interpret the meaning of hemoglobin A1c levels, the most commonly used measure of glycemic control. Diabetes guidelines define good control as having an A1c around 7 or less. But our patients who thought they had “good control” had average A1cs of 9.8, and those who thought they had “poor control” had an average of 10.2. The two readings are close to each other, and neither one is particularly good. Similarly, national calorie-labeling efforts have been largely unsuccessful, perhaps because the reasons consumers fail to control calorie intake are complex and have only partly to do with misunderstanding calorie information.
The assumption has been that an essential early step towards empowering patients is to provide feedback in ways that are easy to understand and that give guidance on what to do with the information. Such approaches grow from models of rational patient behavior in which patients, once informed about how to improve their health, naturally do what’s needed. But the numbers of people injured in car crashes because they weren’t wearing seat belts, who take up smoking even though they know tobacco kills, or who fail to take once-a-day medications that can prevent a second heart attack reveal the limitations of this kind of reasoning. Even if it were easy to provide patients with clear feedback and education, it might not get us far.
Rather than trying to simply educate patients to encourage health-promoting decisions, better to also change the context or architecture of their lives — making good choices more automatic such as with well-chosen defaults. Better to offer carrots with an option to switch to fries, for example, than the other way around. In that way, the decision processes that often lead people toward worse health instead lead them toward better.
It might also seem sensible to simply make it easier or more fun for patients to take better care of themselves. But while an app to manage one’s care may initially seem like fun, the novelty (and effectiveness) wears off quickly. In one pilot in our clinics, we gave patients with poorly controlled diabetes and high blood pressure free wireless devices to measure their blood pressure and blood sugars and asked them to use them each day. Within three months, patients used them only 50% of the time; within six months only 30% of the time.
One proven way of keep patients engaged is by using financial incentives. For many chronic illnesses with effective treatments (such as hypertension, hyperlipidemia, asthma, diabetes, coronary artery disease, heart failure, and chronic lung disease), the cost savings derived from improved medication adherence can be substantial — enough to justify giving patients large incentives. These are most effective in inducing desirable behaviors when patients participate in lotteries. For example, we found that a daily lottery incentive worth only $1.40 increased the daily use rate of glycemic monitoring devices from 58% to 81% over the first three months and was associated with better glycemic control.
Tapping into patients’ social networks can also be an effective engagement tool. Many patients have at least one friend or family member who would be willing or even eager to help them improve their health and support them in their goals. There is lots of evidence that people are influenced by what those around them do, and successful models like Alcoholics Anonymous and a variety of different types of peer mentoring programs suggest that even strangers may be willing to help others improve their health.
There is a large advice industry about time management, designed for busy executives seeking control over the relentless demands they face. Much of the advice aims to help them reallocate fragments of otherwise wasted time toward work — time between meetings, at the airport, in the shower. It’s a Sisyphean task. The real solution comes from learning how to work less, not more, and yet still achieve objectives.
That reframing may be a cliché, but it is directly relevant to delivering care. It’s backwards to think that the solution to the primary care problem is for patients spend more time with primary care providers. As we’ve argued, the solution is for them to spend less time with their providers. New, value-based-payment arrangements that shift financial risk to providers are enabling such innovation in care delivery. Providers have incentive to drive efficiency-enhancing improvements when they are financially accountable for the total cost of their patients’ care. Such payment arrangements can also reduce insurer’s incentives to create obstacles to care such as requiring face-to-face visits. Indeed, the new payment models should incline insurers to support the sorts of innovations we describe here as these reduce patients’ demands on primary care providers.
Health care will never be like tax accounting, but that does not mean it cannot learn from it. New solutions will almost certainly require some form of demand management: replacing things that take a long time with things that take less, and offloading demand for care first to patients and their support networks, then to mid-level professional support, and last to primary care physicians. A vision for the future casts the visit to the primary care doctor not as the solution but as a kind of failure—an inability to accommodate patient needs by any of the less-expensive levels of support.
In this vision, primary care is the new tertiary care. We see this as the kind of oxymoron required for rhetorical emphasis to move us beyond tinkering around the edges of health care delivery toward more sustainable approaches for population health.
This article is based in part on a paper originally published in Annals of Internal Medicine.
David A. Asch, M.D.
David A. Asch, M.D., is the John Morgan Professor at the Perelman School and Wharton School of the University of Pennsylvania where he is also Executive Director of the Center for Health Care Innovation.
Christian Terwiesch, is the Andrew Heller Professor at the Wharton School and the Perelman School of the University of Pennsylvania where he is also Co-Director of the Mack Institute for Innovation Management.
Kevin G. Volpp, M.D.
Kevin G. Volpp, M.D., is the Janet and John Haas Professor at the Perelman School and Wharton School of the University of Pennsylvania where he is also Director of the Center for Health Incentives and Behavioral Economics. Follow him on Twitter @kevin_volpp
This article originally appeared on HBR.org and is being brought to you by Medtronic.