Entitlements make up the greatest share of Federal spending. Barring some improbable major source of revenue enhancement, therefore, long-term reductions in the Federal deficit will require reductions in the growth of entitlement spending. Since the second-largest entitlement program is Medicare, a better system for ensuring the fiscal health of the United States requires a better system for ensuring the personal health of its aging citizens.
Health policy analysts have done a lot of work looking for long-term reductions in health care costs, an objective frequently referred to as “bending the cost curve.” However, much of this analysis takes as fixed the existing modes of service delivery, focusing instead on how to make those modes more efficient. The way to address the particularly high cost of hospital-based health care, so the logic goes, is to reduce the cost of the health care provided in hospitals. Existing evidence from the healthcare industry and experience in other industries suggest that the most effective way to reduce costs is to provide consumers with cheaper, equally effective options, irrespective of the mode of delivery.
How can we give consumers these options? We can do it by focusing public policy on building a “distributed health service delivery system” though four distinct categories of innovations: telehealth, remote medicine, and mobile health (mHealth); medical house calls and home-based care; health agency care and peer-to-peer service delivery; and big data.
Elements of Distributed Health Service Delivery
Not too long ago, most health care services were provided in the home. The capital requirements of the medical profession were minimal, so there was little reason for an office. Since it cost little (if anything) to certify as a physician, barriers to entry and the relative wage paid to physicians were both lower than today. The advent of modern medicine changed all this. In 1930, house calls constituted 40 percent of physician encounters; by 1980 it was just 1 percent. Increased complexity and the ascendency of professional societies led to the emergence of new regulation and certification requirements affecting nearly every area of health care. Doctors increasingly chose to practice in specialized fields that relied on expensive technologies available only in hospitals. Primary care doctors were able to see double the number of patients in offices as they could by going house to house. Norms changed to the point that, until recently, few people could even imagine receiving medical treatment at home.
However, in the past decade, the advantages of hospital-based care have started to erode, for three basic reasons. First, high cost in health care. The financial gains usually enjoyed by economies of scale are running up against the increasingly large investments hospitals have to make to generate revenue (capital intensity). Second, hospital-acquired infections. The concentration of illness in hospitals has made the health care system more vulnerable to hospital-acquired infections (HAIs), with rates of incidence exceeding four out of every hundred admissions. Third, illness caused by medical treatment (iatrogenic injury and death). The very professionalism in the hospital-based model has made health care so complex that incidents of injury or death in the course of medical treatment have become commonplace.
At the same time, technologies and organizational innovations enabling healthcare provision both in the home and at a distance have improved radically in both performance and cost. Together, the four elements listed above (mHealth, house calls, agency care, and big data) can combine to create a real, but as yet unrealized, potential for distributed health service delivery on a large scale, replacing a hospital-based system that is yielding diminishing returns.
Telehealth: Services delivered via information and communications technologies constitute the first dimension of distributed health service delivery. The telephone (over a landline) created some initial opportunities for providing health care at a distance. The internet and personal computing expanded the field of remote medicine dramatically, in particular by creating possibilities for remote monitoring, diagnostics, and even treatment. In the past decade, smartphones and tablet computers have opened up further possibilities for service delivery.
In a 2013 review, Nathan Cortez categorizes mobile health applications as follows. The first kind are connectors, apps that connect smartphones and tablets to FDA-regulated devices. Procedures possible through these apps include measuring blood pressure, performing stable ultrasounds, operating insulin pumps, and visually tracking whether wounds heal. Second are replicators, apps that turn a smartphone or a tablet into a medical tool by replicating the functionality of an FDA-regulated device. Third are automators or customizers, apps that use questionnaires, algorithms, formulae, medical calculators, and other software parameters to aid clinical decision making. Fourth, informers or educators, are apps or digitized medical reference texts that primarily aim to inform and educate. Fifth, administrators, are apps that automate office functions, like identifying appropriate insurance billing codes or scheduling patient appointments. Finally, loggers and trackers allow users to log, record, and make decisions about general health and wellness.
This list suggests the range of services that can be delivered via smartphones and, to varying degrees, computers and fixed phone lines. Venture capital investments in digital health care totaled $1.97 billion 2013, more than double that of 2011 and greater by a comparable margin than venture capital investment in traditional health care sectors such as biotech, medical devices, and medical software.
Medical House Calls/Home-Based Primary Care: The once-obsolescent practice of providing medical services at patients’ homes has recently revived in parallel with the growth of “minute clinics” and other options offering consumers greater convenience and lower cost. Importantly, it’s not telehealth and mHealth driving this resurgence. Rather, the increasing cost of both hospital- and outpatient-based care, and the accumulating evidence of the superiority of home-based care (notably, in terms of patient satisfaction), have spurred this growth.
Even more such growth lies ahead. Existing programs for medical care at home focus on patients suffering from chronic illness. But Johns Hopkins University’s new Hospital at Home program has gone further, providing hospital-level care within patients’ homes.
Health Agency Care/Peer-to-Peer Health Service Delivery: Health agency care in the United States dates back at least to 1813, when wealthy women in Charleston, South Carolina, formed the “Ladies Benevolent Society” to provide care and comfort to poor, sick patients in their homes. By the start of the 20th century, demand for such volunteer services had outpaced their growth. In 1909, Lillian Wald formed the Henry Street Nurses Settlement House in New York City. Wald coined the term “public health nurse” and persuaded the Metropolitan Life Insurance Company to pay for the first visiting-nurse benefit—initially in New York, and, by 1911, nationwide—creating the first national system of insurance for homecare.
When Medicare was signed into law in 1965 it included funding for home care, providing a new and significant revenue source for home health agencies. (To emphasize: These are home health care visits by nurses, so they fall in a different category than physician house calls.) The number of home health care agencies participating in Medicare grew from 1,753 in 1967 to 10,444 in 1997. Substantial fraud and abuse accompanied this growth, however. Stricter enforcement of eligibility criteria for homecare and a number of measures included in the 1997 Balanced Budget Act reversed the growth of Medicare spending on home health care rather dramatically. By 2001, more than a third of all home health care agencies had closed. Medicare continued to account for the largest percentage of spending on home health care.
Nonetheless, for the moment the most powerful impetus for substantive change within the health care system is coming from outside the Medicare and Medicaid systems. Citizens of moderate- to high-income are increasingly dissatisfied with the menu of existing market offerings, and around the country such citizens are organizing Village Networks to pool resources to enable the provision of health care and other life services within the home in their communities. These efforts are nascent, but they signal a market demand for institutional innovations that bypass the existing health care infrastructure.
Social media provides another growing avenue of peer-to-peer health services outside of the traditional model. Aman Bhandari, formerly of the White House Office of Science and Technology Policy has noted that “information exchange, connectedness, and community” are key to an individual’s health, and social media can facilitate all three.
A fourth major trend in technology with significant implications for distributed health service delivery is the advent of Big Data. There are at least three distinct dimensions to the Big Data disruption: data gathering, data aggregation, and data analytics.
The digital devices embedded within human society are generating data at a greater-than-exponentially increasing rate. Information generated by personal computers is a tiny fraction of this total. The overwhelming share is generated from devices and appliances as varied as automobile engines, thermostats, weather balloons, and mobile phones—and now also health-related monitors and sensors of many varieties, including the FitBit and the Apple watch. Both commercial and government interests are increasingly seeing the value that exists in these pools of data, and are seeking to build and structure extremely large databases that can be “mined” for patterns. Finally, the conversion of data from analog to digital, and the search for patterns within data, can be accomplished today at speeds that are orders of magnitude beyond the best attainable even a few years ago.
This convergence of technological capabilities creates opportunities and risks. The opportunities include a potential for cheaper and better services of many types, improved search and matching capabilities, and lower entrepreneurial barriers to entry. The risks include the displacement of human workers in routinized jobs, new forms of exclusion based on algorithms, intensified concentration of wealth, and loss of control over privacy.
Of the four categories of innovations driving distributed health service delivery, Big Data is one that is as relevant to the transformation of care provided in hospitals and outpatient clinics as it is to care outside of those settings. However, the inexorable shift not only toward electronic records, but also toward the use of the varied data sources described above, will further reduce the advantage in data and diagnostic support that doctors working in hospitals and outpatient settings have long held. It represents, in short, yet another example of the shift of power away from health care middle men. It is a stellar example of disintermediation.
Benefits of Enabling Distributed Health Service Delivery
In testimony delivered to the House Committee on Small Business in 2013, Alan Portela, CEO of Airstrip, offered the following summary of the current turning point for the health care industry:
Changes and challenges faced by the healthcare system have been exacerbated as the Baby Boomer generation is reaching retirement age and 16 million formerly uninsured additional patients will be added to the system as part of Healthcare Reform. The change in scope (Patient Centered Model), coupled with the current caregiver shortages the industry faces and the move away from generalist doctors to specialists, will mean a greater reliance on mobile health and shared-medical technologies. The major driver behind this transformation is the prevention of disease and the management of chronic diseases while reducing costs. Approximately 75% of the U.S. population has at least one chronic disease, with cardiovascular diseases representing three of the top five (COPD, hypertension, cardiac heart failure, diabetes and stroke).
Different elements of distributed health service delivery have demonstrated the potential to meet these challenges by reducing costs even as it improves service delivery.
In a 2008 study focused on cost reductions that the expansion of telemedicine can achieve (in particular remote monitoring), Robert Litan found that the greatest opportunities for Medicare cost reductions fall in three categories: chronic illnesses (80 percent of Medicare expenditures); expenditure incurred during the last year of life (25 percent of Medicare expenditures); and mental health (a frequently unrecognized underlying factor correlated with high rates of readmission). Litan observed that “by improving communication between patients and health care professionals (that is, in terms of quantity of contacts as well as the quality of information exchanged), practitioners receive more information, at a greater detail, from which they can base treatment decision.” The result is fewer and shorter hospitalizations related to an array of chronic illnesses: congestive heart failure, diabetes, chronic obstruction pulmonary disease, and chronic skin ulcers. Over a 25-year period, Litan estimates a net savings of $197 billion.
This estimate is conservative, since it considers only one of the four elements described above (telemedicine). For example, recent studies have found that the provision of health services in the home leads to cost reductions of 10–30 percent, while achieving equal or higher levels of patient satisfaction and medical outcomes. These reductions are in a separate category from, and potentially add to, the $197 billion in telemedicine-enabled reductions.
The current system is most dysfunctional at the end of life, when hospitals routinely administer extremely high-cost procedures that yield minimal (or negative) benefit to patients. It is not uncommon for people to use more health care services in the final six months of their lives than they did in the entirety of their lives up to that point. As noted, fully 25 percent of Medicare expenditures are incurred in the last year of life. This percentage would not be a cause for concern if any compelling evidence existed that these expenditures benefitted patients. Rather, considerable evidence exists that tests and treatments do little to lengthen life or increase the quality of the final days.
According to one estimate, a 5 percent reduction in end-of-life costs would amount to Medicare savings on the order of $90 billion over ten years. This estimate holds despite the difficulty of determining when “end-of-life” begins, since distributed health services have the potential to create a new default set of treatment protocols, and as such would plausibly generate savings regardless of our ability to determine when a patient has entered the last year or six months of life. “Death-panel” objections to the contrary, expert judgment informed by evidence-based research certainly can play a role in achieving such cost reductions.
On a national scale, however, neither the imposition of rationing from above nor generalized restrictions to patient choice are either plausible or desirable pathways for reducing unnecessary and unproductive expenditure in end-of-life care. What does constitute a plausible pathway is expanding the range of health services patients can receive at home. Importantly, these services include palliative care—a category that is distinct from hospice (in that it does not presume the end of life is near) but that nonetheless addresses the needs of the chronically ill patients most likely to be frequently readmitted to hospitals.
Enabling Distributed Health-Service Delivery
Overall, what fraction of the services currently provided within hospitals can be offered within the home? We don’t know, but we do know that it is a far greater fraction than is reflected by current practice. To enable the growth of distributed health service delivery, policymakers first must update their conceptions of the nature and extent of impediments to increasing the fraction of health services available in the home.
The central premise here is not new. Indeed, it dates back to the 1940s, and the fundamental rationale for encouraging home care has changed little in the intervening years. Since the 1940s, Federal policy toward home health care has undergone a steady progression, from focusing on post-acute care to gradually incorporating “housekeeping services” as part of post-acute care to envisioning home health as a substitute for nursing home care and, ultimately, to extending that vision to include a range of individuals above a minimum threshold of medically demonstrated need. Concerns over cost and access—driven by the parallel growth of populations of people over 65 years of age or suffering from chronic illness or disabilities—have been constants over this period.
Now, however, things are changing rapidly, and policy isn’t keeping up. We are in the midst of discontinuities with far-reaching implications. These discontinuities, taken together, are defined by a combination of rapid technological advances intersecting with demographic trends. The core challenge for health care policy at present is not in increasing the adoption of one or another existing model of health service provision in the home, but rather in enabling the creation of entirely new business models aimed at helping people become and remain healthy outside of institutional settings. The key to meeting that challenge is a great deal of experimentation. That experimentation must be legal, feasible, and fairly compensated.
Clayton Christensen summarizes the dynamic that invites entry by competitors who can develop cheaper but equally effective ways to meet a market demand:
As companies tend to innovate faster than their customers’ needs evolve, most organizations eventually end up producing products or services that are actually too sophisticated, too expensive, and too complicated for many customers in their market.
Companies pursue these “sustaining innovations” at the higher tiers of their markets because this is what has historically helped them succeed: by charging the highest prices to their most demanding and sophisticated customers at the top of the market, companies will achieve the greatest profitability.
However, by doing so, companies unwittingly open the door to “disruptive innovations” at the bottom of the market. An innovation that is disruptive allows a whole new population of consumers at the bottom of a market access to a product or service that was historically only accessible to consumers with a lot of money or a lot of skill.
Institutions specialized in clinical interventions have well-developed systems for defining and measuring the efficacy of clinical interventions. They do not have well-developed systems for assessing the quality of the sort of integrated health service models needed to realize the full potential of distributed health service delivery. While many hospitals, physician associations, insurers, and private employers are engaged in innovations to advance distributed health service delivery, evidence from other industries suggests strongly that the disruptive innovations with greatest potential to advance health service delivery over the next thirty years will originate from outside the core of the existing health care industry. The potential today is to connect demonstrated demand from the market with technological innovation to enable innovations in the provision of health care services. And the way to do that, in turn, is to reduce the systemic advantages incumbents enjoy.
We need action in three specific policy domains to enable the necessary entrepreneurial exploration to realize its full potential, with the objective in each case being the reduction of barriers to entry that advantage incumbents in three categories: technical/regulatory (for example, regulatory approvals, standards for interoperability); financial/regulatory (for example, methodologies for determining eligibility for reimbursements); and labor market/regulatory barriers to entry (for example, certification requirements). Let’s look at each in turn.
Technical/Regulatory Barriers to Entry: In a study of the early biotechnology industry in Cambridge, Massachusetts (the first municipality in the country to enact guidelines for biotechnology research), Nichola Lowe and Maryann Feldman document how clarity in regulations can help reduce the risk faced by early innovators. They quote a biotechnology entrepreneur who noted that locating in Cambridge “allowed [our firm] to fit into a set of regulations that the university and the community had already accepted. A regulatory framework that provided a social structure for these new activities, [which] really was an important [locational] aspect.”
Lack of regulatory clarity is today a significant impediment to distributed health service delivery. Representative Mike Honda (D-CA), who in 2013 proposed legislation to create an FDA Office of Wireless Health, offers this perspective: “The tech community needs confidence in a consistent, reliable framework for wireless health. The FDA has a critical role to play. Today, there is no confidence [among] industry. It’s nonexistent.”
The fundamental problem is that technology evolves faster than regulators respond. As Representative Greg Walden (R-OR) states:
The collision of worlds in the mobile health, or mHealth, is a study in contrasts. The app economy is characterized by low barriers to entry, quick time to market, and the ability to adapt to quickly changing user needs. Medical devices, on the other hand, face a long and costly premarket approval process at the FDA. We all want to make sure that patient safety is taken care of first, but why would we treat mobile applications the same as a dialysis machine? These are the kinds of questions we need to get answers to about where that sweet spot is.
As with the methodologies for determining eligibility for Medicare reimbursements, the standard operating procedures for device approvals at the Food and Drug Administration are at least an order of magnitude slower than the rate of innovation in distributed health service provision. Institutional innovations to reduce this gap are an urgent priority for Federal action.
Policymakers and regulators in health care at all levels also must get past the notion that health care is exceptional when it comes to privacy. Other industries (notably financial services and education) also face privacy issues that must be balanced against the considerable gains to be had from data aggregation. While these issues are beyond the scope of this essay, a general takeaway for policymakers is well stated in the 2012 report of a Kauffman Foundation Task Force on Cost-Effective Healthcare Innovation: “We believe current data-sharing rules within the medical system are more overprotective than they need to be. The consensus of the task force is that health care data, suitably anonymized, should be treated as a ‘public good’—something that benefits society broadly and whose benefits cannot be restricted to just a few.”
Even when industry and government actors converge on the technical, organizational, and regulatory parameters for open data exchange, its actual implementation is likely to be a slow and laborious process. For this reason, the Kauffman Foundation Task Force proposed an alternative: empowering actors outside the medical system to collect data. Non-profits that combat diseases and employers with access to employee health data are both examples of actors that could help build a better database.
Whether through the existing health care system or outside of it, government at all levels should accelerate the process by which anonymized data is generated and shared across platforms in a minimally restricted manner.
Our Federal system, a strength in many respects, creates another set of challenges for innovators. They must contend with fifty different sets of state laws related to telehealth licensure and privacy. The harmonization of these laws is a priority for policy at the state level.
Finally, in an era of flat or declining Federal government spending on research and development, the Federal government’s own procurement spending can spur innovation. Recent controversies over access to care within the Department of Veterans Affairs (VA) healthcare system highlight endemic provider shortages that are unlikely to be resolved through modifications to existing procedures within institutional settings. The most promising pathway for improving access to care within the VA system without increasing costs is the same as that for the health care system as a whole: distributed health service delivery. By making further, coordinated investments to expand this delivery, the VA will be accomplishing two priority national objectives: providing timely, quality care to veterans and creating prototype systems that could accelerate convergence toward shared technical and organizational coordination in the private sector.
Financial/Regulatory Barriers to Entry: In principle, whether in the case of health care or in other areas, governments should only be paying for privately provided services that are effective and competitively priced. Given a fixed portfolio of clinical interventions, medical science has well-developed methodologies for assessing effectiveness—methodologies that center on the randomized controlled trial (RCT). However, such methodologies are not particularly well-suited either to assessing new integrative models of health service delivery with a social dimension (as opposed to specific clinical interventions) or to establishing the value of diagnostic tools aimed at management of chronic disease or improved general wellness. Consequently, a fundamental challenge for successfully scaling up business models utilizing distributed health service delivery is the widespread perception (encoded in policy) that randomized controlled trials (RCTs) constitute the “gold standard” for assessing the effectiveness of health service innovations in general (and not simply clinical interventions).
In the past fifty years, since research and clinical practice really became connected systematically along these lines, the science of medical intervention has advanced at a remarkable rate. The institutions supporting these improvements in medical intervention—the Food and Drug Administration, the National Institutes of Health, and the Centers for Disease Control; research hospitals across the country and around the world; and a wide array of other health care institutions at the national, state, county, and municipal levels—have all been built on these foundations, with the existing infrastructure of intervention delivery (hospitals and outpatient clinics) generally taken as given.
However, innovations in distributed health services are not conventional clinical interventions because, for starters, they don’t take place in clinics, and they mostly are not interventions. RCTs are simply too costly and too slow to keep up with the pace of innovation in business models focused on providing people with health services. The government should instead focus its efforts on partnerships with the private sector that enable entrepreneurial teams to build businesses that provide people with increased options. Eric Ries has outlined a methodology for using data gathered in real time to assess the effectiveness of entrepreneurial ventures in achieving intended objectives and to adjust to that information.
The lack of clarity of policy at the Federal and state level about reimbursement for the use of diagnostic tools is another barrier to the full realization of the benefits of distributed health service delivery. Population-based approaches to value-focused initiatives in the Affordable Care Act help create incentives for service providers to deploy low-cost, digital diagnostic tools in the communities where they work. But Federal and state governments can do more to support the many companies developing and deploying a new generation of these tools.
Overall, a strong consensus exists that improvements in health service delivery require reimbursements to be value-based rather than fee-for-service based. The acceptance of varied assessment methodologies, notably including ones appropriate to entrepreneurial ventures, accomplishes this goal in a manner that allows for innovation.
One dimension of the Accountable Care Act that is strongly unfavorable to innovation in distributed health services is the 2.3 percent tax on revenue it imposed on medical devices. In recent Congressional testimony, Dr. T. Forcht Dagi, a partner at HLM Venture Partners in Boston, described the distortionary impacts of this tax:
The tax of 2.3 percent sounds modest but it is not. This is a tax on revenue. It is not a tax on profits. The vast majority of entrepreneurial ventures developing MMAs are very small and very early. Some of the companies in which we invest may in fact generate some revenue but very unlikely to generate profit. Revenues are plowed back into the company for growth, and therefore the 2.3 percent tax on small startup companies delays their ability to reach profitability and increases the amount that must be invested before a company can become cash flow positive.
The goal of ensuring that the Affordable Care Act was budget neutral was laudable. However, imposing a tax that could damage a nascent industry of first-order significance to the overall fiscal health of the country was ill-advised. The medical device tax within the Affordable Care Act should either be repealed, or its interpretation by the IRS sufficiently narrowed so that it does not harm the development of inter-related market areas focused on distributed health service delivery.
Labor Market/Regulatory Barriers to Entry: The controversy over access to care in the VA system is simply a reflection of a broader skilled labor scarcity in medicine, particularly in primary care. Part of the solution to this problem may be a relaxation of binding constraints on primary care residency training that Medicare’s decade-and-a-half long freeze on subsidized residency slots has created. However, a more comprehensive approach to addressing medical workforce shortages would adjust reimbursement requirements to allow nurse practitioners to provide a broader range of services, with proper physician oversight and technological support. Yet even that is not an adequate solution, because there are far more homebound adults than there are qualified and willing providers for them.
In the longer term, then, pushing distributed health service delivery forward will require a new category of certified, digitally-empowered health workers, who act simultaneously as health coaches, social support case workers, and front-line diagnosticians. The creation of such a job category could attract new, tech-savvy talent to the health service field and provide a pathway toward formalizing and extending the skills of the home health workers, personal care aides, and family caregivers who are the front line of keeping patients at home. In this way the national objective of “bending the cost curve” in health care will intersect with that of increasing opportunity in the workforce—particularly among young people, who have a comparative advantage in the use of technology, and workers over sixty, who have a comparative advantage in the care of generational peers (or near-peers).
Substantial institutional innovation is occurring in the health care industry today. This innovation is taking place both within the existing healthcare industry and on its boundaries. Hospitals, physician associations, insurers, and private employers—in many cases working in conjunction with the CMS Center for Medicare & Medicaid Innovation and/or with state Medicaid officials—are actively experimenting with distributed health service delivery. In parallel, telecommunications service providers, information technology companies, local communities, and a range of entrepreneurs are innovating a wide range of new approaches.
Trajectories of innovation are unpredictable. There is no absolute basis for ruling out the possibility that modifications to the existing hospital-centered infrastructure of health service delivery could make health care cheaper without sacrificing quality. However, the preponderance of evidence in industrial organization suggests that if the basic inputs don’t decrease in cost (which, in the case of healthcare, they will not), then cost reductions can only be achieved through accelerated institutional innovation. By analogy to other industries, such institutional innovation is most likely to occur as a consequence of disruption caused by the entry of new service models and of new service providers.
Initiatives from existing institutional providers notwithstanding, the pace of change in the healthcare industry lags behind the magnitude of the opportunity to improve the quality and lower the cost of health services provision. An increased commitment to change and more effective coordination by government actors operating at multiple levels is required to realize the full potential of distributed health service delivery.
Now is the time. After all, we’re not getting any younger.