Why Are Telemedicine Systems So Expensive?

This post was originally featured on EMRandHIPAA.

Like many other enabling-technologies in healthcare, telemedicine has vast unrealized potential.

If we make location completely irrelevant and can deliver care virtually, we can address the supply and demand imbalance plaguing healthcare. The benefits to patients would be enormous: lower costs and improved access in ways that are unimaginable in the analog era.

However, one of the many roadblocks to adoption is the cost of the legacy technology powering clinical telemedicine use. In this post, I’ll outline why the telemedicine systems are so expensive, even in the era of Skype and other free video-conferencing systems.

The Telemedicine Industry Is Old…School

Telemedicine as an industry has existed for about 15 years, although uses of telemedicine certainly predate that by another 10-20 years. A decade and a half ago, the foundational technologies that enable video-conferencing simply weren’t broadly available. Specifically, early telemedicine companies had to:

1) Develop and maintain proprietary codecs
2) Design and assemble hardware (e.g. proprietary cameras) and device drivers
3) Deploy hardware at each client site and train end users on its management
4) Build an expensive outside sales force to carry these systems door-to-door to sell them
5) Endure long, grant funding-driven sales cycles

Though some of these challenges have been commoditized over the years, many of the legacy players still manage and maintain the above functions in-house. This drives up costs, which in turn must be passed onto customers. Since many customers initially paid for telemedicine systems with grant money (that telemedicine technology companies helped them write and receive), the market has historically lacked forces to drive down prices. Funny how that seems to be a recurring theme in healthcare!

But, there’s a better way

Today, many startups are building robust telemedicine platforms with dramatically lower cost overhead by taking advantage of a number of technologies and trends:

1) Technologies such as WebRTC commoditize the codec layer
2) The smartphones, tablets, and laptops already owned by hospitals (and individual providers) have high quality cameras built into them
3) Cloud providers like Amazon Web Services make it incredibly easy for young companies to build cloud-based technologies
4) Digital and inbound marketing enable smaller (and inside) sales forces to succeed at scale.
5) To reduce the cost of care, providers are increasingly seeking telemedicine systems now, without wading (and waiting) through the grant process of yesteryear.

In short, telemedicine companies today can build dramatically more cost-effective solutions because they don’t have to incur the costs that the legacy players do.

Why don’t the old players adapt?

The simple answer: switching business models is exceedingly difficult. Consider the following:

1) Laying off hardware and codec development teams is not easy, especially given how tightly integrated they are to the rest of the technology stack that has evolved over the past decade

2) Letting go of an outsides sales force to drive crafty, cost-effective inside sales is an enormous operational risk

3) Lobbying the government to provide telemedicine grants provides an effectively unlimited well to drink from

Changing business models is exceedingly difficult. Few companies can do it successfully. But telemedicine is no different than all other businesses that thought they were un-disruptable. Like all other technologies, telemedicine must adapt from legacy, desktop-centric, on-premise solutions to modern, cloud based, mobile and wearable-first solutions.

The Health Insurance Demand Problem

This post was originally featured on EMRandHIPAA.

A family friend was recently admitted to the hospital after a traumatic motorcycle accident in Colorado. He’s not in great condition, but he’s hanging in there. In light of having just written this post about the cost of highly acute care, I couldn’t stop pondering about his health insurance.

Health insurance is a bizarre creature. Unlike other forms of insurance, people actually want to consume what they’re insured against, defying the very premise of the insurance model!

Confused? Let’s dive in.

No one wants to consume traditional insurance

People never file claims for traditional forms of insurance unless something bad has happened, like car or home accidents, natural disasters, or death (covered by life insurance). In some of these cases (like minor fender benders), the insured customer often elects not to file a claim in order to avoid a premium increase. When people do file traditional insurance claims, that means something sufficiently bad has happened, and the insurance system kicks in place to recoup the damages.

People do want to consume healthcare insurance

Healthcare insurance is a wildly different animal. Only a small percentage of total hospital admissions are highly acute, catastrophic cases. A large majority of the care delivery system services non-catastrophic cases, from preventive care to counseling, scheduled (and elective) surgeries, and skin rashes, for example. Patients want as much (non-catastrophic) healthcare as reasonably possible, and they want their insurance companies to pay for it.

This is a classic principal-agency problem. The person making financial decisions isn’t bearing the cost of those decisions; in fact, the person making financial decisions is empowered to blindly spend without thinking. To make matters worse, many healthcare providers encourage patients to consume costly diagnostics and procedures with little regard for value, knowing that insurance companies will pick up the tab.

Realigning incentives

As it currently stands, this system breaks most of the basic assumptions of capitalism: the principal-agency problem, pricing information, and ability to compare producers/providers.

Reducing demand and utilization of healthcare resources is impossible. Since patients are currently incentivized to demand unlimited care without caring about cost, supply will always find a way to satisfy demand. So, how can we realign the incentives to fix the system?

The only way to reduce demand is to make patients accountable for their own healthcare expenses. With the insurance customer suddenly conscious of the cost and value of their subacute healthcare consumption, providers will be incentivized to compete and offer lower costs.

Thus, insurance companies should provide patients “catastrophe-only” plans. These plans would fully and generously cover highly acute care needs, like trauma, cancer, or stroke care. However, like a vehicle insurance plan without comprehensive coverage, the cost of treating the medical equivalent of a keyed car (e.g. a purely speculative blood test) would fall to the individual.

As CEO of a company in the healthcare space, it pains me to know that I’m contributing to the healthcare incentive problem by providing employees with a traditional healthcare plan. But until healthcare insurers offer catastrophe-only plans, patients will continue to blindly consume. In fact, even the Affordable Care Act failed in this light; the national and state-based exchanges don’t offer a single catastrophe-only insurance plan. They are all bundled and are ripe for unbundling.

You Better Stay Healthy, Or Else...

This post was originally featured on EMRandHIPAA.

As I read Jonathan Bush’s new book, Where Does It Hurt?the most salient problem that Bush discusses is that hospitals can’t effectively measure or attribute their costs. As a result, they can’t make good decisions since they don’t know how to attribute costs and revenues.

Although this has been widely known for sometime, the implications of this are particularly interesting. Since hospitals don’t know how much it costs to actually deliver care (especially multi-faceted, complicated care), their various revenue streams are effectively subsidizing their expenses in an almost random manner. Accounting for costs and attributing revenue is nearly impossible.

Bush notes that more focused care centers – such as standalone labs, imaging centers, and minute clinics – can afford to offer many of the same services as hospitals with equal or greater quality at a lower cost. They can achieve this because they have dramatically less operational overhead than hospitals and have staff performing the same core basic functions repetitively. Indeed, practice makes perfect.

There are hundreds of companies all over the country building healthcare practices based on this very premise: labs, imaging, procedures, home health agencies, ASCs, birthing centers, cath labs, urgent care, retail clinics, and more. Focused-centers are slowly eating away at hospitals by providing better services at lower costs.

Today, hospitals make enormous profits by dramatically marking up routine procedures and services. But that won’t continue forever. As the ACA pushes patients towards high-deductible plans so that patients act more cost consciously, they will seek the more affordable alternatives. Patients will not agree to pay a $300 ER copay and $2000 MRI when the urgent care center down the street offers a $99 copay and $400 MRI. As patients make better decisions, hospitals will lose some of their easiest, most profitable revenues: extremely marked up lab tests, images, procedures, etc.

What will hospitals be left to do when their easiest, most profitable revenue vanishes? They will shift focus to what they do best: performing miracles. Hospitals will compete for high-end services such as-complex surgeries and intensive care. However, because routine services subsidize the hospital’s overhead, they currently offer surgeries and intensive care at a “discount.” When hospitals can no longer subsidize their complex care with routine care, hospitals will raise prices for the highest acuity services that can’t be performed elsewhere. If you thought acute sickcare was unaffordable, think again. The cost of complex care is going to grow dramatically in the coming years.

Understanding Apple Health

This post was originally featured on EMRandHIPAA.

Apple recently announced Health and Healthkit as part of iOS 8, and initial responses have been mixed.

At one extreme, the (highly biased) CEO of Mayo Clinic called Apple Health “revolutionary.” At the other, cynical health IT pundits claim that Apple Health is a consumer novelty and won’t crack the enigmatic healthcare system. As a cynical health IT pundit myself, I’m more inclined towards the latter, but have some optimism about Apple’s first steps into healthcare.

For the uninitiated, Apple Health is a central dashboard for health related information, packaged for consumers as an iOS app. Consumers open the app and see a broad array of clinical indicators (e.g. as physical activity, blood pressure, blood glucose, sleep data). You can learn more about Health and Healthkit from Apple.

The rest of this post assumes significant understanding of modern health IT challenges such as data silos, EMPIs, HIEs, and an understanding of what Health and Healthkit can and can’t do. I’ll address what Apple Health does well, ask some questions, and then provide some commentary.

Apple Health does a few things well:

1) Apple Health acts as a central dashboard for consumers. Rather than switching between five different apps, Health provides a central view of all clinical indicators. In time, Health could help patients understand the nuances of their own data. By removing friction to seeing a variety of indicators in a single view, patients may discover correlations that they wouldn’t have observed before. With that information, consumers should be able to adjust behaviors to lead healthier lifestyles.

2) Apple Health provides a robust mechanism for health apps to share data with one another. Until now, health app developers needed to form partnerships with one another and develop custom code to share information; now they can do this in a standardized way with minimal technical or administrative overhead. This reduces app lock-in by enabling data liquidity, empowering consumers to switch to the best health app or device and carry data between apps. This is a big win for consumers.

Unanswered questions:

1) How does Apple Health actually work? Apple provided virtually no details. Does the patient need the Epic MyChart app on their phone? Is there custom code integrating iOS to Epic MyChart? Is there a Mayo Clinic app that is separate from Epic MyChart? If not, how does Apple Health know that the consumer is a Mayo patient? Or a Kaiser Permanente patient? Or a Sutter Health patient?

2) Does the patient give consent per data value, or is it all or nothing? How long does consent last? Must consent be taken at the hospital, or can the patient opt in or out any time on their phone? Who within the health system can access the consented data?

3) Given that there are hundreds of EpicCare silos and dozens of CareEverywhere silos, how does Apple Health decide which silo(s) to interface with? Does data go to an HIE or to an EMR? If to an HIE, can all eligible connected providers access the data with consent? If a patient has records in multiple HIEs and EMRs (which they likely do), how does Apple Health determine which HIE(s) to push and pull data from?

4) Does Apple Health support non-numerical data such as CCDAs? What about unstandardized data? For example,PatientIO allows providers to develop customized care plans for patients that can include almost any behavioral prescription. Examples include water intake, exercising at a certain time of the day, taper schedules, etc.

5) Can providers write back to a patient’s Health profile? Given that open.epic doesn’t allow Epic to send data out, how could Apple Health receive data from Epic?

7) How will Apple handle competing health apps installed on the same consumer’s phone? For example, if I tap “more diabetes info” in Apple Health, will it open Mayo Clinic’s app (and if so, to the right place in the Mayo Clinic app?) or the blood glucose tracking app that came with with my blood glucose meter? Or my iTriage or WebMD app?

8) Is Apple Health intended to function as a patient-centric HIE? If so, what standards does it support? CCDA? FHIR? Direct?


1) The Apple-Epic partnership is obviously built on open.epic, which Epic announced in September of 2013. It’s likely that Apple and Epic reached an agreement around that time, and asked the public for ideas on how to shape the program to get a sense of what developers wanted.

2) The only way to succeed in health IT is to force the industry to conform to one’s standards, or to support a hybrid of hybrids approach. Early indicators show Apple (predictably) trending toward the former. Unfortunately, Apple’s perennially Apple-centric approach inhibits supporting the level of interoperability necessary to power an effective consumer health strategy. Although Apple provides a great foundation for some basic functions, the long term potential based on the current offering is limited. What Apple has produced to date provides for sexy screenshots, but appears to fall short of addressing the core interoperability and connectivity issues that plague chronic disease management and coordination of care.

3) In a hypothetical world at some indeterminate point in the future, there would be a patient-facing, DNS-like lookup system for provider organizations (Direct eventually?). Patients should be able to lookup provider organizations and share their data with providers selectively. Apple Health provides a great first step towards that dream world by empowering patients to see and, to some extent, control their own data.

You Get What You Ask For

This post was originally featured on EMRandHIPAA.

I recently had a chance to meet Dr. Dave Levin, the first CMIO from Cleveland Clinic, at the Texas HIMSS conference, where I spoke about Google Glass in healthcare. During his keynote, he gave a quick overview of his book – mHealth: Global Opportunities and Challenges – that I’m reading now.

The most important thing I took away from his presentation is that people will do exactly what you tell them to do, not what you’d like them to do. More specifically, people will optimize against what they’re measured against. This is a classic business truism, but one worth repeating.

In order to receive Meaningful Use cash for adopting EMRs, providers are jumping through an excruciatingly difficult series of hoops. Among those hoops is the primary theme of MU Stage 2: patient engagement.

But patient engagement is not an end. Patient engagement is a means to an end. Although there are certainly disagreements on what the end should be (depending on one’s political alignment), the federal government is clearly pushing value-based care delivered through PCMH and ACO models.

So why are we measuring arbitrary metrics such as “5% of patients engaging with their providers” through some sort of patient engagement product? By incentivizing arbitrary usage metrics, we will see little healthcare delivery transformation, despite all the intent in the world. Instead of flipping the clinic by utilizing patient engagement tools as part of a broader healthcare delivery strategy, providers are just going to optimize to barely get by getting 5% of their patients to send them a message through their patient portal.

Consider instead these potential alternative metrics, that better reflect the spirit of the MU regulations:

1) Percentage of patient population cared for under a value-based rather than volume-based model.

2) Percentage of simple visits – script refills, ear infections, etc. – conducted remotely via telemedicine instead of in person.

3) Percentage of visits avoided simply by answering questions via asynchronous secure messaging/pictures.

4) Percentage of complex visits handled by an MD (in which the intention is to hand off simpler visits/procedures to non-physician practitioners to lower costs)

There are certainly problems with some of these proposed metrics. They don’t solve all incentive problems; the system can always be gamed. But compared with existing measures, the above metrics do much more to force providers to rethink care delivery models and flip the clinic.

Some people will interpret these metrics as a way for the federal government to institute socialist control over healthcare delivery. These fears, though, are disproportionate. While a slippery slope argument can be made in this case, the US government has only on a few occasions actually nationalized private functions. In most of those cases, the nationalization was short-lived (such as General Motors 2009).

Given the clout of the AMA and other players, the probability of sliding down this slope seems exceedingly low. History has shown that there is too much friction in the status quo in the US healthcare system for the system to change on its own. At any rate, some change is better than none!

So, Uncle Sam, hear this: you get what you measure. So please measure what you actually want.

The Nurse Will See You Now

This post was originally featured on EMRandHIPAA.

The Atlantic just wrote a piece highlighting the growing trend of non-physicians (commonly referred to as midlevels) providing healthcare. The reason is simple: supply and demand–more precisely, a fixed supply.

For any location where a patient demands healthcare services, there is only a binary result: either there is a qualified healthcare professional available to deliver care, or not. This slide (from Pristine’s investor presentation) illustrates this:


The supply and demand problem is further compounded by an archaic regulatory system. The path toward becoming a physician, at least in the US, is so arduous that the decision to pursue becoming an MD must be made by age 18 or 19. Even if a huge cohort of 18 year olds suddenly decided they wanted to be physicians, the artificially capped supply of available residency slots each year stimies traditional supply and demand economics.

Nursing, on the other hand, has a more varied cohort in terms of age of entry. Many nurses don’t enter the profession until well into their late 20s or 30s. The same is true of physician assistants. This has resulted in a more liquid supply of non-physician practitioners, and these non-physician practitioners are available to respond to the influx of new patients resulting from the ACA, and to the growing number of retiring baby boomer population.

Given the fixed supply of physicians, there are two fundamental ways to solve the supply and demand problem: make physicians more efficient, or substitute physicians with others who can do an equally good job for a given patient’s needs.

The realities of practitioner supply suggest that nurses and other non-physician practitioners will deliver an increasingly large percentage of healthcare services. Physicians will be relegated to the “high end” per Clayton Christensen’s disruption theory. That could manifest itself in a future in which midlevels deliver primary care and triage more acute conditions to “higher end” specialist physicians.

The greatest challenge in the triage-centric model led by midlevels is the (historically quite poor) communication among healthcare providers. We will need a robust technological infrastructure to support the seamless transfer of patient data among providers. Additionally, we’ll need more capable communication tools to empower providers to connect with one another and with patients regardless of location.

Telemedicine seems to be taking hold to power a future in which location is irrelevant. Interoperability is improving within health enterprises, though there are some signs that community health information exchanges (HIEs) are notdoing as well as many had hoped.

At some point down the line, we’ll likely look back and wonder why location mattered so much. It shouldn’t, and because of telemedicine, and liquid data connectivity, it won’t.

Happy Mother's Day! Telemedicine Edition

This post was originally featured on EMRandHIPAA.

In honor of Mother’s Day, I thought I’d write a post highlighting how telemedicine can benefit mothers caring for their children.

Many children can get up to 6 ear infections per a year. Everytime it happens, children complain, and inevitably mothers take their kids to the pediatrician. In most cases, the mother already knows what the problem is (because it’s frequent and easy to diagnose), but yet she has to drag her child to the doctor, leave the kid in a room full of other sick children for half an hour, only to spend 5 minutes with the pediatrician to get a prescription for antibiotics. Then she has to drive with her child to the pharmacy to pickup the medications, and then get her child back home. Meanwhile, she’s falling behind on work.

What a pain. I can’t imagine fighting that battle even once per year, let alone 6 times!

Healthcare shouldn’t be hard. And in my cases, such as the pediatric ear infection, it’s not. What if mothers bought a Cellscope, took a picture of her child’s ear, sent it to her doctor, then received antibiotics from PillPack the next day? Or better yet, what if the medications arrived via drone delivery within an hour?

I’m optimistic that 10 Mother’s Days from now, moms across the country won’t have to deal with so much frustration to solve what is such a simple problem.

The Pristine Story: Springing Ahead

Since the last episode of the Pristine Story, we've been busy as springtime bees!

Since announcing our launch at Brown University, we've gone live at three additional hospitals, with several more coming in May. We're holding back on PR for a bit, but will have some big announcements in the coming weeks, so keep an eye out for those.

We're not just busy at home, though...


Dr. Paul Porter (from Brown University) and Kyle will be presenting at the National Healthcare Innovation Summit in Boston.

They'll discuss Brown's experience using Glass for telemedicine in the ER, and present data on how Pristine's solutions dramatically improve efficiency in the care delivery system.


If you're headed to the American Telemedicine Association Annual Meeting in Baltimore, come visit us at Booth #1814. We'll be the ones in clever T-shirts and Google Glass.

In fact, we're giving away a Google Glass unit to one lucky ATA attendee, so make sure to stop by our booth and enter our raffle!

Engineering Update
Our engineers led the most recent meeting of the Austin Google Developer Group (GDG) diving into the innards of WebRTC on Android.

Keep an eye on The Looking Glass (our engineering blog) over the next few days for more on how Pristine uses WebRTC, our slides from the GDG meeting, and more.
Finally, while it may be busy in the land of Pristine, we still find time to commit our thoughts to prose. To wit, our recent blog post (penned by Kyle) that examines the difficulty in reducing the cost of care.

And if you need more material to for your week's reading. There's plenty more analysis, commentary, and forward-looking thought on our blog. Enjoy!

That's it for this installment of the Pristine Story, stay tuned for more!

Until next time,

Healthcare Entrepreneurship-as-a-Service

This post was originally featured on EMRandHIPAA.

We are witnessing a dramatic unbundling of the services that power business. Almost every aspect of business can be unbundled into a monthly service.

My startup, Pristine, runs on a number of unbundled cloud services that until recently, would have traditionally been outsourced to HR firms or mega-IT companies. We run Pristine on ZenPayrollRelateIQGoogle AppsExpensify,Maxwell HealthXeroResumatorMediaTemple and more. Similarly, we’ve built our flagship service, EyeSight, on top of a broad array of development tools and services (check out the Pristine Engineering Blog to learn more about how the sausage is made). We’ve made it a priority to invent and do as little as possible by utilizing 3rd party stacks and services everywhere possible.

Healthcare is not immune to this trend. There are a number of companies that are unbundling health IT entrepreneurship:

This list should be 10x longer than it is. With all of the capital and startups entering the health IT space, companies providing the infrastructure to accelerate growth will thrive. As the old saying goes, “During a gold rush, sell pickaxes.” The companies listed above are selling pickaxes to proverbial gold miners.

I’ll conclude this post with some areas that can be commoditized as a service. Feel free to leave comments on other areas or companies that I missed.

  • Interoperability as a service (in lieu of HL7)
  • HIPAA compliant videostreaming as a service
  • HIPAA compliant image / video sending as a service
  • Analytics as a service

The Feds Are Supporting Telemedicine

This post was originally featured on EMRandHIPAA.

The Federation of State Medical Boards (FSMB) recently passed a model telehealth policy that promotes virtual visits for first-time encounter. This is notable for 2 reasons: first, many state medical boards liberally borrow from the federal boards, and second, this marks a shift from the old model in which patients were encouraged to see providers in person before engaging in telemedicine consults.

It’s encouraging to see the old, arbitrarily restrictive model fade, in favor of one where patients can begin building a relationship with their physician without travel. Indeed, people meet on the internet all the time; why can’t patients meet their care providers the same way?

The old model was arbitrarily limiting access to care, and thus driving up costs and driving down quality. Under the new model, patients should finally be able to login to a web service and be connected directly to a qualified physician that payers will cover. For telemedicine companies like American WellDoctor on Demand, and others, this is a major coup.

This combination of technology and new guidelines will reduce ER visits, improve access, and ultimately reduce costs. Once it’s easy to get access to preventative medicine, patients will actually partake in preventative care. As a simple example to illustrate this, let’s examine my wellness check up habits.

I’m a healthy young male. I haven’t been to the doctor for a check up in close to a decade and have no intention of going. The process of booking an appointment, leaving my job that I love, and sitting in a waiting room are enough to deter me from ever going to the doctor. But if I could step into a private space and consult with a physician via a video consult for 15 minutes, I might actually get an annual check up. If the physician discovered something concerning and asked me to come, I would actually come in. But I would never come in for an in person visit without an explicit reason to. It’s not worth the pain and headache of going into the doctor’s office unless I have a reason to; the only way to achieve preventive medicine at scale is to make it easy for patients and providers alike.

Ambulances, ERs, and urgent care centers should expect a similar change in their operations. In these environments, specialists can now be reimbursed for first time consults with patients across a range of devices – iPhones, iPads, Androids, Macs, PCs, and even Google Glass. Neurologists can beam into ambulances for strokes, cardiologists for cardiac resuscitations, and trauma specialists for trauma cases. The opportunities are really endless, and my company, Pristine, is proud to lead the way in these new hyper-mobile telemedicine environments.

On the other hand, the new guidelines set forth by the FSMB aren’t all positive. Perhaps most perplexing, the FSMB did  not classify messaging and audio-only phone calls as telemedicine. They didn’t strictly forbid either activity, but they made it clear to payers and providers that live, synchronous video is necessary for reimbursement. In light of the shift to ACOs and value based models, this is perplexing. It’s been suggested that Kaiser Permanente and Group Health physicians reportedly spend up to 2 hours per day interacting with patients through asynchronous messaging.

Despite some setbacks in the new standards set forth by the FSMB, I’m incredibly excited about the future of telehealth across the continuum of care. The new model put forth by the FSMB is just the first of many steps toward a healthcare delivery system in which telemedicine powers the majority of care delivery across the country.