top of page


A new class of healthcare technology, Connected Health, enables data exchange among different user types via a variety of devices. With Connected Health, patients, doctors, other caregivers, insurers, life sciences companies, and employers can collaborate using the web, mobile, as well as dedicated hardware. Deploying healthcare technology to a single type of end-user via a single technology is hard. Deploying multi-user, multi-device Connected Health solutions is much harder. For Connected Health solutions to be effective, patients, providers, and/or other stakeholders must use them. In other words, people must change how they work. Change is never easy. Change in the complex, regulated, and fragmented healthcare industry is daunting. Two groups who appear unfazed by these challenges are entrepreneurs and investors. Digital Health venture funding surpassed $4B for the past three years.[1] Startup Health suggests it was $8B in 2016 alone.[2] Many of these Digital Health investments are Connected Health solutions. The resulting innovations are fascinating. They include AI-based virtual health assistants, a doctor accessible via webcam, wearable wellness bracelets, implanted heart failure monitoring devices, mobile-based patient/provider messaging, semi-closed-loop insulin pumps, various motion tracking tools, and even Bluetooth-enabled tampons. These innovations are already making it to market. As many as 88% of surveyed consumers report using at least one digital health tool, and one in ten use more than five tools.[3] With all this exuberance, it is hard for eHealth business leaders and investors to prioritize their focus. There are many excellent Connected Health segmentation frameworks with associated growth estimates available.[4] I suggest each can be improved by adding one more lens. Are the benefits of a solution worth the pain of deploying it? Failing to make such a calculation can result in some poor Connected Health bets.

Sources of pain in deploying Connected Health

There are at least three main sources of pain in deploying Connected Health: getting patients to use it, getting clinicians to use it, and paying for it. Each of these has hard dollar costs, e.g., hardware and software, consumer marketing, IT integration, and provider training. Just as important are the “soft costs.” These are the expenses that are harder to measure. They include provider morale, patient satisfaction, regulatory risk, and quality-of-care risk. Any decision to deploy Connected Health solutions should include a full accounting of these sources of pain.

Pain #1: Just because you built it does not mean they will come—Consumer adoption. Connected Health solutions that involve patients require patients to regularly use the technology. This is obvious, but it is not easy to accomplish. Consumers need to be made aware of the solution, understand its benefits, deploy it in their real-world settings, integrate it into their daily lives, see value in using it, and potentially even pay for it. Every step along this journey is costly and complex.

Pain #2: Just because it may help them or their patients does not mean they will use it—Provider Adoption. Many Connected Health solutions require some form of provider engagement as well. Clinicians/staff may need to use the platform itself, or at least modify workflow or data systems to support a deployment. Confounding adoption are already busy schedules, rigid IT systems, potential to do harm to patients, complex legal liability and compliance frameworks, tight margins, and lack of technical expertise. Failing to manage any one of these elements can kill a Connected Health deployment regardless of how ultimately useful it may be to all.

Pain #3: Someone ultimately must pay for it—Financing Connected Health. Any change in healthcare requires money. Connected Health programs must finance themselves through some combination of increased revenues or decreased costs, or be sponsored by a third party (e.g., a drug company) who derives some collateral benefit from their use.

Different types of Connected Health require overcoming different pain levels

Connected Health can create at least four sources of value for healthcare industry incumbents. In order of least to most complex, they can: serve as marketing/promotion vehicle, generate new data and insights, improve existing care processes, and create entirely new types of care. Many solutions hold the promise of more than one of these. Some may even create all four.

The easy stuff: Using Connected Health as a vehicle for marketing and promotion. An early incarnation of Connected Health was sponsoring the deployment of relatively simple tools and services to capture patient or physician mind share. They included basic provider reference content (e.g., Skyscape®), consumer self-management tools (e.g., iTriage®), limited care coordination programs (e.g., Text4Baby®), and healthcare shopping apps (e.g., GoodRx®). These tools required little to no training or technology integration, and little change in workflow. Their value to users was immediately obvious. Further, these businesses had a clear economic beneficiary—typically a drug company, insurer, or other third party. Assessing these tools' impact is relatively easy. Sponsoring organizations run estimates on how many users adopt a solution, how using it would change behavior, what the economic value of that behavior change is, and what the cost to promote it is. While few of these solutions profoundly impact the healthcare system, they are useful, and if appropriately capitalized, can be rewarding to investors.

The hard stuff: Using Connected health to generate new data & insights. Healthcare databases as early as 2013 contained over 150 exabytes of data.[5] This is growing at high double-digit rates. However, there is still not enough information. Most importantly, healthcare organizations have little insight into their patients’ lives when they are outside the provider’s four walls. Similarly, understanding clinician mindsets and practice patterns is not easy. Organizations have little actionable data on what providers are doing, how they are thinking, and what they need and want. Connected Health can bridge some of these gaps. Examples of insight-generators include patient self-tracking tools (e.g., PatientsLikeMe®), wearables (e.g., Fitbit®), smart inhalers (e.g., Propeller Health®), real-world evidence gathering platforms (e.g., Health Core®), 3D camera technology for assessing neuro-muscular health (e.g., Atlas5D®), device/drug companion solutions (e.g., Voluntis®), and automated market research (e.g, Sermo®). These tools all generate valuable data for enterprises. This data creates an economic incentive to fund, or at least subsidize, the deployment costs. Such incentives are critical because the cost to capture and then glean insights from data are significant. For example, it is easy to get a provider to use a free online drug reference app. It is much harder to have them spend 10 minutes answering specific questions on why therapy A is preferred over therapy B, even if the question was posed by a colleague. Similarly, getting a consumer to use a 2-minute anonymous symptom checker app is much less of an ask than having them track their weight overtime and share this data over the cloud.

The harder stuff: Using Connected Health to improve existing care processes. Like other industries, social, mobile, analytics, and cloud technologies (SMAC) have already profoundly impacted healthcare. Connected Health is a major component of this, allowing clinicians and patients to better accomplish existing workflows. Messaging platforms (e.g., PingMD®) streamline how doctors, nurses, and patients communicate in between visits. Routine administrative transactions are automated (e.g., pharmacy prior auth via CoverMyMeds®, patient check-in via Phreesia®, or admission notification via Cureatr®). Routine clinical interactions are simplified (e.g., getting an oral contraception prescription via Nurx®, or taking a pre-op assessment via ePreop®). Blood glucometers, blood pressure cuffs, weight scales, etc., have all been made “smart.” These give patients instant feedback on their data while aggregating it for clinicians. Skype®-like technology allows the patient’s living room to serve as the psychologist’s office. Risk stratification analytics, show rate analytics, and even payment analytics, all push new insights to providers to inform care and business decisions.

All these Connected Health solutions require patients and providers to markedly change their workflows and daily activities. Thus, these technologies need to create large benefits to all affected users to justify their cost, risk, and complexity. Quantifying this return-on-investment involves comparing performance pre- and post-deployment, and computing the total cost of roll out. While not a perfect methodology, this at least gives a business leader an estimate of the net benefit.[6]

The really hard stuff: Using technology to provide clinical interventions previously not possible. Finally, Connected Health allows for entirely new workflows and types of care. The “smart devices” mentioned above can be combined with provider interfaces and protocols to allow for continuous patient monitoring (e.g., Health Loop®). This non-encounter based care was not practical fifteen years ago. eVisits allows consumers to initiate a virtual problem-focused clinical encounter either for a common illness (e.g., American Well®), or even with a specialist for a rare disease (e.g., Grand Rounds®). While technically not new, Connected Health has even resurrected the “house call.” Uber-for-doctors if you will. Connected Health devices combined with protocols and analytics can even eliminate entire care encounters. For example, the BabyScripts® pregnancy management platform safely reduces unneeded office visits for healthy expectant mothers. Finally, Connected Health, and its ability to benefit from tremendous amounts of real-time data, allows for the “bot-based” care encounter. Artificial Intelligence solutions like Lark® create nutritionist’s visits without a nutritionist, while Sensely’s® AI-based health assistant serves as a virtual nurse. Integrating any one of these technologies into the clinical workflow is a massive effort. Given the potential to dramatically improve outcomes and healthcare economics, many will certainly be well worth the trouble. Unfortunately, assessing in advance any one of these model's likelihood of success is very hard. It requires envisioning a highly uncertain future where patient demand for the care is initially unknown, as are the resources required to deliver that care.

Some guideposts for finding high-impact Connected Health opportunities

Healthcare’s many inefficiencies will create several Connected Health successes. However, not every Connected Health effort will prevail. The winners will be the ones whose pain of deployment is small relative to the benefits they create. Solutions that are easy to bring to-market and create a lot of economic and clinical value are likely to succeed. Conversely, solutions that require changing lots of well-entrenched behaviors will have a significant risk of failure. If the results of this hard work are only mild benefits, the odds of success plummet. While the pain of deployment versus the potential benefits must be weighed for each opportunity individually, below are some guideposts that suggest a higher likelihood of success:

1. Solutions that target already “wired” patients. Older people represent the largest pool of chronically ill patients. They, however, are the least comfortable with technology.[7] Further, there is a negative correlation between illness and disposable income, reducing their ability to pay for new healthcare services.[8] Thus, solutions that focus on younger patients' illnesses, e.g., type 1 diabetes, asthma, CF, multiple sclerosis, etc. are likely to have an easier path to market than those serving later-in-life illnesses, e.g., CHF, COPD, type 2 diabetes, and arthritis.

2. Solutions that address already well-known consumer or provider pain points. In all industries, healthcare included, it takes time and money to have people seek solutions to problems they did not know they had. For example, many online psychology companies are attempting to convince currently untreated people that they would benefit from behavioral healthcare. This is expensive. Also, while it can create demand, it does not necessarily create demand for the company that paid for the marketing. A similar dynamic is seen in medication reminders and smart pill bottles. While pharmacy non-adherence is a major concern of the provider and payer community, many people do not lose sleep over missing a dose or two of their medications. (Sleep aids, and maybe birth control pills, being possible exceptions to this statement.) For comparison, there are many obvious unmet needs. Many patients know it is a pain in the neck to go to a doctor for an antibiotic prescription for a UTI. Others know they want to schedule a dentist appointment for next Wednesday, but do not have time to play phone tag with a scheduling clerk.

3. Solutions that target a sufficiently large number of patients and providers. All sales is a numbers game. For solutions to achieve a critical market mass there needs to be enough target users. The size of the ideal market is proportional to the complexity of deployment. For clinician adoption, things that target rarer diseases need to function with little workflow modification. Conversely, some workflow modification may be possible for Connected Health technologies touching large swaths of patients. For example, it is reasonable to ask pulmonologists to tell their patients with idiopathic pulmonary fibrosis (IPF) about a new IPF-focused app. It is another thing to ask that practice to put in a remote home monitoring protocol for a disease that only affects a few patients in the panel.

4. Solutions that target concentrated patient and providers populations. In addition to scale, market concentration is critical. For example, solutions targeting specialist-managed conditions, e.g., MS, cancer, ESRD, and pregnancies, are more easily deployed than solutions for diseases which are co-managed with primary care, e.g, hypertension, type 2 diabetes, and CAD. For specialist-managed diseases, solution makers know who to target, and the specialist knows she is responsible. Similarly, patients that self-concentrate either physically or virtually, e.g., pregnant women, MS patients, or patients on hemodialysis, are easier to reach than patients who do not, e.g., patients with depression or obesity.

5. Solutions that do not require inventing the underlying medicine. Implementing Connected Health solutions is inherently innovative. However, effective solutions are built on established, accepted, evidence-based treatments. Consider the many Connected Health successes around pre-diabetes management, e.g., Omada Health® and Canary Health®. Neither of these companies developed the underlying clinical intervention. They are based on the NIH’s Diabetes Prevention Program. These Connected Health companies digitized, scaled, and improved upon the tested approach.

6. Solutions that address already measured, easily compressible costs/wastes. There are many ways to reduce healthcare costs. The first is lowering the costs of the inputs to care (e.g., supplies, non-generic drug use, ensuring clinicians practice at the top their license, streamlining administrative transactions). A second is eliminating known sources of waste (e.g., re-admissions, unneeded testing, and unnecessary use of high-cost care settings.) Finally, one can make patients healthier under the premise that healthier patients spend less. This third category is critical to health system improvement. However, from a Connected Health perspective, all else being equal, the more direct the cost savings the better. "Deploy 'solution A' and 'cost B' falls" is a straightforward sell. "Deploy 'solution C' to patients with 'disease D,' who after a spike in medication spending will on net use less care," is a more complex story to prove. If a Connected Health solution's primary cost reduction mechanism is health improvement then it is better to focus on areas already top of mind among purchasers. For example, tools that reduce CHF re-admissions, a hot topic for many reasons, will get a lot of attention. Those solutions attempting to slow CKD progression to ESRD will get less, as renal health is less of a focus than cardiovascular disease.

7. Solutions that generate actionable, already sought-out data and insights. A common goal of Connected Health solutions is generating new data on consumers and providers. However, not all data is created equal, even if the data is accurate. The data must be useful to whomever is paying for it. Tactically, Connected Health solutions should be able to draw a clear line between the data they are collecting and its economic benefit to the data’s customer. For example, it might be interesting to know what reference information a nursing staff is reading daily, or if patients with high-blood pressure are less socially connected than their healthy peers, but turning this insight into dollars is tricky.

8. Solutions that succeed in current as well as future reimbursement models. I contend most of Connected Health works better in a pay-for-value world. Automating clinical tasks, moving significant amounts of care outside the exam room, generating predictive insights on patients, standardizing delivery, coordinating a multi-disciplinary care team, etc., are not well compensated in fee-for-service. Unfortunately, healthcare payment transformation will be measured in years while Connected Health innovation in plays out in quarters. Thus, there needs to be a market for any one solution today. Ideally, there should be a compelling enough reason to adopt a Connected Health solution in a purely fee-for-service world, e.g., operating cost elimination, new revenue, etc.

9. Solutions that capitalize on macro health economic trends. Healthcare is heavily influenced by macroeconomic forces. While the market is so large there is room for all types of innovation, those that are aligned with the new spending pools are advantaged. For example, in the US there was tremendous investment in employer-health innovation in the 1990s and 2000s, as large employers increased their spending on disease management. On the delivery side, during those years the highly profitable surgical services became the innovators. Obamacare then drove innovation from employer health and surgery to provider-based risk-bearing technologies. Future innovation, I suspect, will be fed by the needs of the senior population, and to a lesser extent, Medicaid-oriented solutions.

10. Solutions that minimize technical integration. Moving data is getting much easier. But it is still not easy. UI/UX integration, user-management integration, and workflow modification are still incredibly challenging. As such, Connected Health solution developers must limit the required integration as much as possible. On the patient side the same is true. Technological change is challenging for many Americans, even as Wi-Fi and smartphone use grows. There are numerous studies that address the technology divide between people of different ages, incomes, and ethnicities.[9] I can speak to this challenge personally, as I often struggle to get my highly educated parents up and running on supposedly “out-of-box ready" technologies. On net, you can never make it too easy.

Few if any solutions will ever be able to check-off all ten of the guideposts above, nor should they. However, Connected Health solutions that do not meet any of these criteria will be have an uphill battle. Regardless, I continue to support the innovators, both the solution developers and early customers, as they endure the pain of innovation. It is a noble effort, and rewards will likely follow.


1. Tecco H. 2016 Year End Funding Report: A reality check for digital health. Rock Health.

2. 2016: The Health Moonshot Movement. Startup Health.

3. Meeker. KP Internet Trends 2017. Kleiner Perkins.

4. For example, see “Top 5 Digital Health Categories Poised for Growth in 2015” at

5. Ibid.

6. Technically, any project evaluation financial decision should employ a discounted cash flow analysis, where the risk-adjusted cost-of-capital accurately reflects the opportunity cost of the investment. In practice, this is very challenging for connected health investments.

7. Smith A. Older Adults and Technology Use. Apr. 3, 2014. Pew Research.

8. Shaw KM et. al. Chronic Disease Disparities by County Economic Status and Metropolitan Classification, Behavioral Risk Factor Surveillance System, 2013. US CDC.

9. Rainie L. Digital Divides—Feeding America. Feb. 9, 2017. Pew Research.

bottom of page