Amy Belt, Class 12
Healthcare spending is out of control and we, as venture investors, may be part of the problem. We can also be part of the solution. I originally began looking into healthcare spending figures in an attempt to quantify the impact the changes in the healthcare insurance market, both public and private, were having on healthcare venture investments. Our investments were suffering from an increase in the data, time, and process required to have new technology not only used but also paid for—often tacking on years (and millions) to the already seven-year time horizon. The choices were denial (which one might argue we’d already been practicing for several years), shooting in the dark (we’ve been accused of doing a fair amount of that as well), or total avoidance (the world could always use another on-line gaming company). Ultimately, what we’re paid to do is assess and quantify risk so that we can value it appropriately.
What I found, through both research and hands-on work, were models of success in getting new technology paid for, as well as a realistic sense of how much it costs to get there. Along the way, I also came to realize that these new reimbursement hurdles are being driven by the massive underlying cost problem in our healthcare system. Despite the various models of insurance and taxation batted around by Congress, no amount of cost shifting will stem the tide of healthcare costs. These costs are swallowing our economy, and attempts to curb them could be our next big wave of healthcare investment. The key to our current healthcare crisis is bending the cost curve. To achieve that end, I argue that rather than stifling innovation, we need to redirect it.
In this article, I discuss the origins of my interest in reimbursement and the effect reimbursement challenges are having on venture capital investment, provide an estimate of the true cost of time and capital for reimbursement, and consider models for successfully obtaining new reimbursement (as well as exploring why reimbursement has become a fundamental challenge for medical device start-ups). This analysis is intended to help others accurately assess the value and risk of reimbursement in medical device start-ups, as well as to implement successful models for obtaining reimbursement in their current portfolio companies. I conclude with a new investment thesis and a call to action for incentives in the reimbursement system that would enlist entrepreneurs and venture capitals to invest in solutions that solve—rather than exacerbate—our healthcare cost crisis.
My Introduction to Reimbursement
I apologize to my reimbursement expert friends, but no other topic in healthcare makes people’s eyes roll back in their head like reimbursement. It has a language so replete with acronyms that it sounds frankly like gibberish, no matter how hard the experts try to simplify it. In fact, when I first started in venture, reimbursement wasn’t even included in initial pitches; if it were, a one-line “we’re using existing codes” was cited and we all moved on, grateful not to have to go into it in any more detail.
However, as a person new to venture capital, I realized that I needed to know something my more seasoned and connected colleagues did not. What I needed was a hook, something meaningful, and something that made me valuable both to the assessment of risk and to the success of the company. Reimbursement was, for reasons stated above, a topic my colleagues didn’t know much about; it also was critical to success because our companies could not sell what no one would pay for.
Given that reimbursement was fundamental, why did so few people understand it? Typically, if something is that important, people will become experts at it no matter how complex it is. The reason we weren’t all experts was that reimbursement was not always a hurdle to success—once upon a time, reimbursement was almost guaranteed with FDA approval.
Origins of the Reimbursement Problem
The first time I saw the change in reimbursement paradigm was in March of 2005. I was still in my position at Guidant, a cardiovascular medical device company that developed the first FDA approved carotid stent, a mesh tube used to open blocked arteries that feed the brain. The FDA had approved Guidant’s carotid stent the previous August, and traditionally CMS (the Center for Medicare and Medicaid Services) would pay for any technology the FDA approved as safe and effective. Not this time. In March 2005, CMS announced that their coverage decision for carotid stent would include patients representing only 1/3 of the population for whom the FDA had approved the technology. Essentially, for the first time in memory, CMS, not the FDA, was determining whether someone could get treated with a technology.
Unfortunately for venture investors, this became a trend with both public and private insurers. Reimbursement was no longer automatic after receiving FDA approval. The process, timeline and capital required to achieve a positive coverage decision continued to grow.
Effect on Venture Capital
This issue was compounded in venture capital because typically we had exited our companies either through M&A or IPO before we needed to drive meaningful revenues, so the area of reimbursement was neither familiar nor budgeted into our returns analysis. A consolidation of acquirers and the closing of the IPO window meant we were suddenly holding companies who were facing a challenge we had not seen before.
Initially, there was a head-in-the-sand reaction among venture investors. We told ourselves it was just a few unlucky companies. Once it sunk in, the next reaction was panic. “Don’t take any more reimbursement risk” was the mantra. Given that more than 80% of healthcare spending is covered through public or private insurance,1 panicky avoidance was also not a rational or sustainable response—there are only so many face-lifts and tummy-tuck procedures to invest in. In addition, firms were already invested in companies and had knowingly or unknowingly taken on reimbursement risk three to five years earlier.
Having raised my hand a year and a half prior to become our in-house reimbursement “expert,” I was put on the case of helping to figure out our portfolio reimbursement situation. I took on the role of Interim Vice President of Reimbursement and got an up-close-and-personal view of what it takes and what it costs to get a new product paid for in the current environment.
Our first clue that something significant had changed on the venture side was that we started to see companies with a product on the market coming around for another round of fundraising. They were allowed to sell but unless someone is paying, they were burning capital. They were also asking for funding of post-market clinical trials specifically designed to address the needs of the insurance companies who were still deciding whether to pay for the device. Worse yet, the process was political and uncertain so it was not known how much time and capital and clinical data were going to be required for the product to be consistently paid.
Ingredients for Successful Reimbursement
If you ask a reimbursement expert what it takes to get reimbursement, you will undoubtedly get the answer, “It depends.” Although there is truth in that answer, as a venture investor, it frustrated me beyond words. Through research and analysis, I arrived at the following rules of thumb. At a high level, there is a three-legged stool of utilization, clinical evidence, and physician society support.
This is one of the more expensive aspects of obtaining reimbursement. Some minimums to be considered for reimbursement are:
- Drive ~3,000+ U.S. case volume (aggregate) to establish that it’s a standard and accepted practice
- Achieve broad-based usage by physicians (i.e. not just academic medical centers)
There is great variation in standards from one specialty to another about how large these trials need to be. The guidance is:
- 5 published, peer-reviewed articles
- U.S.-based studies and journals to get U.S. reimbursement
- At least one randomized, controlled trial
- Ideally, demonstrated non-inferiority to a comparably priced and currently reimbursed product/procedure or superiority to a less expensive product/procedure
Physician Society Support
This is the trickiest aspect of obtaining reimbursement, particularly if you need to obtain a new physician code since it can be the case that where one society (or physician within a society) gains, another one can lose. This zero-sum game is actually set up in the system. Rules of thumb are:
- Influential society leaders supportive of the procedure/ product
- At least one society champion who is willing to speak out in favor of the procedure/product
- Ideally, incorporation in the clinical practice guidelines of a society
- No enemies from within the society or in another society
The Reimbursement Process and Costs
The reimbursement process typically takes at least four years to achieve. Driving utilization, gathering additional clinical evidence beyond what was required for FDA approval, and garnering physician support from key opinion leaders within multiple societies is a multi-year, grass roots effort (see Figure 1). A significant part of this process is getting insurance companies introduced to and educated about a new product or procedure.
Figure 1. Reimbursement: A Three-Stage Process.
Stage 1: Case by Case Reimbursement
This is typically the starting point for a new product or procedure. Before a product has its own code for which coverage has been identified, physicians and hospitals must ask for reimbursement on a case-by-case basis, requiring additional time and documentation from the healthcare provider. This includes appealing denied claims, sometimes reaching a third appeal that is reviewed outside the insurance company.
Stage 2: Plan by Plan Reimbursement
The case-by-case process continues until an insurance plan has seen a sufficient number of claims to believe that it is worth their time to review on a broader scale. Plans will review coverage policies on a local level (often state level) first and typically don’t make national coverage decisions for several years. In these circumstances, a physician who is submitting a large number of claims should request a meeting with the medical director to review the need for the procedure and the clinical data. The more influential the physician is with that particular insurance plan, the more successful those appeals for a policy will be. Smaller, more regional insurance plans are more likely to adopt positive coverage policies first. They can write a coverage policy, even in the absence of a permanent code. Larger, national plans typically differentiate on provider network and cost and therefore can justify being later adopters of technology. This process also takes several years to develop local champions and have hundreds of meetings with insurance companies.
Stage 3: Coverage
After at least three years of driving utilization with the aid of case-by-case and then plan-by-plan reimbursement while simultaneously putting together the clinical data package and society support, it’s time to apply for a new code to get systematic coverage of the procedure. A code doesn’t guarantee coverage, but it goes a long way toward convincing insurance companies who have been denying claims or haven’t established a coverage policy as well as for larger, national plans to establish positive coverage decisions. For physicians, it also initiates a process by which Medicare values and pays for the procedure.
Unfortunately, as may be evident, the process for obtaining reimbursement is both capital-intensive and uncertain. Does this mean another round of financing? The short answer is: yes. Compounding the cost is the uncertainty that comes from politics within the American Medical Association and the art and politics of insurance company reviews of the technology. For all the tough feedback the FDA is getting, FDA approval is a far more straightforward process than obtaining reimbursement.
To further understand the range of cost, I break it down in Figure 2 into direct reimbursement support, clinical evidence collection, and depression of the revenue ramp. Direct reimbursement support is the people hired, either in-house or as consultants, to support short-term and long-term reimbursement efforts, with a rule of thumb range of approximately $3-5M according to my calculations. The clinical evidence cost is based on a trial cost of $10-15K per patient, and ranged from $3.0-7.5M depending on the number of patients required within a particular specialty or disease. Finally, the most expensive part of the process results from the inefficiency of the sales force during the time when the company is fighting for reimbursement. Without reimbursement, the same sales effort can produce as little as a third of the revenue compared to a company whose procedure is well reimbursed. This additional burn can add $7.2-12.6M over 18 months. In total, I calculated a range of $13-25M of additional investment required to get a company to profitability.
Figure 2. Capital Requirements for New Medical Device Reimbursement.
Successful Reimbursement Models
The good news is that obtaining reimbursement for a new product is achievable. There are models of success and a system to work through and clear reimbursement hurdles. In the following real-life examples, both St. Francis Medical and Kyphon were able to navigate through to reimbursement and were rewarded with $725M and $4B acquisitions, respectively. In fact, achieving reimbursement can be viewed as a value creator, not simply a hurdle for which a company gets assigned additional value at acquisition.
A major venture-backed medical device success story, Kyphon developed a novel treatment for the spine, achieved $260M in revenue prior to obtaining reimbursement codes, and was ultimately acquired by Medtronic for $4B. For Kyphon, the key to success was obtaining physician codes, called CPT codes, which allow doctors to get paid for the procedures they perform. A new procedure typically will not yet have a code and therefore must use a catch-all code, called an unlisted code. Unfortunately, since it’s a catch-all code, insurance companies don’t know what’s been done and typically won’t pay for it without further explanation and documentation. That alone is enough to prevent some physicians from performing a new procedure because it requires their time and staff time to gather the documentation and communicate it to the right people in the insurance companies at the right time.
To add to the difficulty, insurance companies typically view new procedures as experimental and investigational, even after they are FDA-approved. They therefore tend to automatically deny claims made for newer procedures as a policy until the procedure is more established in the clinical community. This leaves the company in a catch-22: They can’t drive a significant volume of procedures until the procedure is reimbursed, but they have to drive significant procedure volume in order to get reimbursement.
Kyphon was able to make this happen by removing much of the administrative burden from the physicians and their office staff. They built a reimbursement infrastructure with about 10 staff, giving them about a 1:15 ratio of staff to sales reps.2 This consolidated the expertise so that not all physician offices would have to become experts on the forms, rules, and timing of the unlisted-code reimbursement process. This took a cost burden from the physician’s office and increased the success rate for reimbursement. Kyphon worked this process from 2001 through 2005, ultimately driving revenue to $300 million before they received their own code in January of 2006 (Figure 3).3 They were also building their clinical evidence with trials and publications while they were driving procedure volumes.
Figure 3. Kyphon’s U.S. Sales 2001–2007.
Michael Lachman, “Vertebral Compression Fracture Market Heats Up with New Technologies,” Medtech Insight (2008, April 1).
St. Francis—Hospital Codes
St. Francis Medical had a different challenge. They were introducing a new product and procedure for which the hospital was not receiving payment. Since the hospital typically bears the burden of new technology costs, this would have been a serious deterrent to adoption. St. Francis’s first step was to obtain a new code for the hospital, called an ICD-9 procedure code, and assign it to an existing payment category, called a DRG (Diagnostic Related Group). Using the ICD-9 code to link to the DRG allowed the hospital to utilize a prespecified payment rate for the new procedure, which defrayed some of the costs. Since the payment was not enough to cover the full cost of the new device, St. Francis applied for additional “technology” codes that would give hospitals incremental reimbursement for the device—one each for inpatient and outpatient settings.4 Like Kyphon, St. Francis dedicated substantial resources to reimbursement, with 12 of their 85 employees as part of the reimbursement infrastructure.5 With a trial that demonstrated both superiority to conservative management and non-inferiority to surgery, St. Francis was granted these new technology codes a year after launch. As a result, hospitals received an additional $4,400 in reimbursement each time the device was used in a procedure. St Francis was acquired two months after the inpatient add-on code went into effect. Figure 4 shows the reimbursement timeline for St. Francis.
Figure 4. Reimbursement Timeline for St. Francis Hospital.
Understanding the Healthcare Crisis and Its Effect on the Reimbursement Problem
Unfortunately, successful reimbursement is going to get harder. Not coincidentally, just three months before CMS decided on narrowed reimbursement for the more broadly FDA-approved carotid stent, Congress had passed the 2005 Deficit Reduction Act ordering CMS to cut billions from their budget. Congress had effectively told CMS to value cost over efficacy. Why? Healthcare spending was skyrocketing—spending per person had jumped nearly 5000% from $148 per year in 1960 to $7,421 in 2007.6 Real U.S. healthcare spending, adjusted for inflation, has tripled every 20 years since 1965 (Figure 5).
Figure 5. U.S. National Healthcare Spending per Capita.
Congressional Budget Office, Technological Change and the Growth of Health Care Spending (Author), January 2008.
Although the U.S. economy grew 54% between 1999 and 2008, workers only gained 5% in real wages. Healthcare premiums all but consumed the gains in GDP: All the growth for workers was “paid” to them in the form of healthcare benefits. Since these payments are not typically transparent to workers in the U.S. employer-sponsored system, the enormous change was largely unseen, except by those outside the employer-sponsored system. The rise of the uninsured, driven out by cost, was the most visible sign, the “canary in the coal mine.”
The numbers in Figure 6 reveal an unsustainable trajectory underlying the rise of the number of uninsured in the United States. The rise of the uninsured is not a policy issue—it’s an economic one. Although access to healthcare is the real problem, the cost of care has created the access issue. Broadening access, although important for many reasons, will not solve the underlying problem. The costs of healthcare are eating away at our economy. At this pace, the Congressional Budget Office estimates that healthcare spending will soon consume almost half of our economy (Figure 7). Our economy is suffering and our rising uninsured population is suffering. We should indeed get everyone into the insurance pool, but that is not going to solve the underlying issue behind this crisis.
Figure 6. Cumulative Change in Healthcare Premiums 1999–2008 Relative to GDP, Earnings, and Inflation.
Henry J. Kaiser Foundation, Healthcare Costs; Bureau of Economic Analysis, GDP figures: June 25, 2009 (U.S. Department of Commerce).
Figure 7. U.S. National Healthcare Spending as a Percentage of GDP.
1960–2007 data from Henry J. Kaiser Foundation, Healthcare Costs; 2035–2082 projections from CBO Estimates, “Technological Change and the Growth of Health Care Spending” (Congressional Budget Office), January 2008.
As long as healthcare costs continue to rise, someone has to pay the bill. Money is going to come out of our paychecks to pay an ever increasing healthcare bill, whether it be through reduced wages in employer-based insurance, higher taxes in government sponsored programs (Medicare, Medicaid), or higher out of pocket costs. Worse yet, the higher healthcare costs could result in loss of jobs because a higher relative cost of U.S. employees, which includes healthcare costs, drives business overseas.
What is behind this dramatic rise in healthcare costs? Medical technology has been blamed as a driver of the increase. It is difficult to calculate this cost directly, but one analysis estimates that advances in medical innovation and the care it enables (including drugs, medical technology, and all related patient care) accounts for as much as 50% of the increase in healthcare cost.7 Of course, medical technology is not simply a cost driver; it has been contributing to a steady rise in longevity and quality of life. Slowing innovation would likely result in a trade-off in the advancement of medical care.
Two major demographic trends are also driving healthcare costs: the aging population and the obesity epidemic. Annual healthcare expenditures are $8,776 for someone over 65, compared to $2,330 for someone between ages 25 and 44.8 This cost differential will become even more important as the first Baby Boomers turn 65 this year; by 2030, the population over 65 is projected to rise to 20% of the total population, up from just 12% in 2000.9 The future cost implications are staggering.
The obesity epidemic is having perhaps the most significant impact on healthcare costs. Obesity is clearly implicated in diabetes and in cardiovascular and orthopedic costs. The annual cost of obesity nearly doubled between 1998 and 2008, from $78.5B to $147B.10 Between 1987 and 2001, the rise in U.S. obesity was responsible for more than a quarter of the increase in healthcare spending;11 during that same time, obese people cost the healthcare system an average of $1,429 more per year (Figure 7). Given that a third of U.S. adults12 and 17% of U.S. children13 are obese, the aggregate cost to our system is evident.
A Different Problem Altogether
I originally got into reimbursement to understand how to get medical devices reimbursed and how much that would cost our companies, so I could add value and differentiate myself within venture. I took the job no one wanted because I realized how important it was becoming to our portfolio.
Ultimately, I came away with a new investment thesis: removing costs from our healthcare system. Limiting rising healthcare costs will be an imperative to our economy and therefore could represent the next big wave of healthcare investment.
Cost reduction is a path fraught with political landmines, as evidenced by our healthcare reform debate. Little progress has been made on the cost control aspect of healthcare; the larger effort focused on increased access. Our current systems, as detailed above, do not reward anyone (physician, hospital, or patient) for being good stewards of the healthcare dollar. The physicians who are closest to the patients and best able to make critical decisions are paid for the amount of services they provide. The system encourages physicians to provide services whether they are needed or not, and does not pay them to invest time and money to avoid unnecessary care for their patients. Hospitals have disincentives to move procedures from highly paid inpatient settings to the less costly outpatient setting. No one is paid for coordinating the care of patients with multiple conditions. Patients themselves have no economic incentive to take care of their health, particularly when it is often less expensive to eat poorly (calorie dense foods are cheaper) and exercise less (gym memberships are expensive and the time to use them is often scarce).
There is a lot of talk and even some movement around making changes in the economic incentives for health. Money has been allocated for physicians and hospitals to invest in electronic medical records, which will help move the digital information infrastructure forward somewhat. Medicare has stopped paying for certain hospital-acquired infections, which has incented hospitals to make investments in infection reduction. Medicare is also considering reducing the reimbursement to hospitals if a patient is readmitted for the same diagnosis soon after discharge; this is meant to incentivize providers to appropriately discharge and follow-up with patients.
We must build upon these changes to create appropriate incentives to reduce the costs in our system. People do what they are paid to do. The current system rewards procedures, tests, and volume of care. We need a system that rewards good stewards of the healthcare dollar. With those incentives in place, we have a chance to affect the rising healthcare costs that are threatening our economy.
Currently, we seem to be fighting innovation by continually raising the regulatory and reimbursement hurdles, hoping that slowing innovation will sufficiently slow the rise of our healthcare costs. Instead of stifling innovation, we should redirect it. I propose that instead of fighting innovation in the healthcare sector, we release the power of U.S. innovation and turn our drive to innovate toward solving the problem of healthcare costs. Our current system of healthcare reimbursement does not typically reward technologies and innovations that save costs, so there is comparatively little money invested from the venture capital community toward that goal. By putting the right incentives in place, our government could enlist healthcare entrepreneurs to solve our cost problem for us.
In the meantime, there is a model for getting medical devices reimbursed in the old system, albeit at an increasing cost. Success is dependent on investing in the right infrastructure and clinical data. Data is king in this environment and it is critical to factor in the investment and timeline up front to accurately access the return on your investment. Some improvements to care will no longer be investable given the higher investment thresholds, but large markets will justify the increased investment.
Amy joined Advanced Technology Ventures in 2006 as a member of the firm’s healthcare practice where she focuses on investments in the medical device sector. Before her fellowship, Amy worked in global marketing at Guidant where she led their first Drug Eluting Stent launch internationally. She previously launched and managed the profitability of angioplasty and radiation therapy products in the US, Europe and Asia. Amy has also worked for Bristol-Myers Squibb on the marketing of a variety of cardiovascular, HIV, and neuroscience products. Amy holds an MBA from UC Berkeley and an AB in Economics from Yale.
1 Henry J. Kaiser Family Foundation, Health Care Costs, A Primer: Key Information on Health Care Costs and Their Impact (Author: 2009), 8.
2 Kyphon, Form 10-K: Fiscal Year 2003.
3 Scott Huntley, “Targeting Vertebral Compression Fractures,” Medtech Insight (2007, April 1).
4 David Cassak, “Kyphon Steps Up,” IN VIVO (2007, May 1); David Cassak, “St. Francis Medical: Staking New Ground in Dynamic Stabilization,” IN VIVO (2006, March 1); St. Francis Medical, S-1, Sep 2006; Lisa Sibley, “Reimbursement is Crucial in Life of Medical Products,” Silicon Valley/San Jose Business Journal, June 13, 2008.
5 Cassak, St. Francis Medical.
6 Henry J. Kaiser Foundation, Healthcare Costs.
7 Congressional Budget Office, Technological Change and the Growth of Health Care Spending, January 2008, quoting data from a study by Smith, Heffler, and Freeland, “The Impact of Technological Change on Health Care Cost Increases: An Evaluation of the Literature,” 2000.
8 Henry J. Kaiser Foundation, Healthcare Costs.
9 Credit Suisse, “Eye on the Election,” July 10, 2008, quoting figures from the U.S. Census Bureau.
10 Eric A. Finkelstein et al., “Annual Medical Spending Attributable To Obesity: Payer-And Service-Specific Estimates,” Health Affairs (2009, July 27): w822—31, doi 10.1377/hlthaff.28.5.w822.
11 K. E. Thorpe et al., “The Impact of Obesity on Rising Medical Spending,” Health Affairs (Millwood) 23(2004): w480—86.
12 Katherine M. Flegal et al., “Prevalence and Trends in Obesity Among US Adults, 1999—2008,” JAMA 303, no. 3 (2010): 235–41, doi 10.1001/jama.2009.2014.
13 Centers for Disease Control, “2007-2008 National Health and Nutrition Examination Survey (NHANES).”