With an estimated size of over $133 billion in 2016, the United States medical device market is a multibillion dollar industry1. The U.S. Food and Drug Administration (FDA) is responsible for regulating the medical devices that are sold in the United States. Currently, the 2 main routes of medical device approval are the premarket approval (PMA) process, which requires clinical trials, and the less-rigorous 510(k) premarket notification route, which exempts devices from clinical trials if they prove to be “substantially equivalent” to an existing device2. Class-III (high-risk) devices are subject to PMA authorization, while Class-I and Class-II (low-to-moderate-risk) devices undergo less-stringent approval requirements3.
The 510(k) premarket notification is an expedited process that allows for approval should a device be “substantially equivalent,” or at least as safe and effective, as a legally marketed device2,3. This process was designed to fast-track iterations that were readily able to be demonstrated at least as safe and effective as previously approved products in order avoid costly, time-consuming clinical studies3. Although this process allows medical device innovations to reach patients promptly, safety issues have emerged4,5. Devices approved through the 510(k) route have been shown to result in an 11.5-fold increased risk for recall when compared with PMA-authorized devices. This factor becomes 3.5, 13.2, and 8.5 when recalls are stratified as having an FDA-determined “reasonable,” “remote,” or “not likely” probability, respectively, to cause adverse health consequences4. Furthermore, up to one-quarter of high-risk devices are approved through 510(k) review6.
Unless a device qualifies for the humanitarian device exemption, which is used for devices that treat or diagnose diseases that affect <4,000 individuals per year in the U.S., the only other device approval pathway available is the PMA2,3. Unlike the 510(k), PMA authorization requires clinical evidence, usually on the level of randomized clinical trials or prospective data compared with historical controls, before approval is granted2,3. As a result, devices approved through the PMA process are less likely to be recalled4,5. Although this route decreases the potential for device recalls, it is also associated with substantially higher costs and longer submission-to-approval timelines6,7. These concerns have led medical device manufacturers to ask for more streamlined FDA processes to avoid excessive costs and delays in order to preserve innovation and keep the United States competitive in the global medical device marketplace7,8.
The majority of orthopaedic medical devices are used in joint replacement or fracture management procedures, with projected U.S. markets of $10.3 billion in 2018 and $4.3 billion in 2015, respectively9,10. However, the market size of various devices within these categories can vary drastically11.
While previous studies have assessed the safety and costs of the 510(k) and PMA authorization processes4-7,12, no study, to our knowledge, has addressed the economic feasibility of approving various devices through these 2 routes. Most medical device companies, similar to companies and investors in other industries, would only be willing to invest in the upfront costs of medical device innovation and approval if they could recoup their investment within 7 years13-18. Utilizing financial modeling, our study assesses the financial feasibility of approving orthopaedic medical devices through the 510(k) and PMA processes. We hypothesized that several small-market orthopaedic devices will not be financially viable via PMA authorization because of the large market size needed to cover the increased costs associated with the PMA process and the relative rarity of these procedures.
Materials and Methods
Medical Device Categorization
Medical devices are approved through the FDA as systems. Our analysis focuses on a representative collection of orthopaedic medical device systems and their respective financial approval feasibility based on time to recoup an initial investment. The majority of orthopaedic devices are used in joint replacement or fracture management procedures. Orthopaedic devices used in joint replacement surgeries include hip, knee, shoulder, wrist, and finger implants, while fracture management devices include plates, screws, pins, spinal fixation devices, and others19. However, the usage of these device types varies drastically based on anatomical location11. Because of these variations, we divided orthopaedic devices by the procedure type (joint replacement, fracture management, and spinal fusion) and by anatomical location (hip and thigh, knee and lower extremity, shoulder and upper extremity, and spine). This resulted in a selection of 13 medical device systems for analysis (Table I).
For each medical device system categorized, we used the following statistical model to determine the length of time it would take a medical device company to recoup the investment from either the PMA or 510(k) route after FDA submission.
The difference between the unit revenue per device and the unit cost per device provides the unit profit per device. The unit profit per device, when multiplied by the number of devices sold per year for each system, gives the total profit per year for each medical device system. The PMA or 510(k) concept-to-market investment cost divided by the total profit per year provides the number of years, after FDA approval, that a company could expect to sell their product before recouping its investment. This figure, when added to the average FDA submission-to-approval time for either the 510(k) or PMA process, provides the expected time to recoup investment costs for each medical device system after submission to the FDA.
The expected time to recoup investment was assessed relative to a 7-year return benchmark, a commonly used length of time that investors target to either realize their investment or execute an exit strategy13-18.
The total concept-to-market cost for medical device companies to develop and approve a device through the PMA and 510(k) routes was used to represent investment cost and has been previously determined7. The average total concept-to-market cost of PMA-authorized devices is $94 million, with $75 million spent on FDA-related activities, while the average total concept-to-market cost of 510(k)-approved devices is $31 million, with $24 million spent on FDA-related activities7. In these price points, product testing and clinical studies were considered FDA-related.
Unit Revenue per Device
The purchase price that a level-I trauma center teaching hospital paid medical device companies was used to approximate the revenue per unit sold for each medical device system. The price points reflected amounts paid during the 2014 calendar year.
Unit Cost per Device
Cost per unit sold for each device system was estimated using gross margins. Gross margins, defined as the difference between total sales revenue and cost of goods sold divided by total sales revenue, represent the proportion of revenue that a company retains after accounting for the costs of producing a good20,21. The costs of goods sold represents the variable and fixed costs directly associated with the sale of a good, including material costs, labor costs, shipping costs, and others20.
By rearranging the equation, we are able to solve for the cost of goods sold, equivalent to the unit cost per device, with the unit revenue per device and gross margin data. The unit revenue per device was determined as described above.
To determine the gross margin for orthopaedic medical devices, net sales and costs of sales data for the top 10 largest orthopaedic medical device companies were curated from publicly available company filings from the U.S. Securities and Exchange Commission22. The gross margin for each company was calculated for each year from 2010 to 2014. Because gross margin can vary drastically by industry, we only analyzed orthopaedic medical device companies that predominantly create orthopaedic medical devices20,21. For this reason, Johnson & Johnson (pharmaceuticals, consumer health products) and Medtronic (52% cardiac and vascular devices), which create products outside of the orthopaedic device industry, were excluded from the gross margin analysis. In addition, because our analysis was confined to the FDA and U.S.-approved devices, companies based outside of the U.S. were removed. As a result, Smith & Nephew, a British-based company, was excluded. Arthrex, a private company without publicly available gross margin information, was also excluded.
Gross margin for the remaining 6 companies was averaged across each year from 2010 to 2014 to create the average gross margin for devices sold in the orthopaedic medical device industry (73.6%; Table II). With the average gross margin for orthopaedic medical devices and unit revenue per device data points, we calculated the unit cost per device for each orthopaedic device system.
The difference between the unit revenue per device and the unit cost per device provided us with the unit profit per device (Table III).
Number of Devices Sold per Year (Market Size)
The number of units that a device company could expect to sell per year was determined by multiplying the total number of procedures that utilize a specific device per year in the United States by the market share that a company could expect to capture (Table III).
The total number of procedures per year associated with each medical device system was curated from the Healthcare Cost and Utilization Project of the Department of Health and Human Services (Table III)23. Queries were performed using ICD-9-CM (International Classification of Diseases, Ninth Revision, Clinical Modification) procedural codes for 2012, the most recent year with available data.
Expected market share was determined by averaging the worldwide market share for the top 10 orthopaedic medical device companies in sales (8.77% per company)24.
FDA Approval Time
The average submission-to-approval time for devices through the 510(k) process for fiscal year 2010 was 4.5 months, while the average submission-to-approval time for devices through the PMA route was 27 months over the same time period25.
Our results are summarized in Table III. The unit profit per device varied from $913.41 for the radial plate system to >$6,000 for the anterior lumbar plate system. Furthermore, the total number of procedures per year associated with each medical device ranged from just above 24,000 for the reverse shoulder replacement to >600,000 for the knee implant. From these data, total expected profit per year ranged from around $4 million for the radial plate system to almost $150 million for the knee implant. Joint replacement devices had a higher average expected profit per year ($63.7 million) than fracture management and spinal fusion devices ($22.4 million and $27.4 million, respectively).
The expected time to recoup investment through the 510(k) process ranged from 0.585 years (knee implant) to 7.715 years (radial plate system), with an average time of 2.4 years. For the same set of device systems, the expected time to recoup investment through the PMA route ranged from 2.9 years (knee implant) to 24.5 years (radial plate system), with an average time of 8.5 years (Fig. 1).
Six of the 13 orthopaedic device systems that we analyzed (tibial plate system, shoulder replacement, reverse shoulder replacement, radial plate system, anterior cervical spacers, and anterior cervical plate system) would require longer than 7 years to break even on investment costs if required to go through the PMA route. However, only the radial plate system (7.715 years) would require longer than 7 years to break even if processed through the 510(k) route. If required to go through the same approval process, a radial plate system would require 8.4 times the amount of time to recoup its investment costs when compared with a knee implant.
Because of the high costs and lengthy approval timeline associated with the PMA process, broad requirements for PMA authorization for orthopaedic medical devices are likely not financially feasible. While high-volume and high-margin medical device systems will still be able to attract investors and innovate within a 7-year window, many low-volume and low-margin devices may run into investment hurdles. Each of the 6 devices that would take more than 7 years to recoup investment through the PMA had a market size of <160,000 procedures per year in our model. The other 2 devices with a market size in the low range, the anterior lumbar plate system and the interbody cage system, were buoyed by having the 2 highest calculations for profit per unit sold. Low-volume devices, such as the reverse shoulder replacement, were often more recent creations when compared with the long-standing, high-volume hip and knee implants. These differences may decrease as newer devices gain popularity with more surgeons and patients.
If the PMA authorization route was broadly required, large-market devices such as hip and knee implants, with market sizes of >400,000 procedures per year, and high-margin spine devices would receive a disproportionate amount of investment funding. This would result in a dearth of innovation for less-prevalent orthopaedic conditions, including radial plate systems or reverse shoulder replacements. In these situations, patients would suffer not from safety concerns, but from the lack of novel device development.
With a quicker approval process and cheaper associated costs, the 510(k) process is an attractive option for medical device companies6,7,25. Indeed, 88% of orthopaedic devices cleared in 2012 were approved through the 510(k) process4. However, the safety of approving devices through the 510(k) process has been called into question5,6,12.
With the 510(k) premarket notification, we have a system in which moderate-risk and some high-risk devices enter the market because they are “substantially equivalent” to devices that had been approved years ago. This system unravels, however, if the application of the “substantial equivalence” assumption is unsound3. Furthermore, substantial equivalence can be compounded all the way back to the date that the current approval framework was established in 1976, allowing some products to be grandfathered into the market without ever having been assessed for safety or effectiveness26.
Schemes for new frameworks have been varied, with proponents arguing for increased postmarket monitoring3 and increased regulation5, as well as considerations to ensure a healthy climate for innovation3,7,25. With this uncertainty, proponents have recommended that the FDA obtain adequate information, which does not currently exist, to inform the design of a new, effective clearance process26. By elucidating the financial feasibility of the current approval routes, our study begins to fill in these gaps in knowledge.
It may be best, therefore, to develop a novel, integrated premarket and postmarket regulatory framework. Any effective FDA-approval framework would strive to ensure safety while promoting innovation26. Previous suggestions have included calls for increased postmarket surveillance and collaboration with medical device manufacturers3,26. With this in mind, a small-scale, site-limited rollout of medical devices jointly funded by the FDA and device manufacturers would allow the FDA to accelerate the lengthy PMA process while still promoting safety through a small-scale trial period. Under this framework, the FDA would need to ensure that it possesses adequate postmarket analysis and regulatory tools. The emergence of registries as postmarket surveillance tools can help bridge this transition27.
As with all research, we acknowledge some limitations to our study. The statistical model that we used is limited in the number and scope of medical devices that were considered, and may not be entirely representative of the trend in all of orthopaedics or other specialties. In addition, we used price points from a single center that likely has more bargaining power than other private providers. This can potentially skew the margins that companies expect to receive on each device because of cost variations among and within institutions. We also made the key assumption that average concept-to-market cost is consistent by device type. Furthermore, the margins on each device were estimated using gross margins from the orthopaedic device market. The 27-month PMA authorization timeline could be an underestimation, as other studies have quoted this to be 31 months7. However, a longer timeline would only support our conclusion by underestimating the PMA recoupment time. Finally, the $31 million concept-to-market cost for 510(k) approval could be an overrepresentation. While the cost of industry-wide research and development (R&D) spent per product was estimated at around $50 million for PMA-authorized devices, only $3 to $5 million was spent toward R&D per product for 510(k)-approved devices28. Despite this, a larger real difference between the concept-to-market costs for 510(k) and PMA authorization would only increase the incentive for companies to utilize the 510(k) process, further disincentivizing innovation toward devices that treat less-prevalent orthopaedic conditions and require PMA authorization.
Because of the lack of publicly available data on specific costs and margins per device, we were required to use financial modeling and approximations in our study. These financial data points are often censored and not always peer-reviewed because companies are not willing to disclose their proprietary information for fear of losing competitive advantage. Sensitive financial information on a company’s product margin per device or selling price to various providers would provide ample room for competition and contract negotiation from competing firms. Because of these issues, other studies that estimated the cost of FDA approval did so through complete company anonymity7.
We believe that our paper has considerable value because of the fact that it most closely approximates the real data. Despite its limitations, to our knowledge, this is the first report to assess the economic feasibility of medical devices of any type through the current medical device approval routes. Calculation assumptions used in the model remained consistent throughout the statistical model, promoting the ability to compare across device types and demonstrating the need for more transparent data in the field. With better data, we will be able to improve on this study in order to inform the design of a more effective FDA device regulation process.
This report demonstrates how current regulatory policies can potentially influence orthopaedic device innovation. We hope that it can serve as a catalyst for future inquiries into the economics that will help the design of true regulatory framework reform. With more transparent, mandatory, anonymized data, we will be able to not only advance medical device innovation through more efficient policies, but also ensure that the public and our patients do not incur any health risks.
Investigation performed at Harvard Medical School, Boston, Massachusetts
Disclosure: On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work (http://links.lww.com/JBJS/A141).