New Technology and Health Care Costs — The Case of Robot-Assisted Surgery

Original article in NEJM: http://www.nejm.org/doi/full/10.1056/NEJMp1006602

Technological innovation in health care is an important driver of cost growth. Doctors and patients often embrace new modes of treatment before their merits and weaknesses are fully understood. These technologies can lead to increases in costs, either because they are simply more expensive than previous treatments or because their introduction leads to an expansion in the types and numbers of patients treated. We examined these patterns as they apply to the case of robot-assisted surgery.

Robotic surgical devices allow a surgeon at a console to operate remote-controlled robotic arms, which may facilitate the performance of laparoscopic procedures. Laparoscopic surgery, in turn, is associated with shorter hospital stays than open surgery, as well as with less postoperative pain and scarring and lower risks of infection and need for blood transfusion.

Robotic technology has been adopted rapidly over the past 4 years in both the United States and Europe. The number of robot-assisted procedures that are performed worldwide has nearly tripled since 2007, from 80,000 to 205,000. Between 2007 and 2009, the number of da Vinci systems, the leading robotic technology, that were installed in U.S. hospitals grew by approximately 75%, from almost 800 to around 1400, and the number that were installed in other countries doubled, from 200 to nearly 400, according to Intuitive Surgical, da Vinci's manufacturer.

A wide range of procedures are now performed by means of robot-assisted surgery. Some of these procedures were already being performed laparoscopically before robots were introduced; the introduction of robotic technology affects expenditures associated with such procedures primarily by increasing the cost per procedure. For procedures that were more often performed as open surgeries, the introduction of robots may affect both the cost and the volume of surgeries performed.

Robotic surgical systems have high fixed costs, with prices ranging from $1 million to $2.5 million for each unit. Surgeons must perform 150 to 250 procedures to become adept in their use. The systems also require costly maintenance and demand the use of additional consumables (single-use robotic appliances). The use of robotic systems may also require more operating time than alternatives. In the case of procedures that had previously been performed as open surgery, however, some of the new costs will be offset by reductions in postoperative hospital costs and by productivity gains if patients recover more rapidly and can return to work and other activities sooner.

Estimates of the net per-procedure cost of robot assistance vary with assumptions about the frequency with which a robot will be used. We examined all the cost studies of robot-assisted procedures published since 2005 and found that, on average, across the full range of 20 types of surgery for which studies exist, the additional variable cost of using a robot-assisted procedure was about $1,600, or about 6% of the cost of the procedure in 2007. (For a detailed description of the methods, see the Supplementary Appendix, available with the full text of this article at NEJM.org.) When the amortized cost of the robot itself was included, the additional total cost of using a robot-assisted procedure rose to about $3,200, or about 13% of the cost of these procedures in 2007 (seetableProcedures Performed by Robot-Assisted Surgery, Current Cost per Procedure, and Estimates of Change in Cost, as Compared with Standard Procedure.). Intuitive Surgical is now the sole producer of robotic surgical devices; prices may decline in the future if there is more competition in the market for machines or related consumables.

This additional hospital spending is likely to affect the expenditures of public and private insurers — but only indirectly. New diagnosis-related and procedure codes now exist for robot-assisted surgery, but currently Medicare and most U.S. private insurers do not pay additional fees for this service. Instead, hospitals are likely to increase charges for procedures or diagnoses for which robots are used. Medicare and private-insurer reimbursement rates are computed on the basis of these charges. Thus, increases in charges for robot-associated diagnoses and procedures may eventually generate higher payments that are never explicitly linked to robot-assisted surgery.

Many procedures that can now be performed laparoscopically with robots used to be performed most often as open surgeries. Robot-facilitated substitution of laparoscopy for open surgery has occurred most prominently in the case of prostatectomy for localized prostate cancer.1

Various nonsurgical treatment alternatives for localized prostate cancer exist, with similar long-term outcomes and varying side effects, such as incontinence and impotence.2 Existing analyses suggest that as compared with open surgery, robot-assisted surgery does not diminish the frequency of these adverse effects.1 However, the substantial short-term benefits in terms of postoperative recovery that are associated with robot-assisted procedures may encourage patients with localized cancer to opt for surgical intervention rather than alternative nonsurgical interventions or watchful waiting.

Data from the Nationwide Inpatient Sample show an increase of more than 60% in the number of hospital discharges for prostatectomy (including both robotic and traditional procedures) in the United States between 2005 and 2008 (see graphProstatectomies in the United States, 2000–2008.).3 This increase occurred despite a decrease in the underlying incidence of prostate cancer4 and contemporaneously with a striking increase in the number of robot-assisted prostatectomies performed in the United States. The observed pattern matches evidence from the Surveillance, Epidemiology, and End Results Medicare database, which shows that Medicare beneficiaries (65 years of age or older) who received a diagnosis of prostate cancer in 2005 were about 14% more likely to have undergone surgery by 2007 than were their counterparts whose prostate cancer was diagnosed 3 years earlier.5 These patterns suggest that robotic technology may have contributed to the substitution of surgical for nonsurgical treatments for this disease. In this case, the introduction of the robotic technology may have increased both the cost per surgical procedure and the volume of cases treated surgically. However, the evidence suggests that despite the short-term benefits, robotic technology may not have improved patient outcomes or quality of life in the long run.

It is unlikely that robot-assisted surgeries will completely replace conventional surgeries for the full range of procedures for which cost studies have been done. If such a substitution did occur, however, it would generate nearly $1.5 billion in additional health care costs annually — excluding the (amortized) costs of the robots, which would bring the total to more than $2.5 billion. If robots also led to an increase in the volume of surgery performed, as suggested by the prostatectomy case, their effect on total annual expenditures would be even greater.

Who will consider these costs? The diffusion of robotic technology depends on fragmented, not centralized, decision making. Decisions to purchase robots are made not by payers but by hospitals, which compete with one another to attract surgeons and their patients. Hospitals, seeking surgical volume, find it difficult to resist surgeons' preferences, even without favorable direct reimbursement, and surgeons feel compelled to keep up with market demands so as not to lose patients.

Comparative-effectiveness research — often considered a corollary of regulatory cost containment — can play a critical role in this decentralized process. To date, there have been no large-scale randomized trials of robot-assisted surgery, and the limited observational evidence fails to show that the long-term outcomes of robot-assisted surgery are superior to those of conventional procedures.1 Evidence from well-designed, large-scale, multicenter trials or comparably rigorous nonrandomized evaluations is needed to determine which patients benefit from open surgical approaches and which from robot-assisted approaches. Hospitals could use this information in response to pressure from technophile surgeons; surgeons could use it in discussing treatment options with patients; patients could use it to make treatment choices; and payers could use it in negotiating reimbursements. An efficient health care system must enhance the ability of medical professionals and their patients to make informed choices about the adoption and use of new technologies, even when insurers do not explicitly provide reimbursement for these new technologies.