Dr. Goel is from the Institute for Clinical Evaluative Sciences in Ontario, North York, Ont., and the Department of Preventive Medicine and Biostatistics and the Clinical Epidemiology and Health Care Research Program (Sunnybrook Unit), University of Toronto, Toronto, Ont.
Canadian Medical Association Journal 1995; 153: 1739-1741
[résumé]
In the span of 15 years extracorporeal shock-wave lithotripsy (ESWL) has evolved from a promising new technology(1) to a standard treatment for urinary stones. ESWL is one of many innovative therapies that provide an alternative to more invasive surgical procedures. The introduction of these new treatments usually results in reduced patient morbidity and often leads to shorter hospital stays or to the possibility of same-day procedures -- and thus to a decrease in costs. However, these treatments do come at a price. The technology they require is usually expensive in terms of both capital expenditure and operating costs. Many have unknown risks or untoward side effects. And, more often than not, their use in the health care system quickly becomes diffuse, often before a rigorous scientific evaluation has been made.
New medical technologies do not undergo the sort of review that is expected of new drugs. They fall through the cracks in the system: no statutory responsibility for their evaluation rests with government or with professional colleges.(2) Although such technologies are often introduced for a limited range of indications, there is a tendency for broader indications to be adopted without further evaluation as time goes by. Unfortunately, once a technology is established and accepted, randomized trials become unfeasible or unethical.(3)
In this issue Adrian R. Levy and Maurice McGregor (see pages 1729 to 1736 [abstract]) demonstrate quite elegantly the expansion of the use of ESWL for the treatment of urinary stones in Quebec. Using physicians' claims and hospital discharge data they show that the use of ESWL grew rapidly after its introduction in the mid-1980s and then continued to increase, at a slower rate, through 1992. In the same period the number of surgical procedures performed declined. Most striking is the finding that the total number of people treated for urinary stones in Quebec increased by approximately 59%. There are a few possible explanations for this increase in the use of ESWL for urinary stones. Repeat treatments in different years for the same person is one, but Levy and McGregor show that double-counting would at most account for 3% of the cases. A true increase in the incidence of urinary stones in the population is another. However, although there is evidence of an increase in renal stone incidence over the longer term(4,5) it is hard to believe that this would be observed in the span of 2 years in one province. A third explanation is that the increase reflects the treatment of a backlog of people for whom the risks of invasive procedures had been too great. This is certainly a likely explanation for the sharp increases seen in 1987 and 1988: patients with urinary stones who were either ineligible for surgery or who did not want to take on the risks or morbidity associated with it would certainly be expected to quickly embrace a technology such as ESWL. However, such a backlog would be treated in a relatively short time, and in the absence of other factors, the number of people treated should stabilize or fall. In fact, the rate of use of ESWL in Quebec continued to grow, albeit at a slower rate.
A broadening of the accepted indications for ESWL is the most likely explanation for its increased use. This is not unexpected, given that it has already been noted in the clinical literature.6 Levy and McGregor's population-based study demonstrates that a widening of indications can occur quite rapidly after the introduction of a new technology.
Their study also shows that ESWL costs less than surgery per case treated, largely because of a reduction in hospital stay. This is true even when capital costs are taken into account. However, the total costs for treating urinary stones in Quebec increased somewhat in the study period because of the increase in the number of people treated. Further, the introduction of ESWL led to a shift in costs from hospitals that did not have lithotriptors to hospitals that did.
Given the nature of hospital budgets in this country and the widespread restructuring of hospitals that is occurring, it is important that such shifts in costs be recognized. The funding of new technology extends beyond considerations of capital costs: it has an impact on direct and indirect operating expenses as well. The "savings" that can be realized from new technologies will sometimes appear elsewhere in the system - in this case, in the hospitals that are now doing less urinary-stone surgery. Under current hospital funding formulas, these savings may not actually be realized, as funds are redirected to new programs or the expansion of existing programs.
It is also important to assess the impact of these changes on patient outcomes. In the case of ESWL, the overall health of the population could be improved in two ways: by achieving a better health status for those who receive the treatment, and by redirecting resources to other efficacious and cost-efficient procedures. However, the high retreatment rate for ESWL reported by Levy and McGregor -- 34% -- suggests that further evaluation is required. The documentation of current treatment outcomes is important; the rate of retreatment reported in the present study surely reflects the effectiveness of ESWL while it was still being introduced. A "learning curve" will apply to the introduction of any new technology, and outcomes may be less optimal during the early phases of its use.
Levy and McGregor's study highlights the importance, and limitations, of research that relies on large administrative databases. Such research needs to be conducted longitudinally, such that data files are examined for long-term individual outcomes rather than on a year-by-year basis. The ability to link not just across years but also across different files (e.g., physicians' claims, hospital discharge records and drug benefit records) is crucial to the assessment of the impact of new technologies on a population. This requires further methodologic work, an adequate infrastructure for processing and analysing large data files and appropriate consideration of privacy and confidentiality. It is crucial that the quality of these data, particularly with respect to reliability and validity, be established. For example, Levy and McGregor had to rely on a study of hip fractures that used the same database as they did to document the validity of procedure coding.
New technologies are being introduced into the health care system at a phenomenal rate. Many will come into widespread use before they are thoroughly evaluated. Research such as that by Levy and McGregor will be necessary to document the dissemination and diffusion of such technologies. Ultimately, the use of linked data files will make it possible to assess the impact of such technologies on population health. Although randomized controlled trials will remain the preferred method for assessing health interventions, observational research with administrative data will, realistically, provide the only evaluation of many new technologies.
The evaluation of new drugs is time consuming, and the procedures for obtaining approvals are often regarded as overly bureaucratic. However, this process does help to ensure that drugs put on the market are usually safe and efficacious. The institution of a similar, and perhaps streamlined, approval process for new technologies should be considered.
Dr. Goel's work is supported in part through a National Health Scholar Award from Health Canada.