In the first article of this series (“High Tech, High Cost: Examining the Increasing Costs of Neurosurgical Cares,” Vol. 8, No. 4) I argued the case that new technology is the primary driver of increasing health care costs in the developed world. The second article (“High Tech, High Costs: Examining Technology and Health Care Reform,” Vol. 9, No. 1) reviewed various proposals for health care reform and discussed the impact such proposals would have on developing and deploying new technology. The third article (“High Tech, High Costs: Fostering Neurosurgical Innovation,” Vol. 9, No. 2) stressed the importance of fostering neurosurgical innovation in a health care environment that may become increasingly hostile to technological development. The final article in this series suggests ways in which technology assessment can be used to assure cost-effective neurosurgical care.
Health technology assessment has been defined as “the careful evaluation of a medical technology for evidence of its safety, efficacy, cost and cost-effectiveness and its ethical and legal implications, both in absolute terms and in comparison with other competing technologies.” In practice, investigators in technology assessment have most often employed the methodologies of decision analysis and cost-effectiveness analysis while broadening the scope to include ethical and societal implications.
I believe that this approach will be used more extensively in the future to evaluate medical care. These and two alternate approaches to technology assessment will be briefly reviewed.
Decision Analysis
Decision analysis methodology is used for evaluating the effectiveness of treatment using data generated in clinical trials. In decision analysis, therapeutic strategies are specified and clinically relevant outcomes for each strategy are analyzed in a decision tree model. Probabilities of reaching various clinical outcomes are determined from data collected in clinical trials. Each clinical outcome is then assigned a utility, such as quality adjusted life years (QALYs). Utilities are based on patient or potential patient perceptions.
The value of treatment strategies can then be calculated by multiplying the probability of reaching a clinical outcome by its utility. Stability of the results can be analyzed using sensitivity analysis, whereby the probabilities of reported outcomes are varied over a wide range. Decision analysis can be used to determine the most beneficial treatment approach for an individual patient faced with a difficult treatment decision or for analyzing the value of various treatment modalities for large populations of patients.
Cost-effectiveness Analysis
Cost-effectiveness analysis, a subset of decision analysis, includes the costs of various interventions. Investigators using this methodology attempt to determine the relative value of treatments by dividing the cost of a given intervention by the benefit obtained. Cost-effectiveness can then be expressed as dollars invested per QALY.
To use a neurosurgical example, we can postulate that chemotherapeutic agent X in glioblastoma patients costs $10,000 to achieve, on average, an additional six weeks of life with a utility of 0.06 QALYs. The cost-benefit of this intervention would be greater than $165,000 per QALY. This kind of analysis can be used to compare the relative values of various interventions. Medical technology assessment uses the data generated from such analysis and then tries to address the thornier issues of whether or not this expenditure of resources is warranted.
Limitations of Cost-effectiveness Analysis
Cost-effectiveness analysis is only as reliable as the data on which the analysis is done. Reliable clinical outcomes data, utilities that accurately assess patient desires and realistic cost data are necessary for cost-effectiveness analysis to be meaningful. All of these may be difficult to obtain for a given group of patients.
Alternate Methods foor Technology Assessment
Alternate approaches for technology assessment have been advocated, including appropriateness evaluation and strength of evidence evaluation. The two approaches are summarized here.
Appropriateness Evaluation. This approach employs a panel of experts and formal procedures for judging the appropriateness of a medical intervention in an attempt to reach a consensus on indications for care. This approach is, in my opinion, fraught with difficulties. Consensus criteria are subjective and highly dependent on the composition of the expert panel. The conclusions of any expert panel are suspect unless reliable clinical data are available.
Strength of Evidence Evaluation. This brings us to the second alternative, strength of evidence evaluation or evidence-based practice. Extreme advocates of strength of evidence evaluation recommend that only those interventions with clinical benefit documented by controlled clinical trials be performed and paid for. They point out that if we employed only those interventions whose effectiveness has been documented by controlled trials there would be no need to worry about costs, as many medical and surgical procedures would no longer be done. This approach is also rife with hazards.
Many interventions whose effectiveness has been documented by observational studies would be excluded because the strength of the evidence is limited. However, evidence from observational studies may be so compelling that a randomized trial is inappropriate. Do we need a clinical trial to document the effectiveness of antibiotics in postoperative infections? In addition, prospective, controlled trials are expensive and time consuming and it is unlikely that such trials will ever be done for many interventions.
There are also real constraints to the applicability of data from many prospective, randomized, clinical trials. The NASCET and ACAS carotid endarterectomy trials are an excellent example of this. These trials have documented the benefit of carotid endarterectomy for patients with symptomatic and asymptomatic carotid stenosis. However, the trial results may not reflect how carotid endarterectomy is done in the “real world.” For example, NASCET and ACAS involved only carefully selected hospitals where many carotid endarterectomies were being done. They also involved carefully selected patients and surgeons with documented low morbidity and mortality rates for this procedure. It is unlikely that the results of these trials can be generalized. In fact, outcomes studies using data from the HCFA database demonstrate complication rates for carotid endarterectomy in general practice far exceeding those in NASCET and ACAS.
Ethical Concerns with Technology Assessment
The use of cost-effectiveness analysis for medical technology assessment raises considerable ethical concerns. Using cost as a criterion for determining medical intervention raises difficult questions about rationing of health care.
This is not a new issue for neurosurgeons. We do not employ all of our technological resources for every patient. Do we take a demented 80-year-old patient with a Glasgow coma score of 3 to the operating room to clip his aneurysm? We believe we make such decisions based solely on perceived benefit to the patient without concerns about resource expenditure.
This feels right when we are dealing with acute life and death issues but it is difficult to justify in more nebulous circumstances. Should the adult population be screened with MR or CT angiography looking for unruptured aneurysms? Should everyone over age 65 have a carotid duplex study? These issues become more acute as technology makes new diagnostic and therapeutic interventions available. Unless there are limitless resources for medical care there will sometimes be a fundamental tension between decisions that are best for the group and the desires of an individual patient or physician. We will need to address theese issues.
The Goal of Technology Assessment
John Eisenberg, MD, director of the Agency for Healthcare Research and Quality, said, “Technology is rarely inherently good or bad, always or never useful. The challenge is to evaluate when in the course of an illness it will enhance outcomes, and how it should be implemented.” We all have a stake in creating “a health care system that recognizes value, rewards better outcomes and encourages efficient use of limited resources.”
The Role of the Neurosurgeon
As neurosurgeons we must recognize that our overriding commitment is to do the best for our patients, but we must also be prudent with the resources available to obtain the greatest benefit at the lowest cost. As users of medical technology we must participate in technology assessment by becoming involved in clinical trials, by developing and expanding outcomes assessment databases and by constantly asking ourselves, “What is the evidence that the use of this technology is benefiting my patients?”
It is important to involve the entire community of practicing neurosurgeons in this endeavor, to erase the boundaries between the authors and the readers of clinical data. For the first time advances in information technology make this an achievable goal. Employing online, secure clinical data collection, every neurosurgeon can be involved in generating data for analysis.
Real time analysis and feedback of the data will allow each neurosurgeon to determine the effectiveness of his or her care. With an adequate database, the most cost-effective practices can be determined and these results can be disseminated to the neurosurgical community. This is not a pipe dream. The Outcomes Committee of the AANS and CNS has taken the first halting steps to make these goals a reality. We have a long way to go but the obstacles are not insurmountable.
Technology assessment is a reality we must deal with. I encourage all of you to become involved in the outcomes studies in lumbar discectomy, acute subdural hematoma and treatment of carotid artery disease that are now available online. We can play an active role in technology assessment or we can wait for others to tell us which of our procedures are of benefit.
Robert E. Harbaugh, MD, FACS, is Professor of Neurosurgery and Director of Cerebrovascular Surgery at Dartmouth-Hitchcock Medical Center and Chair of the AANS/CNS Committee for the Assessment of Quality and the AANS/CNS Outcomes Subcommittee. This is the last in a series of four articles that highlight how technology is driving the cost of medical practice. To view the first three articles in this series, visit www.neurosurgery.org/aans/bulletin/summary.html.