Introduction

Efficacy and safety have traditionally been the key evidentiary cornerstones for a new health technology to obtain market access. This evidence is a required part of the licensing of a new drug or medical device. However, over the last decade, with increasing pressure on health care budgets there has been a global trend for health care systems to also provide evidence of cost-effectiveness before they will reimburse or cover a new health care technology.^1,2 Indeed, many countries have established agencies such as the National Institute for Health and Clinical Excellence (NICE) in the United Kingdom (UK), with a mandate to undertake economic evaluations of new and emerging treatments and thus determine if they represent good value for money for that health care jurisdiction. Although often focused on drugs, the consequence of economic evaluation for medical devices, and therefore neurostimulation, are even more potentially challenging given their high up-front cost.

The first part of this chapter is an overview of the different types of economic evaluation and guidelines. The second part of this chapter illustrates these methods by reference to the evidence base for the cost-effectiveness of spinal cord stimulation (SCS).

Methods of Economic Evaluation

The term cost-effectiveness has become synonymous with health economic evaluation and has been used (and misused) to depict the extent to which interventions measure up to what can be considered to represent value for money. Strictly speaking, however, cost-effectiveness analysis (CEA) is one of a number of techniques of economic evaluation.

Economic evaluation is the “comparative analysis of [two or more] courses of action in terms of their costs and consequences”³ and is a tool to assist health care policy makers. A decision maker can then decide whether the intervention is worth paying for by comparing its benefits with the benefits foregone (so-called opportunity cost), in paying for it. Or, to put it another way, given that budgets are finite, economic evaluation seeks to maximize outcomes for the resources available (so-called economic efficiency).

There are four main types of full economic evaluation. Each characterizes costs in the same terms (i.e., in monetary units such as dollars). However, each has a different approach to characterizing health outcomes.

In cost-minimization analysis (CMA), the effectiveness of each of the interventions must have been demonstrated to be equal. Thus the analysis is simply a comparison of costs. The least costly option will be preferred.

In cost-benefit analysis (CBA), both costs and benefits are measured in monetary terms. This has the considerable advantage that the decision to proceed or not is simply a case of measuring whether the value of the benefits exceeds the cost. The disadvantage is that it is difficult (and sometimes objectionable) to effectively put a price on a life. Although often used in areas such as transportation policy, CBA is used infrequently in health care.

Cost-effectiveness analysis (CEA) is concerned with costs and health outcomes and describes health outcomes in naturalistic or disease-specific units. Therefore CEA can tell which strategy maximizes a given objective (e.g., which neurostimulation treatment provides the most pain relief with the lowest cost).

Cost-utility analysis (CUA) is a subgroup of CEA. But, whereas health outcome measures in CEA relate to one aspect of a patient’s well-being, health outcome measures in CUA attempt to capture all aspects of well-being in a single composite (utility) value. The quality adjusted life year (QALY) is the typical outcome measure in a CUA. Calculation of QALYs entails first measuring quality of life (utility) on a scale from zero to one (where zero equates to death and one equates to perfect health). The period of time (in years) over which this quality assessment applies is then multiplied by its quality weighting to give the number of QALYs. Given both the comprehensiveness of the QALY and that it allows policy makers to compare value for money across disease areas (e.g., cost-effectiveness of vagal nerve stimulation for refractory epilepsy vs. deep brain stimulation for Parkinson disease or analgesic drug treatments for the management of neuropathic pain syndromes), CUA has been widely regarded as the gold standard by many health economists and policy makers.

Grounded in economic theory, utility measures reflect the preferences of groups of persons for particular treatment outcomes and disease states and combine many different health domains into a single number, weighting the different domains with the values people have for the particular health states. Utilities can be elicited directly, using the standard gamble or time trade-off (TTO) methods, or indirectly, using questionnaire-based measures (such as EuroQol [EQ-5D] or Health Utilities Index [HUI] for which population preference weights have previously been obtained). Indirect preference measurement techniques obtain utilities using questionnaires to elicit individuals’ valuations of multiple attributes, or domains, of their quality of life (e.g., mobility, emotional well-being, and cognitive ability); responses are then converted into utility values using preestablished formulas. An example of the utility of a male patient age 65 years with diabetic neuropathic pain based on completion of the EQ-5D is shown in Box 24-1. The utility index of 0.228 is contrast to an age-sex matched score of 0.78 for an otherwise healthy member of the UK population.⁴

The result of an economic evaluation is a total cost and total outcome of each of the alternative strategies. However, the results are often presented as an incremental cost-effectiveness ratio (ICER), a summary statistic defined as ΔC/ΔE, where ΔC is the incremental cost (i.e., the difference in costs) and ΔE is the incremental effectiveness (i.e., the difference in health outcomes) in the comparison of a strategy of interest vs. a baseline, or reference, strategy (Fig. 24-1). When one strategy produces greater health outcomes than another but at greater cost, threshold (or trade-off) ratios, or societal willingness-to-pay cut-off points, are used to arrive at a choice among strategies. This threshold may differ among countries (e.g., in the United States it is often quoted as $50,000/QALY,⁵ and in the UK it is £20,000 to £30,000/QALY).⁶

Fig. 24-1 Cost-effectiveness plane (or quadrant).

Incremental cost (ΔC) and effectiveness (ΔE) of new health technology vs. an old technology. In the Southeast quadrant the new health technology is more effective and less costly. When this is the case, the new treatment dominates over the old treatment (i.e., the new health technology would be funded and replace the old technology). In the Northwest quadrant the new health technology is less effective and more costly. When this is the case, the old treatment dominates over the old treatment (i.e., the new technology would not be funded, and the old technology would remain). Since new health technologies are often more effective and more expensive than old technologies, the main focus of economic evaluation is the Northeast quadrant. A new technology with an incremental cost-effectiveness ratio (ICER) below the maximum willingness-to-pay (WTP) threshold would be funded, whereas a new technology with an ICER above this threshold would not be funded. In the Southwest quadrant the new technology saves money compared to the old; however, it could be argued that this is at the cost of health. The maximum acceptable ICER is based on the WTP threshold developed for the Northeast quadrant.

Note that cost analysis is an inherently different type of study. It involves the simple assessment of the cost of a health care strategy (e.g., the cost to treat a person with neuropathic pain for 1 year) and in fact is not a true economic evaluation because it does not involve the side-by-side evaluation of competing strategies. For this reason and because health outcomes are left out of consideration, decisions that rely solely on cost analysis can be erroneous: cheap technologies are not necessarily cost-effective.

Guidelines for Conducting Economic Evaluation

A number of international guidelines for conducting economic evaluations have been published, notably Drummond and associates,³ the U.S. Panel on Cost-Effectiveness in Health and Medicine,⁷ and NICE Technology Appraisal Guidance.⁸

A number of steps are needed to carry out an economic evaluation (Box 24-2). The first is to define the question. It is critical to define exactly what is being compared with what, who the patients are, and what treatments each group is receiving. For example, is the neurostimulatory device to be applied to patients who are otherwise refractory to drug treatment and is the device regarded as an add-on therapy rather than a specific replacement for a current therapy? The scope is equally important. At whose costs are we looking? At whose outcomes? The ideal scope is that of society, meaning that costs incurred by the health service, patients, social services, voluntary sector and all others are included in the analysis. In practice it is often very difficult to estimate some of these; thus analyses are frequently restricted to a health care perspective.

Box 24-2 Steps in Conducting an Economic Evaluation

Define the question.

Define the scope.

Specify the method of economic evaluation.

Define costs to be estimated.

Specify analytical model.

Estimate outcomes.

Summarize results.

Analyze sensitivity.

Cost-Effectiveness of Spinal Cord Stimulation

Over the last 20 years a number publications have included some consideration of the costs of SCS. This economic literature has been the subject to three systematic reviews.^9–11 The literature searches undertaken by these previous reviews across a range of databases (Medline, CINAHL, EMBASE, CDSR, DARE, CENTRAL, NHS EED and HTA) were updated (to March 2010) for the purposes of this article. From these searches, five studies that were full economic evaluations of SCS were identified. Two additional economic evaluations currently in publication were also included. The characteristics of these seven studies are summarized in see Table 24-1.^12–18

The studies were undertaken in a variety of patient groups that included failed back surgery syndrome (FBSS), complex regional pain syndrome (CRPS), refractory angina, and chronic limb ischemia. All studies involved a comparison of costs and outcomes of SCS to usual care, the choice of usual care depending on the medical indication being studied (e.g., analgesia management or spinal reoperation for FBSS and percutaneous myocardial laser revascularization for refractory angina). Studies undertook an analysis of costs and outcomes collected within the trial time horizon (trial-based analysis) or were based on decision analytic modeling to extrapolate beyond the trial data. It is very rare for a single trial to measure all necessary costs and outcomes relevant to decision making; therefore modeling techniques are often used to combine data from different sources (e.g., costs from an observational study and long-term intervention outcomes from a registry) and to project likely future values to arrive at a total cost and consequence from each strategy. Commonly used approaches include decision trees and Markov models.¹⁹ For example, the economic evaluation by Taylor and Taylor¹⁴ used modeling to derive an indirect comparison of SCS vs. medical therapy (from two trials, one comparing SCS vs. reoperation and the other comparing reoperation vs. medical therapy), to convert pain relief into quality of life (utility) and, using observational data on the long-term impact of SCS, to extrapolate 6-month trial data over the patient’s lifetime.

The ICERs of these studies generally show that SCS is cost-effective treatment for FBSS and CRPS (i.e., SCS is either less costly and more effective than the comparative therapy, or SCS has an ICER of less than a maximal willing threshold of £30,000/QALY). Based on these analyses, NICE in 2009 recommended to the UK National Health Service that SCS be made available routinely to selected patients with FBSS and CRPS²⁰:

However, the evidence suggested that the cost-effectiveness of SCS for refractory angina and chronic limb ischemia to be less attractive. The ICER for refractory angina is in excess of the £30,000/QALY threshold. The analysis of SCS for critical limb ischemia demonstrates the disadvantage of using CEA: is saved good value for money? Given this evidence NICE concluded²⁰:

The last stage in an economic evaluation is sensitivity analysis. These “what if” analyses test the robustness of the evaluation and thus the degree of confidence that can be placed in the results. The importance of such analysis is illustrated by the following statement from the Dyer and associates’ study assessing the cost-effectiveness of SCS for refractory angina¹⁷:

In other words, after a period of clinical learning and familiarity, SCS was a potentially cost-effective therapy (i.e., <£23,000/QALY).

One-way sensitivity analysis has traditionally been used to assess uncertainty in economic evaluations, one variable at a time being changed through a range of values. The resulting set of ICERs is then examined to assess the influence of this variable. Fig. 24-2 shows the one-way sensitivity analyses for SCS compared to medical therapy for FBSS. For example, although the implantable pulse generator (IPG) longevity was assumed to be 4 years in the base case analysis (shown in fifth row of Table 24-1), across a range of plausible range of IPG longevities the cost-effectiveness of SCS could range from being dominant to medical therapy (assuming an IPG longevity as long as 10 years) to be being dominated by medical therapy (assuming an IPG longevity of only 2 years). However, recent methodological advances have motivated the gradual adoption of more sophisticated approaches to measuring and reporting uncertainty in CEA. In probabilistic sensitivity analysis, all model parameters are varied simultaneously over a large number of draws, or simulated trials called Monte Carlo simulations.²¹ The incremental cost and QALY values arising from 1000 Monte Carlo simulations are shown a cost-effectiveness quadrant in Fig. 24-3. That the majority of these simulations lie in the Southeast and lower part of the Northeast quadrants below the £30,000/QALY is evidence that SCS is cost-effective in this case. The cost-effectiveness acceptability curve (CEAC) shows the probability of a health technology being cost-effective over a range of maximum willingness-to-pay thresholds.²² As shown in Fig. 24-4, taking into account all parameter uncertainty, there is a 93% probability that SCS (compared to medical therapy for FBSS) is cost effective at £30,000/QALY.

Fig. 24-2 One-way sensitivity analysis—spinal cord stimulation vs. medical therapy for failed back surgery syndrome. ICER, Incremental cost-effectiveness ratio; IPG, implantable pulse generator; SCI, spinal cord stimulation.

Fig. 24-3 Sample cost-effectiveness quadrant—spinal cord stimulation vs. medical therapy for failed back surgery syndrome.

Fig. 24-4 Cost-effectiveness acceptability curve—spinal cord stimulation vs. medical therapy for failed back surgery syndrome.

Summary

Evidence of cost-effectiveness is increasingly required to secure coverage and market access in many countries. For medical devices, including neurostimulation, because of high up-front costs of the technology, demonstrating value for money is particularly important to ensure patient access. Therefore clinicians working in the field of neurostimulation need to be able to understand the methods and interpret and critique the findings of economic evaluations. The purpose of this chapter was to provide an introduction to these issues and, in doing so, provide greater familiarity with a set of research approaches used increasingly to investigate the relationship between costs and the health consequences of health-related policy decisions.