Proof or Consequences: Who Shall Pay for the Evidence in Pain Medicine?: Table 1

Abstract

Interventional pain medicine faces a crisis. In the interests of bringing relief to their patients, pain specialists practice a variety of invasive procedures that interrupt pain pathways in order to stop pain, or modulate physiological processes in order to reduce pain perception. Few of these interventions have been subjected to randomized controlled trials. Meanwhile, in an effort to reign in increasing costs, third-party payors are implementing “evidence-based medicine” (EBM) standards. Unfortunately, the version of EBM that they apply is increasingly demanding multiple, randomized, controlled trials before interventions are recognized and reimbursed. Conversely, when randomized controlled trials are lacking, their absence is becoming accepted as proof of ineffectiveness of a procedure, and reimbursement is being denied. As a result, a conflict has developed. Practitioners unfortunately are faced with patients today whose treatment cannot be ignored while waiting for someone to generate the evidence. Meanwhile, payors cannot afford to recognize and reimburse every intervention when they cannot tell if it amounts to no more than a source of income for practitioners, with no benefit to patients. Preventing resolution of these tensions is an obstacle that has not attracted attention and consideration in debates and treatises on EBM. To test and validate a single intervention requires enormous funding. For example, a double-blind, placebo-controlled trial of radiofrequency neurotomy for neck pain [1] cost some $500,000. The costs of a placebo-controlled trial of intradiscal electrothermal therapy [2] were estimated at over $1,000,000. But single trials are not enough. Pundits do not credit single controlled trials; they require replication studies. Therefore, costs are doubled. Furthermore, various authorities require evidence for their particular populations. Thus, some insurers are interested in workers' compensation patients, whereas the Centers for Medicare and Medicaid Services wants data on Medicare patients. Therefore, costs double again. Table 1 lists many of the contentious procedures in interventional pain medicine that lack randomized prospective studies. Producing a single study of each would result in an estimated conservative cost of $19.5 million. Reproducing these studies could result in a potential total cost of $58.5 million. Agencies such as the National Institutes of Health do not have either the funds or the disposition to fund the research required. Nor is it a solution that industry should share the burden of costs. Agencies who adjudicate evidence consider industry-sponsored research to be suspect and do not accept it as valid evidence [3]. Nor can the burden be shared overseas, because many insurers and agencies such as Current Procedural Terminology (http://www.ama-assn.org/ama/no-index/physician-resources/3882.shtml) currently only accept studies conducted on U.S. patients and published in U.S. journals. The responsibility for funding would seem to fall back to independent medical societies. The social reality is that it is economically impossible to satisfy the ad hoc academic standards demanded by regulatory authorities. Yet, at stake is the welfare of millions of patients. A selection of interventional pain procedures that require randomized controlled trials A selection of interventional pain procedures that require randomized controlled trials A resolution of this situation is possible by adopting a set of principles from jurisprudence that define the burden of proof. For serious criminal matters, the criterion is “beyond all reasonable doubt.” For lesser matters, the standard is “on the balance of probabilities.” For interventional pain medicine, randomized controlled trials are analogous to proof beyond all reasonable doubt. The corresponding device for “on the balance of probabilities” are well-designed, properly conducted observational studies. This is not an unreasonable alternate conceptually, academically, or economically. The crucial requirement is that the observational studies be well designed and properly conducted. Poor studies do not provide evidence, but well-designed observational studies do. A particular virtue of observational studies is that they can refute the efficacy of an intervention, without requiring a control. If an intervention fails to work when properly audited, it clearly does not work, and it does not require an expensive randomized controlled trial to refute it. Therefore, ineffective practices in interventional pain medicine can be eliminated relatively cheaply by conducting good observational studies. However, if observational studies are positive, they provide prima facie evidence of efficacy. Subordinate questions might then be raised, as to why the intervention works, or if it works better than other treatments. These amount to academic or socioeconomic questions, and do not vitally affect the prima facie status of an intervention. Accordingly, we propose a two-sided resolution to the crisis in interventional pain medicine. In consideration of regulatory authorities countenancing observational studies as evidence, proponents of interventions should provide well-designed, properly conducted observational studies. To assist both sides in this enterprise, the American Academy of Pain Medicine and the International Spine Intervention Society have undertaken to develop a set of guidelines that define the necessary components of an observational study, and how they might be assessed. In brief, those components are independent monitoring of the following: 1) pain scores before treatment and at a sufficient number of times after treatment until the outcomes have stabilized; 2) the proportion of patients who achieve reductions of pain greater than the minimal clinically important change, and reductions to zero and other end points; 3) the proportion of patients who genuinely reduce or eliminate the need for other health care; 4) the proportion of patients who achieve clinically meaningful—and personally meaningful—improvements in quality of life measures; and 5) the incidence of side effects, harms, and cessation of treatment. Depending on the objectives of the intervention, other components may be relevant, such as emotional or psychological functioning. Studies that gather this sort of information do not need to be complicated or expensive. They have two cardinal requirements: an assessor, such as a research nurse, independent of the practitioners who perform the intervention; and the willingness of practitioners to expose themselves to independent audit. Adoption of well-designed observational studies is an economically realistic, and achievable, means of bringing evidence to bear on behalf of the patients who suffer pain. Unless this is carried out, under the current academic nihilism that has been applied to interventional pain medicine, soon no patient with pain will be able to get any form of treatment.