The Development of Algorithms for Pain Care Including Neuromodulation Therapies: Introducing the SAFE Principles

Introduction

Algorithms provide a set of general rules or instructions to solve a specific problem in a particular instance. In contrast to mathematical or engineering algorithms, where a rigid workflow provides a precise prediction, medical algorithms for appropriate interventions for a particular disease need more flexibility to accommodate the needs and perspectives of the physician and patient and incorporate psychosocial constraints, efficacy, and cost considerations (both personal and societal) of those interventions or therapies.

As we have seen in the preceding chapters, robust algorithms for providing treatment for chronic pain conditions are especially critical, as chronic pain is among the leading causes of disability in the United States. The Center for Disease Control (CDC) recently communicated that of the 53 million Americans carrying a diagnosis of arthritis, over 22 million report a significant limitation of their function, including 66% who are unable to work or disabled ( ). Therefore, efficiently treating chronic pain conditions such as arthritis and their sequelae is a prerequisite to not only improving the quality of patients’ lives but also the appropriate distribution of resources and reintegration of our patients as participants in society.

Neuromodulation therapies, including implantable drug delivery systems (IDDS) and spinal cord and peripheral nerve stimulator systems, have high variability in how they are offered, often because of their high up-front costs. Additionally, depending on the training or education of individual physicians from disciplines other than pain management, there is often a lack of knowledge of the appropriateness or risks involved with such therapies. So an effective algorithm to implement and choose neuromodulation therapies for appropriate patients must not only include levels of invasiveness of the therapy and assessment of the safety and efficacy of the intended therapy, but also gauge appropriateness of the therapy for a particular patient or populations of patients and the relative costs to the patient and society. We can then contrast what we have learned of this therapy using these tools to evaluate other competing therapies. We must also strive to overcome roadblocks to care, such as patients’ access to capable physicians who are familiar with such emerging technologies and the misconceptions of referring physicians and third-party payers regarding these treatments.

Prior Algorithms to Chronic Pain Treatment

The World Health Organization (WHO) released its first algorithmic approach to managing cancer pain in 1986, including a guide to the stepwise introduction and escalation of analgesic medications, culminating in a recommendation for opiates as a last resort. This “WHO analgesic ladder” has subsequently been updated and extended to include noncancer pain; so for an individual patient, effective analgesics with the least propensity to do harm would be trialed before considering other therapies that may be more costly or have a greater propensity to do harm (more side-effects). Importantly, the WHO ladder does not explicitly incorporate the field of neuromodulation. Furthermore, the relevance of the WHO ladder to the treatment of chronic noncancer pain is likely inadequate, because in contrast to acute or cancer pain, patients with chronic noncancer pain have increasing components of pain that are linked to cognitive or affective influences rather than purely biologic, nociceptive, or peripheral influences ( ).

In Krames proposed a treatment algorithm positioning neuromodulation therapies such as IDDS, spinal cord stimulation (SCS), and peripheral nerve stimulation (PNS) at the end of a continuum for the treatment of chronic pain—a “last-resort” treatment option after conservative measures had failed ( ). The main guiding principle of this approach was “KISS” (keep it sweet and simple), which values noninvasiveness and low up-front costs together with the time-honored dictum of doing the least harm to a patient as possible. According to this serial model, for example, a patient would first use exercise and over-the-counter analgesics plus cognitive and behavioral therapies before trialing adjuvant medications such as tricyclic antidepressants or other neuropathic pain agents, if indicated. If needed, according to this continuum of care, one could consider interventional procedures such as steroid injections or nerve blocks, before considering oral opiates. Finally, if all less invasive and less costly interventions failed, patients would be offered implantable options such as IDDS, PNS, or SCS. By this algorithm or continuum of care, therapies would be offered in serial fashion. Because neuromodulation therapies were still nascent at the time of that publication, however, and clear data on safety and efficacy was lacking, any rational dialogue on appropriateness of this therapy was precluded. shortly thereafter recommended a more flexible approach allowing selection of different therapies at different time frames, including neuromodulation, depending on the clinical situation for a particular patient. Krames et al. in more recent publications have suggested that while such reasoning was previously acceptable, it is now outdated and should be supplanted with a more nuanced and rational algorithm, bringing to bear current evidence on neuromodulation over the last 15 years ( ).

These publications by Krames et al. outlined a novel rational approach to determine when and in what patient we should consider neuromodulation therapies. The new guidelines have been labeled the SAFE (safety, appropriateness, time to fiscal neutrality, and effectiveness) principles for medical algorithmic thinking. By expanding the scope of factors relevant to selection of optimal therapy for patients, this pattern of thinking allows individualization of algorithms that are tailored to each patient by incorporating contemporary evidence at multiple parallel levels from the literature. In many instances, the up-front costs of a specific therapy can be offset by reductions in future costs of care, and may be more effective if offered earlier. With this new perspective, neuromodulation therapies, as we shall see, should routinely be offered earlier in the treatment continuum for chronic pain as we reconfigure the algorithms.

The SAFE Principles

This section outlines the SAFE principles as they apply to creating algorithms for medical care that involves neuromodulation therapies. However, it is important to highlight that the SAFE principles are generic tools of evaluation and may provide a general framework for organizing thinking about any medical therapeutic treatment intervention or algorithm, not just the algorithm of care for chronic pain management. Of note also, the SAFE principles have yet to be validated by studies.