The National Pain Strategy,1 which was developed by the Interagency Pain Research Coordinating Committee in response to the Institute of Medicine report Relieving Pain in America (IOM, 2011b), highlighted several challenges to accessing high-quality pain care, said Robert Kerns. These challenges include the fact that pain care is often not evidence based or team based; that it is often limited to pharmacological treatment delivered by a single provider in a single setting; and that it is usually procedure oriented and incentivized for specialty care. Another challenge, said Kerns, arises from the fact that people with pain often have overlapping chronic pain conditions and other medical morbidities, especially mental health and substance use disorders.
To address these challenges, the National Pain Strategy promoted the concept of
integrated, patient-centered, evidence-based, multimodal, interdisciplinary care, with systematic coordination of medical, psychological, and social aspects of care delivered by primary care, mental health, and sometimes specialty care providers.
Various models of care may provide integrated care, said Kerns, including stepped care, risk stratification, matched care, collaborative care, care
1 For more information about the National Pain Strategy, see https://iprcc.nih.gov/National-Pain-Strategy/Overview (accessed February 6, 2019).
management, integrated care in a co-located setting, use of telecare and technology-assisted care, peer-delivered care by informal caregivers, and whole-health approaches that promote the uptake of complementary and integrative health (CIH) approaches.
Kerns mentioned a recent study that reviewed evidence from nine diverse models of multimodal approaches to pain management (Peterson et al., 2018). All these models provided robust evidence of improvements in measures of pain intensity and pain-related functioning. Most importantly, said Kerns, the review cited four key ingredients in effective models of care: (1) decision support to enhance provider education and treatment planning, (2) additional care coordination resources, (3) improved patient education and activation, and (4) increased access to multimodal care. He highlighted two studies that exemplify the high-quality evidence that has been obtained for multimodal pain approaches. The first of these—collaborative care delivered in five primary care clinics at one Department of Veterans Affairs (VA) Medical Center—showed significant improvements compared with usual treatment with an intervention that included “clinician education, patient assessment, education and activation, symptom monitoring, feedback and recommendations, and facilitation of specialty care” (Dobscha et al., 2009). The second, which used a stepped-care approach for affective disorders and musculoskeletal pain included pharmacotherapy, phone-delivered pain self-management, and an intervention like cognitive behavioral therapy (CBT), also showed robust effects, particularly on depression outcomes, but also on pain intensity (Kroenke et al., 2009).
Although these studies provide strong evidence to support the efficacy and effectiveness of multimodal approaches to pain management, Kerns cautioned that concerns remain about access. He advocated paying more attention to the social determinants of pain and pain management. The Centers for Disease Control and Prevention (CDC) recently published a study that showed a higher than average prevalence of chronic pain and high-impact chronic pain among women, those previously but not currently employed, less educated persons, those living in or near poverty or in rural settings, and those with public health insurance, after adjusting for age. Chronic pain was also more prevalent among non-Hispanic whites and veterans after adjusting for age (Dahlhamer et al., 2018). These social determinants are important to keep in mind, said Kerns, because they describe the same groups of people who may be disadvantaged in terms of accessing high-quality pain care.
Kerns added that delivery of integrative approaches may also be hindered by limited coverage for certain modalities. For example, a recent analysis showed that only 3 out of 15 Medicaid plans covered acupuncture or psychological interventions for low back pain (Heyward et al., 2018). However, despite these challenges, Kerns said the VA, a highly integrated health system, has successfully implemented multimodal integrative approaches; although, there is great variance across the system and still room for improvement (Carey et al., 2018). Kerns emphasized that successful implementation of these approaches will be more challenging in private practice and community settings.
Kerns called for more high quality research on the effectiveness of pain interventions, particularly using integrated approaches and reimbursement innovations. More effective strategies are also needed to disseminate and promote uptake of evidence-based findings into routine clinical practice, he said.
Giving patients a voice in choosing the type of care they receive may improve compliance and outcomes; however, David Atkins mentioned the subset of patients who might not naturally gravitate to certain types of therapies, such as CBT, even though such approaches may be best suited to their needs. He asked how studies could be designed to test whether outcomes could be improved by particular strategies to match therapies to subsets of patients based on their preferences or other characteristics. Robert Edwards, a clinical psychologist in the Pain Management Center at Brigham and Women’s Hospital/Harvard Medical School, said he could envision a study using something like cluster randomization testing where one group receives treatment based on their preferences and another group receives treatment based on a risk assessment. The research question to be answered in that study would be whether the extra time and expense of assessment and phenotyping results in better long-term outcomes. Alicia Heapy, associate professor of psychiatry at Yale University School of Medicine and associate director of the Pain Research, Informatics, Multimorbidities, and Education (PRIME) Center of Innovation at the VA Connecticut Healthcare System, added that this model is consistent with the stepped-care approach, discussed below.
Lynn DeBar, senior scientist at the Kaiser Permanente Washington Health Research Institute, commented that among the biggest challenges encountered in a study aimed at bringing interdisciplinary care to patients on chronic opioid therapy were (1) getting patients interested enough in the kinds of treatments being offered, and (2) creating a culture of care
grounded in the biopsychosocial model. She added that the language practitioners use to talk about nonpharmacological therapies is important to ensure that patients believe they are being offered the best treatments possible. The facts, she said, are not nearly as compelling as the powerful experiences people share.
Stepped care has been used for many conditions, purportedly to provide patients with low-risk, low-cost care early on, followed by higher risk and more invasive treatments if symptoms persist, said William Shaw, associate professor and chief in the Division of Occupational and Environmental Medicine at the University of Connecticut Health Center. Unfortunately, he said, stepped care often means starting with easy care such as pharmacological treatments before implementing more difficult approaches such as psychological treatment if the problem continues.
Shaw said the VA developed a stepped-care plan that begins with screening, assessment, and self-management, followed by secondary consultation with specialists in pain management and a third step of multidisciplinary rehabilitation or advanced diagnostics and interventions (Kerns et al., 2011). Julie Fritz, distinguished professor in the Department of Physical Therapy and Athletic Training in the College of Health, University of Utah, added that subsequently, another step was added at the beginning of the pathway that involves self-care. This step was added, according to Benjamin Kligler, when it became apparent that pain management is most successful when people are empowered to take care of their health. Shaw noted that this approach works to limit costs and reduce iatrogenic effects in health care systems such as the VA that have complete control of care. However, for the treatment of acute and subacute back pain, the stepped-care model may make it difficult to start nonpharmacological treatments early, he said. The model presumes that it takes time to establish psychological and other kinds of treatment approaches and ignores the fact that psychosocial risk factors can be assessed in a patient within hours of back pain onset.
A stratified-care model for musculoskeletal conditions may help solve this problem, said Shaw. This approach begins with the new-onset patient completing a one-page self-report questionnaire that assesses the risk of chronicity, disability, or other poor outcomes. Based on this single measure, patients can be assigned to low-, medium-, or high-risk groups that
receive conservative to more aggressive care, respectively. This approach provides early intervention for high-risk cases and may prevent unnecessary treatment, said Shaw.
The best example of the stratified-care model, he said, is the STarT Back trial, in which participants filled out a nine-item questionnaire and then were assigned to the three risk groups. Low-risk participants received advice on appropriate levels of activity, exercise, and self-help. Medium- and high-risk participants were referred for physical therapy and more intensive treatment. Outcomes for all treatment groups improved compared with the control group, said Shaw (Hill et al., 2011). The magic of this approach, he said, was the cost savings that resulted from determining at the outset who needed care and who did not.
The disadvantage of this approach is that it may be difficult to assign people to risk categories with this simple questionnaire. Shaw and colleagues proposed a matched-care model that uses a more detailed questionnaire to assess the nature of risk based on specific types of risk factors, such as mood problems, functional limitations, or workplace issues. They then assign individuals to treatment groups that match their individual problems (Shaw et al., 2013). Shaw acknowledged that this approach, which is still in an experimental phase, requires substantial resources and evidence that it improves outcomes. He suggested that a middle road is needed between specificity related to the nature of the patient’s problems and the feasibility of delivering care within the payer-based systems currently in place. Edwards added that many studies would be needed before clinicians are able to match specific treatments to individual patients.
The importance of first contact was mentioned by Steven George in Chapter 2 and was reinforced by Fritz, one of many researchers who contributed to a recent Lancet study focused on how to reorder the way care is delivered so that first-line treatments offered for conditions such as low back pain are grounded in a biopsychosocial framework that enhances self-management (Foster et al., 2018).
Training providers to behave differently and reforming payment policies are long-term strategies that might accomplish this goal, said Fritz. However, she advocated more strategic ways of reordering care, for example, by sequencing care and adding explicitly a step that precedes primary care and is focused on self-management and self-care. For example,
a pilot study by the National Health Service in the United Kingdom introduced a primary care triage and treatment practitioner—typically a nurse or physiotherapist—as the first contact for patients with spinal and/or radicular pain. This practitioner provided encouragement, reassurance, educational materials, and advice on self-management. Patients with signs of significant compression or inflammatory back pain were referred to primary or specialty care, while others who were nonresponsive to the first line of care or who had physical or psychological risk factors for poor prognosis were referred for evidence-based nonpharmacological treatment (Foster et al., 2018).
In the United States, Fritz said the first provider contacted for low back pain is most often a chiropractor or primary care physician. Data show that receiving care from a nonphysician, reduces the risk of opioids being either initiated or used over time (Kosloff et al., 2013; Weeks and Goertz, 2016; Wheldon et al., 2018). Claims data from her own institution also show that opioid prescription fills and long-term use of opioids are much lower when patients start their journey of care with a nonphysician provider, said Fritz (Fritz et al., 2016).
She acknowledged that selection bias may a play a role in these results, but emphasized the importance of understanding why patients choose a particular type of provider and how to encourage them to choose nonphysician providers more often. Fritz’s research suggests that patients are willing to choose a physical therapist visit as the first of line care if they can be ensured that their health care providers are working together as a team and if their concerns are appropriately addressed.
Nonpharmacological interventions have shown promise in treating patients with complex and high-impact chronic pain, said Edwards. One reason for this is that there are many clearly delineated risk factors for developing and maintaining high-impact chronic pain, including psychosocial risk factors such as catastrophizing. In the context of pain, Edwards defined catastrophizing as a negative cognitive and emotional response characterized by rumination about pain, magnification of the threat value of pain, and feelings of helplessness. Many studies have shown that the highest catastrophizing patients get the least benefit from pharmacological treatment and are most likely to discontinue treatment, he said (e.g., Toth
et al., 2014). Just as important, said Edwards, catastrophizing changes with treatment and shapes treatment-related outcomes. For example, in one study examining the effectiveness of tai chi for reducing pain and disability, about one-third of the effect on pain and two-thirds of the effect on disability were attributed to a reduction in catastrophizing (Hall et al., 2016).
Edwards and colleagues have been studying the neural mechanisms that may contribute to catastrophizing’s effects on pain. They have found that patients who selectively activate brain networks associated with catastrophizing also report the most severe daily pain and have the greatest self-reported catastrophizing (Lee et al., 2018). Catastrophizing also appears to contribute to the maladaptive network connectivity that is observed in patients experiencing chronic pain.
Importantly, catastrophizing is a modifiable risk factor, noted Edwards. CBT, for example, has been shown to reduce both catastrophizing and pain in patients with fibromyalgia (Lazaridou et al., 2017). Some evidence suggests that the benefits of CBT are greatest in patients with high baseline catastrophizing scores, who are the most resistant to pharmacological treatment (Schutze et al., 2018). Recent research indicates that high catastrophizing may be associated with enhanced benefit from certain interventions and reduced benefit from others, which suggests that this phenotype may be useful as a factor for selecting personalized treatments, said Edwards (Schreiber et al., 2018).
In 1991, the National Institutes of Health established the Office of Alternative Medicine, which was renamed the National Center for Complementary and Alternative Medicine in 1998 and renamed again in 2014 as the National Center for Complementary and Integrative Health (NCCIH). The evolution in the name of the center reflects the recognition that although interest in alternative approaches has increased over the past 20 years, there is an even greater interest in integrating them with conventional care, said DeBar.
Complementary approaches include the use of natural products and mind and body practices (NCCIH, 2018). A 2012 survey by the National Health Interview Survey showed that among mind and body practices, yoga, chiropractic and osteopathic manipulation, and meditation were
most popular. Other mind and body approaches include acupuncture, massage, tai chi, and the Alexander technique2 (Clarke et al., 2015). A more recent study showed that these practices continue to grow in popularity, said DeBar. From 2012 to 2017, the number of adults over age 18 using yoga grew from 9.5 to 14.3 percent (Clarke et al., 2018).
The Agency for Healthcare Research and Quality set a somewhat higher bar in a recently published comparative effectiveness review by looking at short-, intermediate-, and long-term treatment effects of these approaches, said DeBar. For low back pain, yoga, manipulation, massage, and acupuncture were shown to improve function and decrease pain intensity at least in the 1- to 6-month range (Skelly et al., 2018). A limitation of this research, according to DeBar, is that many (if not most) people with pain have chronic overlapping pain conditions, yet researchers have tended to examine effectiveness in a siloed manner. One exception she cited was a study conducted by Davis and colleagues (2016), who demonstrated that the symptom cluster of sleep disturbance, pain, anxiety, depression, and low energy and fatigue—the SPADE pentad—correlated with functional outcomes in a stepped-care intervention program.
Complementary and integrative health modalities have also been studied in more partitioned ways, disconnected from conventional care, said DeBar. For example, Cherkin and colleagues conducted a series of rigorous clinical trials that examined the effectiveness, compared to usual care, of acupuncture, massage, yoga, and mindfulness-based stress reduction for the treatment of low back pain (Cherkin et al., 2009, 2011, 2016; Sherman et al., 2011). Interestingly, said DeBar, all these modalities showed similar impacts. She suggested that perhaps these effects derive from non-specific contextual effects of the care experience rather than specific characteristics of the intervention. Which modality is selected may be less important than what is available to patients in a particular area or what they are excited about and will commit and adhere to, said DeBar. There might be ways to sequentially employ both active and passive modalities in clinical trials.
DeBar said there have been some provocative studies indicating that the placebo effect is real and can be powerful both for pharmacological and nonpharmacological therapies. She pointed to a study by Crum and colleagues that demonstrated how in the real world, psychological and so-
2 An educational process aimed at retraining poor movement and posture habits.
cial forces of healing contribute to and underlie the magnitude of a treatment effect (Crum et al., 2017). We should harness those effects, said DeBar, particularly in pragmatic trials and studies of real-world treatments.
Karen Sherman, senior scientific investigator at the Kaiser Permanente Washington Health Research Institute and affiliate professor of epidemiology at the University of Washington School of Public Health, recommended future research priorities for CIH. First, she advocated including more trials in older adults even though their interest in CIH may differ from younger adults. Chronic pain conditions associated with comorbid conditions are relatively more common in older adults, and standard treatments may work less effectively or be associated with worse side effects in seniors. Nonpharmacological approaches may reduce these adverse events, said Sherman. Currently, many randomized controlled trials of chronic pain treatments, including CIH interventions, exclude older adults, especially the oldest old and those who are institutionalized, she said, noting that CIH interventions may need to be modified for these populations. Sherman also recommended:
- Conducting studies on implementing and disseminating various mind–body and self-care therapies, which can increase self-efficacy, but often are unreimbursed and require sustained, consistent practice.
- Determining the optimal combinations of treatments and establishing principles for combining and/or sequencing treatments.
- Exploring the effectiveness of CIH treatments in people who use opioids.
A diverse array of technologies offer the potential to improve access to care by delivering self-management approaches to patients, according to Heapy. These technologies range from text messaging and other mobile apps to virtual reality and artificial intelligence, and include wearable devices, interactive voice response systems, and telemedicine, she said. They have the capacity to collect high-intensity, longitudinal data, which Heapy said has been shown to be valid and reliable and less vulnerable to recall bias than data collected farther from the time the pain occurred. Some of these technologies have adapted theory- and evidence-based interventions in which small to moderate effects have been demonstrated, such as CBT
and mindfulness approaches. Heapy described others as “modular treatments,” where multiple approaches are combined to promote self-monitoring, goal setting, skill acquisition, education, assessment, patient–provider communication, and social support. Superimposed on these diverse technologies are differences in the amount of contact that patients have with clinicians.
Heapy said that while these technologies appear promising, limited quality evidence is available regarding the effectiveness of these interventions relative to in-person care; which technologies are best or which components are most important or effective; and to what extent clinician contact increases engagement and improves outcomes and the necessary frequency of that contact (McGuire et al., 2017). Greater understanding of how these treatments work is also needed, she said, as well as the temporal sequence of changes, causal factors that are related to changes, and common mechanisms that may underlie the effectiveness of different treatments. To improve the quality of evidence, researchers have recommended conducting more randomized controlled trials with fewer single-arm studies, reducing variability across studies to enable comparability, and focusing on interventions with a theoretical basis, said Heapy (McGuire et al., 2017; Sundararaman et al., 2017; Wethington et al., 2018). She added that cost-effectiveness studies are also needed of interventions using these technologies. Heapy advocated including clinical experts and end users in the development of technology-based treatments and adapting treatments for special and underserved populations, including older adults.
Little is known about how to implement technology-assisted treatments for pain either to the patient directly or within a health system as part of routine care, said Heapy. Internet-based interventions for anxiety and depression are the most well developed. A recent study of Internet-based CBT found general agreement among therapists and service managers on the need for technology-assisted interventions to increase access, but skepticism existed from referring providers about the quality of interventions (Folker et al., 2018). There were also concerns about the need for additional therapist training to deliver these treatments. Heapy suggested that hybrid implementation and effectiveness trials might achieve two aims with one trial.
Many questions remain about how to advance the use of technology-assisted treatments and integrate these treatments into existing care, said Heapy. Among these: Who will pay for these treatments? Will treating patients across state lines raise licensure issues? How can risk be managed (e.g., for patients at risk for suicide)? How can patient data confidentiality
be ensured (Hill et al., 2018; McGuire et al., 2017)? Heapy added that there is a need to systematically examine barriers and facilitators to treatment at the patient, provider, and system levels. Education and marketing are also needed to explain to patients how and why nonpharmacological approaches such as CBT may help, and to providers why early referral is important, she said. Heapy added that partnerships will be needed between investigators and commercial entities to address these concerns and develop a consensus, and that integrated health systems like the VA, the military health system, and Kaiser Permanente have an important role to play because they have the infrastructure, electronic health records, and large expert workforces capable of developing, testing, and implementing novel treatments into care.
According to Kurt Kroenke, there is some, but limited, evidence to support combining pharmacological and nonpharmacological modalities to treat pain. He cited several combination trials that focused on pain or pain and depression. A study that compared the combination of the antidepressant nortriptyline plus stress management to either therapy alone or placebo, for example, showed that the combination was marginally more effective than monotherapy (Holroyd et al., 2001). The ESCAPE trial (Evaluation of Stepped Care for Chronic Pain) in veterans of the Iraq and Afghanistan conflicts combined medication with self-management strategies for the first 12 weeks, followed by 12 weeks of CBT. The results showed statistically significant reductions in pain-related disability and pain severity (Bair et al., 2015). Another study conducted by Kroenke and colleagues compared analgesics, mood treatments, and self-management to automated self-management only (see Kroenke et al., 2018, for study design and sample characteristics). Interestingly, they found that about one-third of participants in the usual care arm worsened over the 12-month trial period compared with about one-sixth of participants in the active arm, said Kroenke. This suggests that patients may actually be harmed if they are not offered the right combination of nonpharmacological and optimized analgesia treatments, which he said is often not currently feasible in primary care.
Kroenke noted that it is difficult to unbundle the effects of the different components in a combination trial to assess synergistic effects. Other fac-
tors that make studying combinations of modalities difficult are the numerous medication and nonpharmacological options with only modest benefit and the multiple common pain conditions against which those treatments could be tested. Adding devices to the equation further complicates the situation. However, “imperfect treatments do not justify therapeutic nihilism, and a broad menu of partially effective treatment options maximizes the chance of achieving at least partial amelioration of chronic pain,” said Kroenke in a recent editorial (Kroenke and Cheville, 2017, p. 2366). He suggested that future research on combined modalities should examine patient preferences, optimal sequencing, differential effectiveness for chronic pain versus site- or mechanism-specific pain, long-term effectiveness, and alternative trial designs.
Neuromodulation—the alteration of nerve activity through targeted delivery of a stimulus to a specific neurologic site—represents another nonpharmacological approach to the treatment of pain, said Richard Wilson, associate professor of physical medicine and rehabilitation at Case Western Reserve University and director of the Division of Neurological Rehabilitation at the MetroHealth Rehabilitation Institute of Ohio. The stimulus may be electric current or chemical and can be delivered either invasively or non-invasively, he said, with the goal to reduce pain and improve function and quality of life.
Among invasive approaches, Wilson said spinal cord stimulation is one of the most commonly used methods for pain control. In this modality, electrodes are surgically placed within the spinal canal in close proximity to the spinal cord and then connected to an implantable pulse generator. Dorsal root ganglion stimulation is a newer form in which the electrode is placed near the dorsal root ganglion to deliver more precise stimulation, he said. Other invasive methods he described include peripheral nerve stimulation (PNS), in which electrodes are placed underneath the skin and connected to an implanted pulse generator. This approach, said Wilson, has been used to reduce chronic head pain. A less invasive option involves placement of an electrode underneath the skin and connecting it to an external stimulator, which can provide long-term benefits.
Brain stimulation may also be achieved through both invasive and noninvasive approaches, said Wilson. In transcranial direct current stimu-
lation (tDCS), electrodes are placed on the scalp. A current flowing between two electrodes stimulates the brain tissue underneath. With transcranial magnetic stimulation (TMS), a coil placed outside the skull generates a current within brain tissue that can alter brain activity. Among invasive neuromodulatory approaches, deep brain stimulation requires neurosurgery to plant an electrode within brain tissue to deliver more precise stimulation.
Wilson said there is evidence supporting the use of tDCS and PNS to treat migraine, and for TMS to treat phantom limb pain, neuropathic pain, and spinal cord injury (Boldt et al., 2014; Chen et al., 2015; Corbett et al., 2018; O’Connell et al., 2018; Shirahige et al., 2016). Some evidence also suggests that spinal cord stimulation may be more effective than other treatments for failed back surgery syndrome (Frey et al., 2009).
All these approaches have modest or small effects, said Wilson. He suggested that a multimodal approach using interventions that target different mechanisms or that have different mechanisms of action might provide better pain relief. For example, exercise and neuromodulation reduce pain in different ways. To treat knee osteoarthritis, Wilson hypothesized that exercise may reduce pain by improving biomechanics, altering endogenous pain mechanisms, and possibly promoting neuroplasticity. Concurrent treatment with tDCS, meanwhile, may improve somatosensory function and stimulate the cortex. The hope is that these effects might be additive or synergistic, said Wilson; however, there is little evidence at this point to guide how to use these approaches in practice.
He cited three randomized controlled trials of combined modalities for pain. The first showed that spinal cord stimulation and physical therapy were more effective than either approach alone in the treatment of complex regional pain syndrome (Kemler et al., 2000, 2004). Another trial showed that tDCS plus aerobic exercise was better than either alone for fibromyalgia. A third showed that tDCS plus exercise is better than exercise and sham stimulation for low back pain.
Wilson’s lab has developed a treatment for hemiplegic shoulder pain after stroke that involves placing an electrode between the middle and posterior deltoids of the affected shoulder and delivering electrical stimulation from an external stimulator every 30 seconds for 6 hours per day for 3 weeks. They compared this approach to physical therapy designed to improve range of motion and strength and showed that both groups had reduced pain, with greater improvement in the PNS group (Wilson et al., 2014). As a next step, Wilson and colleagues are now conducting a randomized controlled trial to test combining these two approaches. One group will get the multimodal approach, another will get PNS and sham
physical therapy (physical therapy aimed at something other than shoulder pain), and a third will get sham stimulation and physical therapy. They have added other important outcomes to this trial to help them understand mechanisms and patient variables that influence response (e.g., catastrophizing, fear avoidance, or anatomic abnormalities).
Wilson mentioned one other example of multimodal research: a study comparing mirror therapy compared to tDCS for phantom limb pain (Pinto et al., 2016). Mirror therapy uses mirrors to create the illusion of movement in the phantom limb, which leads to a reorganization of cortical networks. Understanding mechanisms such as this will help investigators learn how to improve these approaches in future trials and understand which combination of treatments will likely yield the greatest effects, said Wilson.
This page intentionally left blank.