The treatment lottery of chronic back pain? A case series at a multidisciplinary pain centre

Introduction

Defined as pain exceeding three months [1], chronic pain is associated with several somatic and psychiatric comorbidities, which in turn affect pain experience and complicate treatment [2], [3], [4], [5]. Behaviors such as fear-avoidance, catastrophizing, and low self-efficacy, as well as work-related factors such as poor job satisfaction, are indicative of worse prognosis in chronic pain patients [2, 6], [7], [8]. In both Europe and United States, approximately 20% of the adult population is affected by chronic pain [9, 10]. Moreover, 40% of European chronic pain sufferers were unsatisfied with the treatment they were receiving [10]. Low back pain and neck pain are ranked 1 and 4, respectively, among the leading causes of years lived with disability globally [11]. Chronic, disabling, low back pain is more prevalent in people with low income and short education [2], indicating the need for a biopsychosocial approach [12, 13].

The Pain and Rehabilitation Centre (PRC), Linköping, is an academic, tertiary, specialist pain center for treating patients with analgesics [14], invasive interventions such as for instance spinal cord stimulation or non-neurodestructive pulsed radiofrequency (RF) treatment [15], or interdisciplinary pain rehabilitation [16], as appropriate. When treating chronic (back) pain, analgesics alone rarely lead to adequate pain relief [17], [18], [19]. The only group of analgesics currently recommended by the English National Institute for Health and Care Excellence (NICE) is non-steroidal anti-inflammatory drugs (NSAIDs) [20].

The existence of a “facet syndrome”, that would in theory benefit from RF treatment, is generally accepted [21]. However, the Mint study showed that RF treatment had no clinically important, additional effect to physiotherapy for pain originating from the facet joints, sacroiliac joints, or intervertebral discs [22]. A smaller randomized controlled trial, comparing RF treatment to placebo-RF for facet joint pain, showed similar results [23]. RF treatment is currently recommended by NICE for certain treatment-refractory back pain conditions, only after a positive diagnostic test block [20], although the predictive value of such a test block is poor [15, 23].

The interdisciplinary pain rehabilitation programs offered at PRC are intense group-based treatments that include pain education, exercise therapy, cognitive behavioral therapy (CBT), and work-related interventions, as described by Ringqvist et al. [24]. Such programs show moderate long-term effect sizes for musculoskeletal conditions on pain intensity, interference in daily life, and overall health, and small effect sizes for many other outcome measures [24]. Evidence suggests that patients with worse self-reported disability and suffering yield better outcomes of interdisciplinary pain rehabilitation [3, 24].

For patients with “unspecific” chronic back pain without a clear component of neuropathic pain, there are currently no guidelines at PRC as to which patients should be offered which treatment. Rather, it is up to the individual physician to choose the most suitable treatment, also considering the patient’s wishes. The present study aimed to provide an overview of which treatment chronic back pain patients are being provided in clinical practice at this tertiary pain center. More precisely, this real-life study aimed at answering the following research questions:

1. What is the distribution of the three main forms of treatment available for chronic back pain in clinical practice?

2. Do the three treatment groups differ, with regards to

   1. Patient demographics?

   2. Pain characteristics?

   3. Patient-reported outcome measures?

   4. Medication use?

   5. Previous treatment?

   6. Organisational factors?

Methods

Study population

As of today, the Swedish version of International Classification of Diseases version 10 is used in Sweden (ICD-10-SE). All patients with a primary chronic back pain diagnosis, defined as ICD-10-SE-diagnosis M53.0–M54.9 but excluding M54.1 and M54.3, were extracted from the new visits to the PRC, University Hospital, Linköping, Sweden, in 2015, provided that the patient had answered the Swedish Quality Registry for Pain Rehabilitation (SQRP) questionnaires. Two patients were excluded during the medical record review, one due to many blocked notes and the other because all visits and treatment focused on pain other than back pain. A third patient was excluded after the analysis to determine treatment groups, as the initial treatment strategy was impossible to categorize (Figure 1).

Figure 1:

Flowchart of referral requests leading to analysis of 103 patients. SQRP, Swedish Quality Registry for pain Rehabilitation. Diagnoses do not include M54.1 and M54.3.

Results

Summary of main findings of Table 2

For most of the 55 variables listed in Table 2, omnibus testing revealed no differences between groups. For instance pain intensity, psychosocial or psychological distress, or medication did not differ between groups. However, we found that 5 of the 55 variables were statistically significant at 0.05 level, and 4 out of 5 variables remained significant at FDR 25%. The 4 variables were age, DaysNotWork, PainDurPer, and PrevInvasive:

1. Age: As shown in Figure 2, the rehabilitation group was younger than both the invasive intervention group (p=0.008, Bonferroni-corrected) and the no treatment group (p=0.002, Bonferroni-corrected), with large effect sizes (d=1.3 and d=1.2, respectively).

2. DaysNotWork: As shown in Figure 3, the rehabilitation group reported much fewer days since last in work or studies, compared to the invasive intervention group (p=0.025, Bonferroni-corrected), with a large effect size (d=1.3).

3. PainDurPer: As shown in Figure 4, the no treatment group had been in persistent pain for much longer than the analgesics group (p=0.029, Bonferroni-corrected), with a medium effect size (d=0.6).

4. PrevInvasive: As shown in Figure 5, the invasive intervention group had been treated much more often with an invasive intervention previously than all 3 other groups (Bonferroni-corrected p-values were 0.001, <0.001 and 0.006 compared to the no treatment, rehabilitation, and analgesics groups, respectively).

Figure 2:

Age of the patients in the four treatment groups. After Bonferroni correction, p-values between the no treatment group and the rehabilitation group, and between the rehabilitation group and invasive intervention group, remained significant (0.002 and 0.008, respectively).

Figure 3:

Number of days out of work or studies in the four treatment groups. After Bonferroni correction, the p-value between the rehabilitation group and the invasive intervention group remained significant (0.025).

Figure 4:

Duration of persistent pain (PainDurPer) in the four treatment groups, measured in days. After Bonferroni-correction, the p-value between the no treatment group and the analgesics group remained significant (0.029).

Figure 5:

Percentages of patients previously treated with an invasive intervention. After Bonferroni-correction, the p-values for comparisons between the invasive intervention group and the other 3 groups remained highly significant, see text.

Discussion

This real-life study illustrates the wide range of treatment strategies available for chronic back pain [17]. Overall, the rehabilitation group was younger, had fewer number of days out of work or studies, and had shorter duration of persistent pain – factors that can obviously be inter-related, at least to some extent. Most variables listed in Table 2, however, did not differ between groups. Moreover, organizational factors such as for instance the specialty of the physician were associated with choice of treatment (Table 3). The question therefore arises: Was the choice of chronic back pain treatment essentially a “lottery” for the individual patient? In other words, was the treatment of chronic back pain more a reflection of where and by whom the patient was assessed, rather than the condition itself? Even if, due to the nature of the present study, it is not possible to give a definite answer, it is nonetheless important to reflect upon these important questions.

Each referral request was assessed beforehand by a senior physician, and we therefore think that there was substantial selection bias involved. Patients fitting general principles for rehabilitation at PRC (e.g. fewer days since last in work, and indirectly hence lower age) would to a greater extent be assigned to a rehabilitation physician, whereas other patients (and especially those who had previously been treated by invasive methods) would to a greater extent be assigned to an anesthesiologist. Hence, the statistically significant group differences in Tables 2 and 3 probably reflect (at least in part) a selection bias. However, there is a clear limit to the amount of information a referral letter can contain, and it is difficult to believe that especially the clear differences found in Table 3 would solely be the result of a well-informed pre-visit selection based on medical facts found in the referral letter.

Altogether, we do think that it is fair to ask the “lottery” question, and we also think that it is sensible to postulate that the situation at PRC is a reflection of the treatment of chronic back pain in general, i.e., that patients are often treated more according to the individual physician’s “armamentarium” and local treatment availability rather than according to objective evidence-based criteria, amounting to a gap between evidence and practice [17, 35, 36]. Simply put, even in a large multidisciplinary pain center with a wide range of possible treatment options, it seems that physicians tend to use the tools they are themselves accustomed to from their previous professional experience. As in pain medicine in general, there is a need for more precision medicine in the context of chronic back pain. Even if this vision is only partially possible to attain today (more research is indeed needed concerning the mechanisms of chronic back pain and how to assess them with high validity and reliability), it is still very important that pain centers create a comprehensive and structured way of caring for chronic back pain patients – not least in order to avoid what we have described as a “lottery”.

The evidence for RF treatment is poor [15, 22, 23], and one of the research questions defined by UK-based NICE [37] is: “What is the clinical and cost effectiveness of radiofrequency denervation for chronic low back pain in the long term?” According to NICE clinical recommendations [37], RF treatment can be considered for chronic low back pain patients when:

1. non-surgical treatment has not worked for them and

2. the main source of pain is thought to come from structures supplied by the medial branch nerve and

3. they have moderate or severe levels of localised back pain (rated as 5 or more on a visual analogue scale, or equivalent) at the time of referral.

NICE also stresses the importance to only treat patients with RF after a positive response to a diagnostic medial branch block [37]. Hence, if one views the problem from a primary care perspective, RF treatment should be viewed more or less as a last resort. But if one views the problem from a pain center point of view, the judicious use of rehabilitation vs. RF treatment is a more complicated question. For instance, should RF in that setting really be viewed as a last resort, or could one conceptually view diagnostic medial branch blocks for facet joint pain [21] as an integral part of the diagnostic workflow? Is it possible to integrate the best from the world of the anesthesiologist with the best from the world of the rehabilitation physician? As already mentioned, one option could perhaps be to use diagnostic medial branch blocks as one tool in a multimodal assessment of the patient. Once again, what is needed is a “reunion” of the two worlds of the anesthesiologist and the rehabilitation physician in both assessment and treatment. Concerning the analgesic treatment group, it is important to stress that the evidence for analgesics in chronic back pain is generally low, not least concerning opioids.

It may seem strange and even perhaps ethically worrying that there was a “no treatment” group. However, it is important to understand what this means. In this context, “no treatment” almost always meant that the pain physician gave the referring physician pharmacological advice, or advice about unimodal rehabilitative strategies available in primary care – such as physiotherapy or contact with a primary care psychologist. In other cases, the patient was referred to another specialist such as neurologist, addiction specialist, etc. In a few cases, advice to the referring physician even focused on the importance of tackling obesity issues. The latter shows that the physicians generally took a holistic approach. All in all, “no treatment” should not be understood as physician passivity or dismissiveness.

A few additional issues are worth discussing. Although this was not statistically significant by omnibus testing, there was a tendency for higher educated patients to rarely be part of the no treatment group (Table 2). Our study is relatively small, implying low statistical power to detect real differences, but it would indeed be ethically worrying if it turned out to be the case that patients with lower education level are treated at a multidisciplinary tertiary pain center to a lesser extent than others just because of their lower education level. There are, however, several possible confounders, one being that education level is associated with occupation, which in turn often affects back pain aetiology and therefore treatment options [2]. On the same note, it is reassuring that there does not seem to be any discrimination based on gender or country of birth but once again, the relatively low power of this study is a caveat.

Although the hypothesized treatment classification proved useful in this study, the medical record review itself can at best be regarded as semi-objective. Medical record reviews, while important tools for gaining an overview of current medical practice and patient safety [38], have questionable inter-rater reliability [39]. The internal validity of the medical records review also depends upon correct and comprehensive reporting by each physician. The external validity of this study is also limited, the study population coming from a single multidisciplinary pain center (tertiary care) at a university hospital, implying a selected group of the most complicated pain patients. None of the results of the present study can therefore easily be generalized to other populations of chronic back pain patients. Another obvious methodological limitation is the cross-sectional and observational nature of the study. Correlation does not imply causation, and one should therefore be cautious when interpreting the results of this study.

To conclude, our findings study are congruent with the described need for a “redesign of clinical pathways” [17] for chronic back pain patients. This is clinically highly relevant, not least because of the fact that back pain is one of the leading causes of years lived with disability globally [11]. Treatment guidelines are often based on the results of randomized controlled trials (RCTs), resulting in a gap between trial efficacy and real-world effectiveness [40]. This study provides insight into factors affecting choice of treatment for patients with chronic back pain at a tertiary pain clinic. Better implementation of existent evidence-based guidelines as well as further “precision medicine” research are needed for this very large group of patients, the aim being a personalized assessment that takes into consideration relevant biomedical and psychosocial factors (i.e., the biopsychosocial model), leading to a judicious use of the different tools available to help this large and heterogenous group of patients. The results also highlight the need of better integration of treatments aiming to decrease nociceptive transmission vs. treatments primarily directed at changing the perception and/or consequences of chronic pain.