The "logical plan" for patient care includes sparing use of medications, providingpain relief, and quick resumption of work.
ABSTRACT: Shared decision making for guiding medical/surgical management of patients with low back pain (LBP) involves concerns about regional variation in health care, patient autonomy versus
physician paternalism, and the degree of patient involvement. The evidence base about chronic LBP is addressed in several questions (eg, how common is LBP? what treatment is appropriate?). The "logical plan" for patient care includes sparing use of medications, providing pain relief, and quick resumption of work. Regional variance may be overcome by recognizing that not all patients are best served by fusion, that individual patient and physician factors/preferences need to be weighed carefully, and that providing unbiased presentation of the best available evidence on the patient's therapeutic options facilitates such careful weighing. (J Musculoskel Med. 2008;25:421-427)
Shared decision making as a preferred method to guide the medical/surgical management of patients is a somewhat new concept. At its foundation rests the contemporaneous emergence of the following 3 independent principles of modern health care: (1) that regional variation in health care is bad; (2) that patient autonomy trumps physician paternalism; and (3) that patients are consumers who should be actively involved in deciding what they purchase.
Although each principle may be and has been debated, together they have been accepted as guiding ideals over the past 30 years. Each principle has a unique and interesting history, but their histories before the 1970s are but trickling streams compared with the river formed with their confluence. In this article, we discuss the application of shared decision making principles to spine care, particularly management of patients with low back pain (LBP).
Regional variation and spine care
Concerns about regional variation in health care may be traced to surgical rates of tonsillectomy in the United States in the 1920s, when American Child Health Association researchers monitored 1000 11-year-old New York City public school students. At the start of the study, 61% of the students already had had tonsillectomies. The others were sent to a group of school doctors, and 45% of them were told that they needed a tonsillectomy. The remaining students were redistributed to the school doctors, and 46% of them were referred for tonsillectomy, and so on. In the end, only 65 of the original 1000 children were not recommended for tonsillectomy-if the children eventually saw the "right" doctor, they would be recommended for surgery.
A decade later in England, Glover reported his classic findings that tonsillectomy rates vary widely based on geographic region. Half a century later, Wennberg published his classic findings that such variation is true not only for tonsillectomy but also for multiple interventions and pathologies (including spine care, where operative rates were found to vary dramatically among well-defined small geographic regions) and that such variation appears to have more to do with the number of physicians in the region, their medical/surgical specialties, and the procedures they prefer to perform than with the health of presenting patients.
This "conclusion" was not popular among doctors, and it was alternatively suggested that the variation problem might not lie with the physicians/surgeons but rather with the lack of high-quality evidence about what is the best available treatment for any particular pathology. Hence, the "outcomes movement" and "evidence-based medicine" arose to provide quality evidence for medical/surgical decision making.
The evidence base and spine care
Although spine care practitioners as a group often are characterized as aggressive, they have emerged as leaders in the development of a solid evidence base. Since the mid-1970s, randomized controlled trials (RCTs) studying the efficacy of interventions aimed at LBP have numbered in the thousands. Also, the "per-year" production and quality of such studies continue to rise as a result, not in small measure, of the emphasis placed here by the leading spine societies and in the journal Spine under the guidance of James Weinstein at Dartmouth.
Based on these studies, the evidence base about chronic LBP (not including sciatica, disk herniation, spinal stenosis, and gross deformity) is addressed in the following questions:
• How common is LBP? The point prevalence is about 33%, and the 1-year prevalence may be as high as 73%.
• How do persons with acute (shorter than a few weeks' duration) LBP generally do? The great majority do quite well. Pain resolution often occurs within a few weeks of onset.
The patient's primary care physician should reassure him or her about his prognosis. Bed rest should be limited to at most a day or two, a quick return to reasonable activities should be encouraged (including work with short-term restrictions, if necessary), and short-term symptomatic treatments may be used (eg, low-dose NSAIDs, superficial heat, and short-term care provided by a physical therapist or chiropractor). The old days of best rest, extended time off work, and use of narcotic analgesics and muscle relaxants should be long gone. This approach is not supported by the evidence base as beneficial and it probably is detrimental (especially pertaining to return to work).
• What if the patient is just not getting better over a few weeks or the pain is severe/incapacitating? The primary care physician should conduct a focused history and physical examination and look closely for signs of radiculopathy, spinal stenosis, and underlying pathology (the "red flags" of neurological compression, infection, tumor, and trauma). If suspicion is high that something bad might be going on, lumbar x-ray films should be ordered. If the above are detected, referral to a spine surgeon makes sense. If not, patience, persistence, reassurance, and continued symptomatic interventions are key. Most patients improve within 6 weeks; if the patient does not, referral should be considered.
• After referral, MRI often is ordered. What MRI findings are important? The aim with MRI is to rule out significant neurological compression caused by disk herniation, spinal stenosis, instability, underlying infection, tumor, and trauma, not to rule in degenerative disease. Disk degeneration (loss of height, loss of fluid signal, contained bulging) and facet joint degenerative disease are normal parts of aging-the "gray hair" of the spine-and are found in most persons older than 40 years, the majority of whom have no symptoms. How these findings are associated with LBP currently is unclear.
• What do you mean? I thought degenerative disk disease and facet joint arthritis were the source of pain in patients with persisting LBP (and no significant neurological compression or underlying pathology). As noted above, degenerative findings are the norm in asymptomatic persons-if they were a pathognomic source, shouldn't we all have persistent LBP? Accordingly, other more subtle findings on MRI have been studied (eg, annular tears and bone edema). However, these subtle findings are less common and, again, are not clearly sensitive or specific to those who have LBP compared with those who do not.
• If MRI does not provide a clear pathoanatomical source of pain in patients with LBP, what about using provocative tests? Multiple interventional diagnostic tests have been used over the years to discern the source of pain in these patients. Anesthetic/corticosteroid blocks of the lumbar facet joints and diskography remain popular. However, these tests are not especially sensitive or specific; reliance on these techniques to define the "pain generator" may be misguided, and the results do not appear to correlate with eventual outcomes.
Therefore, the diagnostic summary of persisting LBP is this: The pain appears to be associated with degenerative changes in the lumbar spine, but it is unclear what the specific pathoanatomical pain source is. To date, MRI and provocative tests have not been able to differentiate patients in whom pain will persist from those in whom it will resolve as expected or from those who are asymptomatic. Indeed, the strongest correlations to persistence of pain are not found in these studies but rather in a history of psychological distress, chronic pain elsewhere in the body, or occupational/legal problems-the "yellow flags" of LBP.
• What about treatment? Assume I send the red flag patients to spinal surgeons and the yellow flag patients for psychosocial counseling. This still leaves a fair amount of patients who have degenerative changes in their backs and persisting LBP. I let them know that I do not see anything terrible on their MRI scan, but what do I do to help them get better? The evidence base is large and continues to grow. Consider the following:
NSAIDs. Based on more than 40 RCTs, NSAIDs appear to have moderate effectiveness for short-term symptomatic treatment of patients with acute LBP. Patients with chronic, persisting LBP appear to have a much smaller response that often is minimally detectable. This finding, coupled with the well-documented risks of the long-term complications of NSAID use, suggests that they should be used cautiously, if at all.
Muscle relaxants. Despite their widespread use for chronic LBP, only a handful of placebo-controlled RCTs have been undertaken. Muscle relaxants have shown a statistically significant but clinically marginal effect in reducing pain.
Antidepressants. Older RCTs demonstrated no statistically significant effect on LBP, and a few newer RCTs have shown a small effect. Again, caution is wise given a 20% risk of adverse effects.
Narcotic analgesics. These drugs afford short-term moderate symptomatic relief. However, tolerance, adverse effects, and the potential to negatively influence long-term outcome have led to a considerable pendulum swing away from using these medications for chronic LBP.
SUMMARY: The medications frequently used for chronic LBP have less effect and greater risks than commonly appreciated. Their use should be considered with care and should be short term.
Exercise. The results of more than 20 RCTs suggest that exercise provides a moderate reduction in pain and a moderate increase in function in patients with chronic LBP. They also suggest that the specific type of exercise (eg, McKenzie, usual physical therapy, chiropractic, aerobics, or machine training) is unimportant-all appear to be effective to some degree.
Back schools. Intensive classroom-style education for patients with chronic LBP appears to have a minor benefit for return to work.
Behavioral therapy. Operant conditioning, cognitive treatments, and progressive relaxation have been studied in a handful of RCTs. The results have been conflicting.
Manual therapy. This intervention, often undertaken within chiropractic, appears to offer reasonable short-term relief but limited long-term effect. Because the adverse effects are minimal, however, manual therapy is a reasonable alternative.
Biofeedback. About 10 RCTs have been performed. The evidence suggests no effect on chronic LBP.
Traction. A handful of RCTs have shown no effect of traction on chronic LBP. More recent studies on "decompression machines" (machine-driven traction devices) suggest the same.
Orthoses. There is some evidence to indicate that a lumbar corset provides some mild subjective improvement.
Transcutaneous electrical nerve stimulation (TENS) units. There have been a handful of RCTs on TENS therapy. The results conflict, and no meaningful information has been gathered.
Acupuncture. Again, several RCTs have been undertaken, with conflicting results. Compared with sham, acupuncture appears to have a minimal effect on chronic LBP.
Functional restoration programs. A moderate but consistent response has been noted with the combination of physical therapy and cognitive-behavioral therapy (CBT).
SUMMARY: Exercise, manual therapy, a light corset, and functional restoration programs provide moderate relief for some patients.
Injections and percutaneous therapeutic interventions. They have some effect in patients with radiculopathy. In multiple RCTs of patients with isolated chronic LBP, however, epidural corticosteroid injections, lumbar disk injections, facet joint injections, trigger point injections, radiofrequency ablation techniques, and focal intradiscal electrothermal treatments (IDETs) demonstrated no effect or minimal short-term effect in highly selected patients.
SUMMARY: No breakthroughs have been found here-and these treatments are invasive!
Lumbar fusion surgery. Given that patients with chronic LBP have no significant neurological compression, laminectomy/diskectomy designed to decompress the nerves serves no role in surgical treatment. However, based on the idea that abnormal motion at the level of a degenerative disk is the source of pain, fusion aimed at eliminating the motion has arisen as the surgical procedure of choice.
Although several studies have evaluated the efficacy of fusion in this patient population, 3 high-quality studies have been most influential. In separate studies, Brox and Fairbank demonstrated a moderate improvement in pain and function after fusion. However, the results were similar to those for patients who were randomized to nonoperative CBT rehabilitation programs. Fritzell also found benefits to surgical intervention; surgical outcomes were superior to "usual" nonoperative care, but the study did not have a group undergoing CBT rehabilitation.
Disk replacement surgery. Several RCTs have compared outcomes of fusion and disk replacement surgery. There currently is no evidence to indicate that the outcomes with disk replacement are better than those with fusion; the risks/learning curves are higher, and the costs are higher. Accordingly, at this point there is just theory but no evidence to support the use of disk replacements in the lumbar spine.
SUMMARY: Fusion appears to afford better outcomes than usual care for patients with chronic LBP but similar outcomes to those achieved with intensive CBT. Because fusion is invasive, it should be reserved for patients who have no red flags and no or minimal yellow flags; have intolerable pain; have not succeeded with appropriate therapy; and have single-level, severe degenerative disk disease.
Regional variation and spine care revisited
Based on the fairly well-sized and high-quality evidence base outlined above, the following "logical plan" for the care of a patient with chronic LBP makes sense:
• Medications should be used sparingly and for the short term.
• Exercise, physical therapy, a light corset, and chiropractic may provide some relief.
• At some level,work should resume as soon as possible.
• Reassurance should be given that activities might cause some pain but no harm and, eventually, some benefit.
• Truly resistant cases of LBP should be referred for intensive CBT rehabilitation.
• Fusion might be considered as a last resort for a select group of patients.
However, although this logical plan has been recognized for the past decade or so, the most recent studies clearly demonstrate that regional variation in the care of patients with chronic LBP persists. Wennberg's theory that variation is the result of physician factors rather than an insufficient evidence base appears to be correct. Medication use (especially narcotic analgesics), injections (especially lumbar epidural injections and facet joint blocks), interventions (especially radiofrequency ablations and IDET procedures), and surgery (fusions and disk replacements) are being used and performed at ever-increasing rates despite the evidence, and which treatments the patients receive depends mostly on which physician they happen to see (just like with tonsillectomies in 1920).
Shared decision making revisited
How can we overcome such regional variation? How can we avoid putting our therapeutic personal preference stamp on our patients based solely on a broad diagnostic category or ICD-9 code? The answers appear to lie in recognizing the following:
• That not all of the patients who present with chronic LBP and degenerative disk disease are best served by fusion (or injection, or narcotic analgesics, etc) and accordingly, that individual patient and physician factors/preferences need to be weighed carefully and will have a great impact on patient outcomes.
• That the best way to facilitate such careful weighing is to provide unbiased presentation of the best available evidence on the patient's therapeutic options (personally or via use of accurate, up-to-date educational tools, including DVDs and online resources) and to listen carefully to the patient's expression of desires and expectations. This process may be one-on-one or, given common difficulties and the time commitment required to afford proper patient understanding, shared decision centers may be used. Then, a mutually decided plan of action may be put forth. Because the plan will be individualized, the problem of variation will be lessened. Also, the plan will be directed by the patient. Therefore, the modern ideals of autonomy and consumerism are allowed to trump the paternalistic and potentially profit-driven decisions of the past.