Low back pain (LBP) with or without leg pain is an extremely common, potentially disabling, nonfatal public health problem with significant direct and indirect costs. The efficacy of many of the available therapies for this condition is questionable (1,2). Although the causes of this disease are currently not well understood (2), complaints are typically organized into three clinically distinct syndromes: transient, acute, and chronic. Patients whose complaints are considered transient typically report pain of sudden onset, often without an identifiable cause, that subsides within a few days with little or no medical intervention. Patients whose complaints are considered acute typically report either single or multiple episodes of disabling pain that persist for a longer period, usually no more than 3 months; these episodes are also self-limiting. Acute complaints often do not require medical intervention, but symptomatic treatments tend to speed up return to function (3). Some patients with acute LBP with clearly defined intractable radicular signs and symptoms can benefit from surgical treatments (1). For these patients, surgery is generally expected to alleviate pain completely and return the patient to nearly full function. Patients whose complaints are considered chronic typically report pain that persists unabated for more than 3 months, often for years, despite multiple medical, surgical, and psychologic interventions. The International Association for the Study of Pain recognizes chronic pain as any pain that has persisted for longer than 3 months (4). There is a subgroup of patients with chronic LBP who can be characterized by significant behavioral, emotional, and physical comorbidities (5). In the LBP literature, they are often referred as “failed back syndrome” patients (6–8). Patients in this subgroup are often depressed, anxious, and angry; they commonly use analgesic medications inappropriately; and their vocational, recreational, and social activities are severely limited (9). These patients appear to require a multidisciplinary treatment program with predominantly physical and psychologic components (8,10,11).
We identified another large and important subgroup of patients with LBP whose pain persists longer than 3 months, often for as long as several years, but without the behavioral, emotional, and physical comorbidities that characterize the “failed LBP” patients (12,13). We have referred to these patients as patients with “persistent” LBP to distinguish them from patients with acute and chronic “failed LBP”(12,13).
The typical patient with persistent LBP has had pain intermittently for 10 years. The pain is usually well localized, but its severity varies considerably. Most patients with patients with persistent LBP report significant decreases in their vocational, recreational, social, and interpersonal activities because of LBP. A majority report undergoing multiple ineffective treatments including limitation of activity, weight control, transcutaneous electrical nerve stimulation, massage, heat, chiropractic manipulation, and acupuncture. A few patients also report more aggressive ineffective treatment measures such as surgery and intradiscal therapy (13). When these treatments fail, these patients seek additional care from recognized specialists in the evaluation and treatment of spinal disorders; mainly tertiary care neurologic and orthopedic surgeons.
Despite the personal, economic, and social impact of persistent LBP, data on the effectiveness of the treatments prescribed by the specialists are sparse and contradictory. One possible reason is that patients with persistent LBP have not been considered a distinct group and thus have not been studied independently. They have been viewed as either acute or chronic and may have contaminated studies of both the acute and the chronic LBP populations. Another possible reason is the heterogeneity of the patient population with persistent LBP. There is much variability in the medical, psychologic, behavioral, social, and demographic characteristics of patients with persistent LBP, and the variability has a considerable impact on both choice and outcome of treatment (13). A third possible reason is the lack of standardized, multidimensional procedures for evaluating treatment effectiveness. Many studies have used physicians' opinions rather than patients' reports to assess effectiveness of treatments.
In this study, we evaluated the short- and long-term effectiveness of the treatments that are currently prescribed for patients with persistent LBP by tertiary care specialists in spinal disorders. Our objective was not to compare the relative efficacy of the various treatments. Rather, it was to evaluate the effectiveness of the treatments, as they are currently prescribed in actual clinical practice, by using patient-reported outcome measures including pain severity, physical symptoms other than pain, functional disability, psychologic distress, and health care use (14).
MATERIALS AND METHODS
The data reported in this article were gathered as part of the National LBP Study, a multicenter, prospective, longitudinal study of patients referred to a tertiary care neurologic or orthopedic surgeon for evaluation and treatment of persistent LBP (13). One of the aims of the study was to evaluate the short- and long-term effectiveness of the treatments that are currently prescribed for patients with persistent LBP by specialists in spinal disorders. The methods and procedures of the study have been described in detail elsewhere (13), and are briefly summarized below.
Eight university-affiliated tertiary care centers, selected to yield a demographically heterogeneous group of patients, participated in the study. The appropriate human subject authority at each study center independently approved the study protocol. All patients who participated in the study gave informed consent.
Each eligible patient completed a baseline evaluation consisting of questions about pain severity, physical symptoms other than pain, functional status, previous treatments for LBP, and current usage of health care resources. Patients also completed a battery of standardized psychologic tests consisting of: the Eysenck Personality Inventory (15), the Symptoms Checklist 90R [SCL90R](16), the Beck Depression Inventory (BDI) (17), and the Spielberger State-Trait Anxiety Inventory (STAI) (18).
Before the start of the study, participating physicians collectively developed standardized procedures for conducting and recording the results of the study's required musculoskeletal and neurologic examination. At enrollment, every eligible patient underwent the required examination, and the treating physician recorded its results on a standardized study form.
Physicians also collectively developed standardized procedures for evaluating diagnostic imaging studies, and for recording their interpretation of the studies. Physicians were not required to obtain any specific imaging studies, but those who used diagnostic imaging studies in their evaluation were required to report the findings on a standardized study form.
Physicians also developed a list of possible diagnoses and agreed on a standard definition for each. The pretreatment diagnoses postulated for each patient were recorded on standardized study forms.
Follow-up evaluations were scheduled at 3 and 6 months (short term), and at 12 and 24 months (long term). At 3 and 6 months, patients completed an evaluation consisting of a small subset of the questions included in the baseline evaluation. They answered questions about pain, physical symptoms other than pain, and functional disability. At 12 and 24 months, patients completed an evaluation consisting of a large subset of the questions included in the baseline evaluation. They answered questions about pain, physical symptoms, functional status, and use of health care resources. As part of the long-term evaluations, patients also completed three of the four psychologic tests they completed at enrollment: the SCL90R, the BDI, and the STAI.
Enrolled in the study were 2,374 eligible for follow-up. Their demographic and medical characteristics have been described in detail elsewhere (13). Patients were mostly between 30 and 50 years of age (mean = 45.3 years, SD = 12.79). The percentages of male and female patients were 45.8% and 54.2%, respectively. The majority of patients were white (92.2%), 4.4% were African American, and the rest were mostly Hispanic. Eighty-eight percent of the patients reported graduating from high school, and 60% reported attending college. At the time of enrollment, 66.9% of the patients were working either inside or outside their home, but one in five (20.1%) had stopped working because of LBP. Almost half of the patients reported household income greater than $35,000. About one patient in six (16.3%) reported receiving financial compensation from unemployment, social security, or workmen's compensation benefits, and about the same percentage (16.7%) reported involvement in litigation because of LBP.
The average patient had had well-localized pain intermittently for 10 years, during which time its severity varied considerably. The duration of the current episode of LBP ranged from 1 month to more than 20 years, with a mean of 2.52 years (SD = 4.88) and a median of 9 months.
Nearly all patients had already consulted multiple health care providers and had received a variety of treatments for their complaints, and almost one in five had had either a surgical or an intradiscal procedure before enrollment.
The neurologic and musculoskeletal examinations of the patients were generally unremarkable. Nonspecific abnormalities, such as muscle spasm, tenderness, and trigger points were common, but motor weakness, sensory deficits in the lower extremities, and reflex changes in knees and ankles were much less common. Diagnostic imaging studies revealed that most patients had spondylotic abnormalities involving root compression or lumbar instability or both (13).
The patient and his/her physician determined each patient's treatment plan; study personnel were not involved. Study personnel kept a detailed record of all treatments prescribed by the study physician for each eligible patient, for the duration of the study.
At follow-up, patients were asked to report any treatments they received from anyone other than the study physician, and these treatments were categorized and recorded. Patients who reported receiving additional treatments from physicians who were outside the study were asked for a detailed description of the treatment, which was used to determine the nature of the treatment.
Patients were divided into five treatment groups on the basis of 1) the nature of the prescribed treatment, 2) whether the treatment was administered as an initial or as an additional treatment, and 3) whether the treatment was prescribed and performed by a physician who was associated with the study or by a physician who was outside of the study. The resulting five treatment groups were:no-treatment, conservative care, immediate surgical care, delayed surgical care, and outside surgical care.
Patients were assigned to the no-treatment group if they were prescribed no treatment as initial treatment (treating physicians were required to indicate on the treatment form that they were prescribing no treatment), and reported no additional treatment at all follow-up evaluations.
Conservative Care Group
Patients were assigned to the conservative care group if they were prescribed an active physical therapy program such as exercise; a passive physical therapy program such as massage, heat, or transcutaneous electrical nerve stimulation; some form of nonsurgical extradiscal invasive therapy such as nerve blocks, facet denervations, epidural injections, facet blocks, or trigger point injections; or any pharmacologic treatment. All conservative treatments prescribed any time during the course of the study, by a study physician or a physician outside the study, either as an initial or as an additional treatment, were included.
Surgical Care Groups
Patients were assigned to one of the three surgical care groups if they were prescribed and had received a surgical treatment sometime during the course of the study. Surgical care patients were assigned to one of three surgical subgroups based on when they received the surgical treatment and who performed it. Patients were assigned to the immediate surgical care group if a study physician treated them surgically during the initial treatment period (within 91 days from enrollment). Patients were assigned to the delayed surgical care group if a study physician treated them surgically after the initial treatment period for any reason. Patients were assigned to the outside surgical care group if they were treated surgically by a physician outside the study anytime after the initial treatment period.
Patients who received intradiscal therapies any time during the study, including injections of chymopapain, collagenase, or steroids, and percutaneous discectomies, have been excluded from the analyses reported below. Their number (n = 26) is considered to be too small for any meaningful analyses of treatment outcomes.
Table 1 shows the number of patients who were assigned to each treatment group. Three hundred twenty-two patients were prescribed no treatment during the initial treatment period and reported no additional treatments at follow-up.
The majority of patients, 1,441, were prescribed conservative care. Seven of every 10 of these patients were prescribed an active physical therapy program, generally exercise. About one of every four was prescribed a passive physical therapy program such as massage, heat, and transcutaneous electrical nerve stimulation, and only 6% of the patients were prescribed some form of nonsurgical extradiscal invasive therapy such as nerve blocks, facet denervations, epidural injections, facet blocks, or trigger point injections. Fewer than 1% were prescribed either acupuncture or chiropractic manipulation therapy, compared with 43.7% who were prescribed the same treatments before enrollment in the study.
Immediate surgical care was prescribed for 331 patients: 198 of these patients had a discectomy, 139 had a laminotomy, 159 a laminectomy, 137 received a foraminotomy, and 61 required fusion. Thus, most immediate surgical care patients underwent complex procedures rather than simple discectomies. In addition, a small number of these patients (n = 38) were prescribed some form of conservative care before undergoing surgery.
Delayed surgical care was prescribed for 126 patients. Fifty-six of these patients were initially prescribed conservative care, and the rest (n = 70) were initially prescribed no treatment. The majority of the patients in this group (n = 87) received surgery sometime after the 3-month evaluation but before the 6-month evaluation period. An almost equal number of patients received surgery between the 6-month and the 1-year follow-up and the 1- and 2-year follow-up (n = 25 and n = 24, respectively.) Again, many of the patients in this group underwent complex surgical procedures rather than simple discectomies: 14 had a discectomy, 30 had a laminotomy, 61 a laminectomy, 56 a foraminotomy, and 66 required a fusion.
One hundred twenty-eight patients reported receiving outside surgical care. Of these patients, 101 were initially prescribed conservative care by their study physician, and the rest (n = 27) were initially prescribed no treatment. Physicians associated with the study considered only 13 as surgical candidates. A majority of these patients reported receiving their surgery sometime after the 6-month evaluation period but before the 1-year evaluation period (n = 80). The rest (n = 48) reported receiving their surgery sometime between the 1- and 2-year evaluations.
Five composite outcome measures were derived for each patient and for each evaluation period:pain severity, with scores ranging 0 to 10;functional disability, with scores ranging from 0 to 10;psychologic distress, with scores ranging from 0 to 10;physical symptoms, with scores ranging from 0 to 4;and health care use, with scores raging from 0 to 5. The procedures for deriving the composite measures have been described in detail elsewhere (12,14,19,20). We have shown that these measures are highly reliable, valid, and responsive (14,20).
Baseline differences between groups were evaluated using the chi-square test in the case of discrete variables and the t test in the case of continuous variables. Differences between baseline and follow-up values on individual composite outcome measures were evaluated separately for each treatment group by using analysis of variance for repeated measures.
Table 2 shows means and standard deviations of patients' enrollment scores on all five evaluation variables, for all patients as well as for the separate treatment groups. The average patient in the current study had a pain severity score of 5.14 at enrollment, indicating pain of substantial severity, but the standard deviation (1.60) also indicates considerable variability in pain severity scores between patients. The five different treatment groups also show differences in average pain severity at enrollment. The surgical groups had higher mean pain severity scores at enrollment than did the conservative and the no-treatment groups.
Table 2 shows that the average patient in this study had a functional disability score of 5.2, also indicating substantial disability, and again the standard deviation (2.35) indicates a great deal of variability between patients. Patients in the immediate surgical care group reported the highest average level of functional disability, with patients in the outside surgical care group reporting the next highest average. Patients in the delayed surgical care group reported nearly the same average level as did patients in the no-treatment group. Patients in the conservative care group reported the lowest overall level of functional disability.
The average psychologic distress score was 3.76, indicating little or no serious psychologic distress, but again, there was considerable variability in psychologic distress scores between patients. Differences between treatment groups were small, with patients in the delayed surgical group and those in the outside surgical care group reporting slightly higher average levels of psychologic distress.
At enrollment, the average patient reported experiencing two of the four possible symptoms of 1) numbness or tingling in the lower extremities, 2) weakness in the lower extremities, 3) difficulties with bowel or bladder, and 4) trouble falling asleep or being awakened from asleep because of pain most or all of the time. There were substantive differences between patients within and across treatment groups in the number of reported symptoms. On average, patients in the immediate surgical care group and patients in the outside surgical care group reported the highest average number of symptoms, whereas patients in the conservative care group reported the lowest average number of symptoms.
At enrollment, the levels of health care use activities reported by patients in all treatment groups were about the same. The average patient reported engaging in two of five health care use activities to manage his/her pain.
Of the 2,348 patients eligible for follow-up and included in the analyses reported below, 1,487 (63.3%) completed the 3-month short-term follow-up, 1,253 (53.4%) completed the 6-month short-term follow-up, 1,441 (61.4%) completed the 1-year long-term follow-up, and 1,334 (56.8%) completed the 2-year long-term follow-up.
Patients who completed no follow-ups when compared with patients who completed at least one follow-up were more likely to be younger (42 versus 46 years, t = 5.50, df = 2346, p = 0.0001), more likely to be male (52% versus 44%, chi-square = 6.99, df = 1, p =0.008), more likely not to have completed high school (18% versus 11%, chi-square = 16.98, df = 4, p = 0.002), more likely to have a lower household income (33% versus 23% with household income less than $20,000, chi-square = 19.20, df = 4, p = 0.001), more likely to have been out of work because of LBP (26% versus 20%, chi-square = 17.99, df = 4, p = 0.001), and more likely to have been involved in litigation (21% versus 16%, chi-square = 7.223, df = 1, p = 0.007). However, patients who completed no follow-ups did not show important differences on the evaluation variables from patients who completed at least one follow-up. Enrollment scores for pain severity, functional disability, psychologic status, physical symptoms, and health care use for both groups were practically identical.
Outcomes for Patients in the No-Treatment Group
Table 3a shows the means and standard deviations for the five composite measures for patients in the no-treatment group. Average pain severity scores for these patients decreased moderately over the course of the study. There was a substantial decrease between enrollment and the 3-month follow-up evaluation, followed by a significant but much smaller decrease between the 3- and 6-month follow-up evaluation. Minimal changes occurred after the 6-month evaluation. The table also shows little or no systematic change in average functional disability scores during the duration of the study. A slight, but not significant, increase in average psychologic distress scores from enrollment to 1 year was observed, with essentially no change from the 1-year to 2-year follow-up evaluation. Patients reported slightly fewer symptoms at 3 months than they did at enrollment, but the number of reported symptoms changed little thereafter. Finally, patients in this group reported fewer health care use activities at the 1- and 2-year follow-up evaluations than they did at enrollment.
Outcomes for Patients in the Conservative Care Group
Table 3b gives the same information for patients in the conservative care group. Average pain severity scores for these patients decreased moderately during the course of the study. The decrease proceeded in small steps, with the largest step occurring between enrollment and the 3-month follow-up evaluation. Patients reported considerable functional disability at enrollment, with average scores remaining essentially constant after a moderate improvement between enrollment and the 3-month follow-up evaluation. This group showed little sign of serious psychologic distress throughout the study, with average scores changing little from enrollment through all follow-up evaluations. There was also little or no change in the average number of symptoms reported except for a small drop that occurred between enrollment and the 3-month follow-up evaluation. Health care use decreased considerably from enrollment to the 1-year follow-up evaluation and remained constant thereafter.
Outcome Status for Patients in the Immediate Surgical Care Group
Table 3c shows the means and standard deviations for the five composite measures for patients in the immediate surgical care group. Patients in this group improved greatly on all five measures. Average pain severity scores decreased from a high of 5.56 at enrollment to a low of 2.75 at the 2-year follow-up evaluation. Most of the decrease was observed within 3 months after surgery. Average pain scores remained relatively unchanged after the initial decrease. Patients in this group also reported considerable improvement in functional disability at 3 months and again at 6 months after treatment. The improvement was maintained with little change at 1 and 2 years after treatment. Average psychologic distress scores of patients in this group decreased from enrollment to the first-year follow-up and then remained constant at a low level thereafter. The patients in this group also reported notably fewer symptoms after treatment than they did before treatment. Again, the change occurred mostly within the first 3 months after treatment, with the average number of reported symptoms remaining remarkably constant thereafter. Health care use activities decreased dramatically between enrollment and the 1-year follow-up evaluation, and remained at about the same level thereafter.
Outcome Status for Patients in the Delayed Surgical Care Group
Table 3d shows the means and standard deviations for the five composite measures for patients in the delayed surgery group. Most of the patients in the group underwent surgery during the period between the 3-month and the 6-month follow-up evaluations. The typical patient showed little or no change in pain severity between enrollment and the 3-month follow-up evaluation. At the 6-month follow-up evaluation, the typical patient showed a moderate improvement in pain severity, followed by an additional but smaller improvement at 1 year, and little or no change thereafter. Average functional disability scores showed a slight, nonsignificant increase during the 6-month period after enrollment, but they decreased moderately between the 6-month and 1-year follow-up evaluations, remaining relatively unchanged thereafter. There was a slight but insignificant improvement in the psychologic distress score of the average patient in this group between enrollment and the 1-year follow-up evaluation, with little or no change thereafter. The average number of symptoms reported by patients in this group remained relatively constant during the 6-month period after enrollment, decreased moderately between the 6-month and 1-year follow-up evaluation, and remained relatively constant thereafter. There was also little or no change in the average number of health care use activities between enrollment and 1 year, followed by a moderate decrease between the 1-year and the 2-year follow-up evaluations.
Outcome Status for Patients in the Outside Surgical Care Group
Table 3e shows means and standard deviations for all five composite measures for patients in the outside surgical care group. Most of the patients in this group (about two in three) underwent surgery during the period between the 6-month and the 1-year evaluation. A substantial number, however (about one in three) underwent surgery during the period between the 1-year and 2-year evaluations. Overall, patients in this group showed little or no improvement on any composite measure except pain severity. There was a gradual and relatively small decrease in average pain severity between enrollment and the 2-year follow-up evaluation. Average functional disability scores remained high and relatively unchanged throughout the study. Changes in average psychologic distress scores were small and insignificant. The average number of symptoms reported by patients in this group also remained relatively unchanged throughout the duration of the study. There was a small increase in average health care use scores between enrollment and the 1-year follow-up, with little or no change thereafter.
Differences Between Treatment Groups on Outcome Variables
Figure 1 shows average pain severity scores for each of the five treatment groups at each of the five evaluation periods. The figure shows that average pain severity scores for patients in the no-treatment group and those in the conservative care group were practically the same. Patients in the surgical care groups showed higher levels of pain severity at enrollment than patients in the nonsurgical care groups. However, patients in the immediate surgical care group showed considerably greater improvement in pain severity than did patients in all other treatments groups; their average pain severity scores for all follow-ups were markedly lower than those of patients in the non–surgically treated groups. Patients in the delayed surgical care group, however, did not show the same level of improvement; their average pain scores decreased to about the same level as the non–surgically treated patients. Patients in the outside surgical care group consistently reported higher average pain severity scores than patients in all other treatment groups. The figure also shows much more variability in average pain severity scores between treatment groups at the 2-year evaluation as compared with enrollment.
A similar but less compelling pattern is apparent for the average functional disability scores displayed in Figure 2. At enrollment, non–surgically treated patients had lower average functional disability scores than did the surgically treated patients. Patients in the no-treatment group had higher average functional disability scores than patients in the conservative care group throughout the study. Patients in the immediate surgical care group again showed most improvement; by the 2-year evaluation, their average functional disability scores were lower than those of the no-treatment group and as low as those of patients in the conservative care group. Improvements for patients in the delayed surgical care group did not appear until the 1-year follow-up. At the 2-year evaluation their average score was approaching those of the immediate surgical group and the conservative care group. Patients in the outside surgical care group had consistently higher levels of functional disability than did patients in all other treatment groups. Again, there was greater variability in average functional disability scores at the 2-year evaluation when compared with enrollment.
Figure 3 shows average psychologic distress scores for each of the five treatment groups at enrollment and at 1-year and 2-year follow-up evaluations. Even though there were only small changes in average psychologic distress scores from enrollment to follow-up, the group differences are similar to those observed for pain severity and functional disability. Patients in the nonsurgical care groups had, on average, lower psychologic distress scores than did patients in the surgical treatment groups, with patients in the conservative care group having a lower average score than patients in the no-treatment group. At the 2-year evaluation, patients in the immediate surgical care group showed the lowest average psychologic distress score, whereas patients in the outside surgical care group showed the highest. Average scores for patients in the delayed surgical care group again fell between those in the no-treatment group and those in the conservative care group. Variability in average psychologic distress scores was higher at the 2-year evaluation than at enrollment.
Figures 4 and 5 show information equivalent to that of Figures 1, 2, and 3 for the symptoms and health care use measures, respectively. For both measures, patients in the surgically treated groups had higher average scores at enrollment than did patients in the non–surgically treated groups, with patients in the conservative care group being slightly better off than patients in the no-treatment group. Also for both variables, patients in the immediate surgical care group showed the greatest improvement. Their average scores at 2 years were as low as or lower than those of patients in all other groups. Patients in the outside surgical care group showed the least amount of improvement. Their average scores on both measures were consistently higher than those of patients in all other treatment groups. Again, average scores for the five treatment groups were more diverse at the 2-year follow-up evaluation than they were at enrollment.
Patients with LBP who seek treatment from recognized specialists in spinal disorders appear to be different in several important ways from patients who have traditionally been classified as patients with acute or chronic LBP. The medical, psychologic, sociodemographic, and behavioral characteristics of this clinically important group of patients, whom we have called patients with persistent LBP, together with some of the clinically relevant ways in which we believe these patients differ from patients with acute and chronic LBP, have been described elsewhere (12,13).
Most patients with persistent LBP in this study were prescribed some form of conservative care including active or passive physical therapy, or nonsurgical invasive extradiscal therapy. A smaller, but still substantial number of patients with persistent LBP underwent surgery. Most operations were complex procedures involving discectomy, laminectomy, foraminotomy, and/or fusion and not simple discectomy alone.
Initial status on the evaluation variables as well as change in status after treatment differed substantially among three subgroups of surgically treated patients. Most of the surgical interventions were prescribed and performed by study physicians and were carried out within 91 days from initial evaluation. The patients who underwent these operations formed the immediate surgical intervention subgroup. The delayed surgical subgroup comprised patients for whom surgery was performed later, either because of nonmedical reasons such as scheduling, waiting for second opinion, and waiting for insurance approval, or because the patient's medical condition had changed after the initial evaluation. For this subgroup, the surgery was usually preceded by a trial of physical therapy. There were no differences between patients for whom surgery was delayed for nonmedical reasons and those for whom surgery occurred later because of change in medical condition. Thus, we treated them as a single delayed surgery subgroup. A third subgroup of surgical patients consisted of patients who were advised not to have surgery by the study physicians, but were later evaluated and treated surgically by physicians who were not associated with the study. These patients appeared to be clearly different on the evaluation variables from the other surgically treated patients. Therefore, we evaluated the immediate, delayed, and outside surgical interventions separately.
The data revealed that at the 2-year follow-up, the typical patient of the no-treatment group had improved slightly in terms of pain severity and health care use, with little or no improvement in the functional status, physical symptoms, and psychologic status outcome variables. Spontaneous improvement was typically minimal, and for the duration of the study, the average patient showed as much physical, functional, and psychologic disability as he/she did at enrollment.
The majority of the patients (61.4%) in the current study were prescribed some form of conservative care. Of these, 70% were prescribed active physical therapy, 24% passive physical therapy, 6% nonsurgical extradiscal procedures, and a very small number (less than 1%) were prescribed nontraditional treatment such as acupuncture or chiropractic manipulation. Outcomes did not differ noticeably among these forms of conservative care. The average patient reported small improvements in pain severity, functional disability, physical symptoms, and health care use, with no change in psychologic status. The reported improvements occurred within the first 3 months after enrollment, with essentially no change thereafter. The course of improvement closely paralleled that observed for the no-treatment patients, although the average status of the conservative care patients was slightly better on all outcome measures from enrollment through the duration of the study.
The number of patients who were initially treated surgically was surprisingly low at 14.1%. These surgical procedures were carried out without mortality or serious morbidity. The effect of surgery on the status of patients on the outcome measures was dramatic. The average patient in the immediate surgical care group showed substantive improvement on all of the outcome measures. The observed improvements were evident shortly after treatment and were maintained for the duration of the study. Surgical treatments for persistent LBP appear to be highly effective for the typical patient selected for surgery by the specialists participating in the study. During the course of the study, 86.6% of the patients reported a decrease in the severity of their pain and 75.7% reported an improvement in their functional status.
There was no difference in outcome between delayed surgery for nonmedical reasons, such as scheduling, waiting for a second opinion, or waiting for insurance approval, and late surgery because of change in the patients' conditions. Both yielded outcomes that were less dramatic than those observed for immediate surgery, but substantially better than those observed in the conservative care or the no-treatment group.
The patients who were treated surgically by physicians outside of the study did not improve over time. Their status on the outcome variables throughout the 2-year evaluation period suggests that these patients had characteristics commonly associated with the chronic “failed back syndrome.” Their pain on enrollment was severe and remained so throughout the evaluation period. They also remained functionally impaired, psychologically distressed, and continued the frequent use of health care resources.
It is generally believed that patients who are treated surgically tend to experience much greater improvement than do patients who are treated nonsurgically in the short term, but this difference in degree of improvement tends to diminish in the long term. This belief is based on the study reported by Weber (21), who found a statistically significant difference in outcome between surgically and non–surgically treated patients 1 year after treatment, with the difference in outcome of the two groups no longer statistically significant after 4 years. In the current study, we found that the five treatment groups were considerably more different from each other at the 2-year follow-up than they were before treatment. Although our results can neither confirm nor refute Weber's findings, they suggest that the differences in improvement between treatment groups increased throughout the first year and remained the same after 2 years.
Implications for Clinical Practice
The study reported here was a prospective, observational effectiveness study of patients who were treated for persistent LBP by surgeons who are recognized specialists in spinal disorders. Because the study was not a randomized clinical trial, we cannot attribute the observed differences in outcome between the treatment groups to differences in their treatments. We believe, however, that our results are generalizable to patients who are typically referred to orthopedic and neurologic surgical specialists for diagnosis and treatment of persistent LBP. The patients enrolled in the current study are representative of that population, as judged by demographic characteristics, physical and radiographic findings, and complaints. The physicians who contributed patients to this study also are representative of specialists in spinal disorders.
Patients with persistent LBP typically show no spontaneous recovery. The typical patient who is prescribed no treatment of persistent LBP shows no discernible improvement in status either immediately or after an extended period. Patients with persistent LBP are thus quite different from those in the acute LBP group, who often recover spontaneously within a relatively short period from the onset of their pain.
Conservative care measures, as currently prescribed by surgeons who specialize in spinal disorders, do not appear to be any more effective than no treatment of the typical patient with persistent LBP. Currently available conservative care treatments, including alternative nonconventional procedures, result in little more than minimal improvement in the status of patients for whom they are prescribed. These treatments failed to reduce pain severity, alleviate symptoms, restore function, or decrease health care use both before and after the patients were enrolled the study. Although other forms of conservative treatment may lead to better outcomes, the failure of the currently available procedures, as they are now prescribed, to produce changes that are superior to those seen in patients receiving no treatment suggests that there is little value in recommending these measures for patients with persistent LBP.
The outcome of surgery for persistent LBP varied from dramatic for one subgroup of surgical patients to dismal for another subgroup of patients. Patients who were selected by experts immediately for surgical treatment improved substantially on all outcome measures shortly after surgery and maintained their gains for at least 2 years. Those treated surgically by physicians not associated with the study fared less well. We have yet to identify which patient characteristics are predictive of highly successful outcome and which characteristics are associated with less successful outcomes. The relationships between patient characteristics and treatment outcome will be explored in a subsequent report.
There are important conclusions that can be drawn from this evaluation of the effectiveness of current therapies for persistent LBP. Patients with persistent LBP cannot be expected to improve spontaneously. A minority of these patients are candidates for surgery. The typical patient chosen for treatment by expert spinal surgeons will be substantially improved by surgery.
These findings are at variance with accepted medical practice. Some of the discrepancies are because of a failure to distinguish the persistent LBP syndrome from the acute and chronic “failed” LBP syndromes. The conservative treatments now viewed as required practice do not, on average, seem to influence complaints or function. Surgery prescribed and practiced by experts is highly successful, although used by them on a minority of their patients with persistent LBP. The results of surgery in patients who were not considered surgical candidates by these experts was abysmal, a finding with important implications.
The authors thank Ms. Jane Eisinger and Ms. Bertina C. Wilson for their assistance in the preparation of the manuscript. This study was supported by a grant (NS 21718) from the National Institutes of Health, D. M. Long, Principal Investigator.