Study Design. A prospective, consecutive, cohort study of patients with acute low back pain classified into subgroups based on examination data and treated with a specific treatment approach.
Objective. To calculate the interrater reliability of a classification system, and to compare initial patient characteristics and outcomes of physical therapy treatment when a classification approach is used.
Summary of Background Data. Classification of patients with low back pain into homogeneous subgroups has been identified as a research priority. Identifying relevant subgroups of patients could improve clinical outcomes and research efficiency.
Methods. Consecutive patients referred to physical therapy for treatment of acute low back pain were evaluated and classified into one of four subgroups (immobilization, mobilization, specific exercise, or traction) before treatment. Physical therapy treatment was based on the patient's classification. The classifications were compared for initial patient characteristics, frequency and duration of physical therapy, and improvement in Oswestry scores.
Results. In this study, 120 patients were evaluated and classified. Analysis of interrater reliability showed a kappa value of 0.56. Differences were found among the classifications for age, initial Oswestry score, history of low back pain, symptom distribution, and average change in Oswestry score with treatment.
Conclusions. Reaching a consensus regarding relevant patient subgroups requires data on the reliability and validity of existing classification systems. Further work is required to validate improvement in treatment outcomes using a classification approach.
The optimal treatment for patients with acute low back pain (LBP) remains largely enigmatic. Randomized clinical trials have failed to find consistent evidence for improved treatment outcomes with many treatment approaches commonly used by physical therapists including exercise, manual therapy, and traction. 50
One explanation offered for the lack of positive research findings from randomized trials is that patients with nonspecific LBP are regarded as a homogeneous group, with all patients equally likely to succeed or fail with any particular treatment. 26,42 Several authors have theorized that patients with LBP actually are a heterogeneous group consisting of several smaller homogeneous subsets. 13,29,43 As proposed, each homogenous subset, or classification, is more likely to respond to a type of treatment unique to that classification. A recent report from an international forum of primary care researchers on LBP ranked the accurate and reproducible characterization of subgroups of patients with LBP as the top priority for primary care research. 6
The classification of patients with acute LBP may be performed on the basis of pathoanatomy. 4,33 However, the difficulty in identifying a relevant pathoanatomic cause for most patients with LBP 1 makes this approach questionable. Others have proposed the classification of patients based on characteristics such as the presence or absence of sciatica, 5 the duration of the symptoms (acute, subacute, or chronic), 51 or work status. 43 Although these characteristics may have some prognostic value, their ability to direct clinicians to specific treatments that improve outcomes has not been established. 8
Other classification systems have incorporated impairments identified during the physical examination into the process of classification, linking specific treatments with each classification. 13 The system described by McKenzie 29 has been reported as the most commonly used classification system by physical therapists for this purpose. 2
More recently, Delitto et al 13 have proposed a treatment-based classification (TBC) system for use in the evaluation and treatment of patients with acute LBP. This system uses information gathered from the physical examination and from patient self-reports of pain (pain scale and pain diagram) and disability (modified Oswestry questionnaire) to classify the patient. The classification then guides the treatment of the patient.
The TBC system is designed for patients judged to be in the acute stage, 13 with the determination of acuity based on the nature of the patient's symptoms, the degree of disability, and the goals for management instead of strictly on the elapsed time from injury. Patients in the acute stage are those with higher levels of disability (Oswestry scores generally greater than 30) and substantial patient-reported difficulty with basic daily activities such as sitting, standing, and walking. Management goals are to improve the ability to perform basic daily activities, reduce disability, and permit the patient to advance in his or her rehabilitation. Patients judged to be in the acute stage are assigned to a classification, which guides the initial treatment. Patients judged to be in a more chronic stage are treated with a conditioning program designed to improve strength, flexibility, and conditioning, or with a work-reconditioning program. 13
The classification process for patients in the acute stage was the focus of this study. Seven classifications are described: immobilization, lumbar mobilization, sacroiliac mobilization, extension syndrome, flexion syndrome, lateral shift, and traction. 13 The classifications along with their corresponding key examination findings and recommended treatments are shown in Table 1. To facilitate comparisons among classifications, these seven classifications may be collapsed further into four classifications based on similarities in the prescribed treatments (Figure 1): immobilization, mobilization (either sacroiliac or lumbar), specific exercise (flexion, extension, or lateral shift correction), or traction.
The immobilization classification is purported to identify patients with lumbar segmental instability. Key examination findings are gathered primarily during history-taking and include a history of frequent episodes of symptoms precipitated by minimal perturbations, frequent use of manipulation with short-term relief of symptoms, trauma, or reduced symptoms with the prior use of a corset.
Many of these findings have been proposed in the literature to indicate possibly lumbar segmental instability. 20,24,38 Physical examination findings may include aberrant movements during lumbar flexion (i.e., an instability catch) 36,38 or generalized ligamentous laxity. 3 Treatment focuses on strengthening exercises for the back extensor and abdominal exercises 28 as well as stabilization exercises designed to improve dynamic control of the lumbar spine. 37
The mobilization classification includes patients believed to have indications for either sacroiliac or lumbar region mobilization or manipulation. Sacroiliac region mobilization is indicated by asymmetries of the pelvic landmarks (anterior superior iliac spines, posterior superior iliac spines, and iliac crests) with the patient in the standing position and by positive results in three of four tests as follows: 1) asymmetry of posterior superior iliac spine posterior superior iliac spine heights with the patient sitting, 2) standing flexion test, 3) prone-knee flexion test, and 4) supine to long-sitting test. These tests are described in detail elsewhere. 11
Acute-stage treatment involves a manipulation technique proposed to affect the sacroiliac joint region, 17 muscle energy techniques, 32 and range of motion (ROM) exercises for the lumbosacral spine. Lumbar mobilization is believed to be indicated by the presence of 1) unilateral paraspinal pain in the lumbar region and 2) asymmetric amounts of lumbar side-bending ROM with the patient standing in either an opening pattern (limited and painful flexion and side-bending ROM to the side opposite the pain) or a closing pattern (limited and painful extension and side-bending ROM to the same side as the pain). Treatment consists of lumbar mobilization or manipulation techniques 7 and ROM exercises for the lumbosacral spine.
The key examination finding that places patients into a specific exercise classification is the presence of centralization with movement of the lumbar spine. 29 Centralization occurs when the patient's pain or paresthesia is abolished or moves from the periphery toward the spine. The presence of centralization has been linked to prognosis by other researchers. 16,22 When either lumbar flexion or extension is found to produce centralization, the patient is treated with specific exercises in the direction producing the centralization. Patients also are educated to avoid positions found to peripheralize symptoms during examination. Patients noted to have a lateral shift in which the shoulders are offset from the pelvis in the frontal plane 29,39,47 also may be given specific exercise treatment.
The primary examination findings that lead to a classification of lateral shift are a visible frontal plane deformity and asymmetric side-bending ROM in standing. If correction of the deformity produces centralization, the patient is taught specific exercises designed to correct the lateral shift (i.e., pelvic translocation). 29
Traction classification is reserved for patients with signs and symptoms of nerve root compression who are unable to centralize with any lumbar movements. Acute stage treatment involves the use of mechanical or autotraction 35 in an attempt to produce centralization.
Although several classification systems have been proposed, 14,29,49 few data have been published attesting to the reliability of these systems or describing outcomes of patients treated with a classification approach. 40 Without these data, judging the effectiveness of a classification system or making comparisons among competing systems cannot be undertaken.
The purpose of this study was twofold. First, the interrater reliability of the TBC system described by Delitto et al 13 was investigated by having different physical therapists evaluate and classify the same patient without any intervening treatment. The second purpose was to describe the use of the TBC system in consecutive patients referred to physical therapy for treatment of LBP. Comparisons were made among the classifications for the initial patient characteristics and the outcomes of physical therapy.