Factors related to the inability of individuals with low back pain to improve with a spinal manipulation.

Many interventions used by physical therapists in the management of patients with low back pain (LBP) lack evidence supporting their effectiveness. (1-4) For example, interventions such as thermal modalities, electrical stimulation, and biofeedback have not been studied sufficiently, whereas interventions such as transcutaneous electrical nerve stimulation, mechanical traction, and ultrasound have been studied and found to be ineffective. (3,4) The evidence regarding exercise interventions for patients with LBP has been equivocal. Exercise has generally been reported to be ineffective for patients with acute LBP, but is usually recommended for patients with chronic LBP. (1,2,5) Spinal manipulation is one intervention for patients with LBP that is supported by evidence. (1) Clinical practice guidelines published in the United States, (4) the United Kingdom, (6) and New Zealand (7) recommend manipulation for patients with LBP of less than 4 to 6 weeks' duration who do not have signs of nerve root compression.

The American Physical Therapy Association's Guide to Physical Therapist Practice (Guide) defines mobilization/ manipulation as "a manual therapy technique comprising a continuum of skilled passive movements to the joints and/or related soft tissues that are applied at varying speeds and amplitudes, including a small-amplitude/high-velocity therapeutic movement." (8(p688)) The Guide lists mobilization/manipulation as an intervention appropriate for the care of patients with spinal disorders. (8) In clinical practice, and in systematic reviews of research evidence, spinal manipulation is usually distinguished from spinal mobilization. (9,10) Spinal mobilization involves low-velocity, passive movements of a joint within or at the limit of its range of motion (ROM), whereas spinal manipulation involves a high-velocity thrust to a joint beyond its restricted ROM. (9,11) Some evidence supports the use of spinal manipulation rather than mobilization. (1) In 2 studies, (12,13) spinal manipulation was compared with spinal mobilization for patients with LBP. Superior results with the use of manipulation were found in both studies.

Despite evidence supporting the use of spinal manipulation for patients with LBP, the intervention appears to be underutilized by physical therapists. Jette and Jette (14) reported on the interventions used by physical therapists in clinics located in the United States in the management of over 1,000 patients with LBP prior to the generation of most clinical practice guidelines recommending the use of manipulation. Mobilization/ manipulation was utilized during the course of care for 35% of the patients. The authors did not distinguish between manipulation and mobilization, and therefore it is likely that spinal manipulation was actually utilized in a much smaller percentage of patients. Several interventions lacking evidence were used at much higher rates, including flexibility exercises (81%) and thermal modalities (86%). (14) Studies conducted outside the United States, and after the publication of clinical practice guidelines, reflect similar patterns. Li and Bombardier (10) recently surveyed 569 physical therapists in Canada regarding their treatment beliefs and recommendations for patients with LBP. Only 30% of the therapists surveyed reported that they believed spinal manipulation to be an effective intervention in the management of most patients with LBP. The percentages of respondents expressing beliefs in the effectiveness of several interventions without evidence were higher, including ice (82%), spinal mobilization (80%), heat (66%), electrical stimulation (53%), and mechanical traction (36%). (10) Gracey et al, (15) in a study of 1,062 patients managed for LBP in Ireland, reported utilization of spinal manipulation in 9% of patients, compared with mobilization (44%), electrotherapy (30%), heat (19%), and traction (15%).

Several reasons have been offered for the apparent dissonance between current practice guidelines and the low utilization of spinal manipulation by physical therapists. Chief among these reasons appears to be a concern about the risk-benefit ratio for spinal manipulation. (16,17) The benefits of lumbar spinal manipulation have been well documented; however, little research has been conducted on the risks of spinal manipulation, particularly for patients with LBP. The most serious risks associated with spinal manipulation occur when techniques are performed on the cervical spine. (18) The most serious risk of manipulation of the lumbar spine is cauda equina syndrome. (19) The level of risk, however, appears to be extremely low, with an estimated occurrence of less than once per 100 million lumbar manipulation procedures. (20) Haldeman and Rubenstein (19) reviewed the literature over a 77-year period and found 10 reports of cauda equina syndrome occurring as a result of lumbar spine manipulation. The risk of cauda equina syndrome from manipulation of the lumbar spine is thought to increase when manipulation is performed under anesthesia or in the presence of sciatica. (20) Avoiding these situations likely reduces the risk of cauda equina syndrome.

Although serious complications resulting from lumbar spine manipulation appear to be very rare, less serious adverse effects have been reported to be more common. (21) Senstrad et al, (21) in a survey of 1,058 patients receiving spinal manipulation by chiropractors in Norway, reported no instances of cauda equina or other severe complications, but they did note that 55% of the patients reported at least one adverse effect related to the intervention. These adverse effects represented a worsening of the patient's clinical status, at least in the short term. The most common adverse effects were local discomfort (53%) or radiating discomfort (10%), headache (12%), and fatigue (11%). Leboeuf-Yde et al (22) reported on 1,858 spinal manipulations performed on 625 patients by chiropractors in Sweden. In that study, 44% of the patients reported experiencing increased symptoms, including local discomfort, headache, and fatigue, following manipulation, and in 19% of these cases, the increased symptoms persisted for greater than 48 hours. No serious complications were reported. (22) These studies did not distinguish manipulations performed on the lumbar spine from those performed on other regions of the spine, making it difficult to assess the prevalence of these adverse effects in patients receiving only lumbar spinal manipulation.

Based on the extremely low risk of serious complications from lumbar spinal manipulation reported in the literature, it appears that the primary risk that therapists need to consider when weighing a decision to use lumbar manipulation techniques is the possibility that the patient's symptoms will worsen or, at least, fail to improve. Researchers reporting on the prevalence of adverse effects of spinal manipulation did not attempt to describe the clinical presentation of patients whose status was worsened as a result of lumbar spine manipulation. (21,22) If the clinical presentation of patients unlikely to benefit from lumbar spine manipulation could be characterized, this information could be particularly useful for clinicians considering the risk-benefit ratio of utilizing manipulation for a specific patient with LBP.

We have previously reported on a cohort of 71 patients with LBP, each of whom underwent 1 or 2 sessions of lumbar spinal manipulation performed by physical therapists. (23) Our initial purpose in performing this study was to develop a clinical prediction rule that could assist physical therapists in identifying patients with LBP who are highly likely to experience a dramatic (50% or more) decrease in self-reported disability following 1 or 2 manipulation interventions. Thirty-two patients (45%) reported dramatic decreases in disability. We believe that patients with a high likelihood of experiencing 50% or more reduction in disability should be managed with spinal manipulation, and therefore we developed a clinical prediction rule to assist clinicians in identifying these patients. (23) As a secondary analysis, we were interested in characterizing the clinical response of those patients who did not experience a dramatic decrease in disability.

None of the patients in this small sample experienced any serious complications as a result of the manipulation. Many patients experienced a decrease in disability--measured using the Modified Oswestry Questionnaire--that exceeded the minimum clinically important difference of the instrument (24) but did not reach our threshold for a dramatic decrease. We believe that these individuals could be managed with spinal manipulation; however, because treatment effects were smaller, the possibility that other interventions might be more effective could not be excluded. A smaller group of patients in our initial study did not appear to show any clinically meaningful improvement in disability with the manipulation, and a few patients even experienced an increase in disability. We became interested in characterizing the group of patients who did not show any improvement with spinal manipulation, because we believed this could help clinicians identify which patients may be better suited to a different intervention. The purpose of the analysis in this article is to identify variables that are associated with a lack of improvement, or worsening, in the clinical status of patients with LBP who are managed with spinal manipulation.

Method

Subjects

This study involved a prospective cohort of patients with LBP recruited from 2 outpatient facilities: Brooke Army Medical Center and Wilford Hall Air Force Medical Center. All data were obtained in a previous study, (23) and in this report the data are analyzed to answer a different question. The primary purpose of this study was to identify clinical variables associated with a lack of improvement with a spinal manipulation technique. All subjects met the following inclusion criteria: age between 18 and 60 years; referral for physical therapy with a diagnosis related to the lumbosacral spine; a chief complaint of pain or numbness in the lumbar spine, buttock, and/or lower extremity; and a baseline Modified Oswestry Low Back Pain Disability Questionnaire (OSW) score (24) of at least 30%. Exclusion criteria were: current pregnancy, signs consistent with nerve root compression (positive straight-leg-raise [SLR] test at less than 45[degrees] or diminished lower-extremity force, sensation, or reflexes), prior lumbar spine surgery, or a history of osteoporosis or spinal fracture.

Seventy-five subjects were participants in this study, and 71 (95%) completed the intervention protocol. Four subjects did not return after the initial session. Two of these subjects left the study due to personal or work-related circumstances, 1 subject dropped out due to complications from an ongoing episode of gastrointestinal distress, and 1 subject did not return for follow-up. Data from these 4 subjects were not included in the analysis. Of the 71 subjects who completed the study, the mean age was 37.6 years (SD = 10.6, range = 19-59), and the mean initial OSW score was 42.4 (SD = 11.7, range = 30-86). The mean OSW score at the conclusion of the study was 25.1 (SD = 13.9, range = 0-64). Based on the criteria used in this study, 20 subjects (28%) were categorized as not improved with manipulation, and 51 subjects were improved. The mean change on the OSW for the subjects who improved was 25.1 points (SD = 14.2, range = 6-72, median = 22). For the subjects who did not improve, the mean change on the OSW was 0.0 points (SD = 3.6, range = -10-4, median = 0) (Fig. 1). The mean percentage of change on the OSW was 57.2% (SD = 25.0%, range = 16%-100%, median = 58.9%) for the subjects who improved and -0.03% (SD = 0.10%, range = -29%-13%, median = 0.0%) for the subjects who did not improve.

[FIGURE 1 OMITTED]

Descriptive statistics for the total sample, and for subjects who improved or did not improve, at baseline are contained in Tables 1 through 4. Interrater reliability of data obtained for the physical examination variables was assessed on 55 subjects. These subjects did not differ from the entire sample with respect to age, sex, initial OSW score, or duration of symptoms (P>.05). The results for both the categorical and continuous variables are contained in Table 5.

Therapists

Eight licensed physical therapists participated in this study. All therapists were associated with the US Army-Baylor University Post-professional Doctoral Program in Orthopaedic and Manual Physical Therapy. Four therapists were residents in the program, and 4 therapists were instructors. This 18-month program is designed to provide physical therapists serving in the US military with advanced training in orthopedic and manual physical therapy. All participating therapists had at least 3 years of clinical experience and used manual therapy, including lumbar spine manipulation techniques, in patient care. A 1-day training session covering the examination and intervention aspects of the study was conducted for all therapists.

Baseline Measures

All subjects initially completed several self-report measures and underwent a physical examination by the treating physical therapist during the initial session. The following self-report measures were completed:

* Numeric pain rating--Subjects rated their current level of LBP intensity using an 11-point pain rating scale ranging from 0 ("no pain") to 10 ("worst imaginable pain"). (25)

* Pain diagram--Subjects indicated the location of their symptoms on a body diagram. (26) The body diagram was used to categorize the symptom location as low back, buttock/thigh, or distal to the knee based on the distal-most extent of the symptoms indicated. Werneke et al (27) found high levels of interrater reliability (kappa = .92) for physical therapists coding the distal-most extent of symptoms from a pain diagram.

* Modified OSW--Disability due to LBP was measured with a modified version of the OSW, a 10-item scale originally described by Fairbank et al. (28) Each item is scored from 0 to 5, and the final score is expressed as a percentage, with higher numbers indicating greater disability. The modified OSW substitutes a section regarding employment/homemaking ability for the section related to sex life. This modified version has been shown to possess high levels of test-retest reliability (intraclass correlation coefficient [ICC] = .90 for 23 subjects with LBP whose condition remained stable over 4 weeks), construct validity (Pearson correlations with global patient ratings and other region-specific disability measures greater than .80), and responsiveness (effect size = 1.8 in patients receiving physical therapy interventions for LBP). (24) These findings are similar to those reported for the original version.

* Fear-Avoidance Beliefs Questionnaire (FABQ)--The FABQ quantifies the level of fear of pain and beliefs about avoiding activity in patients with LBP. (29) The FABQ has 16 items, each scored 0 to 6, with higher numbers indicating increased levels of fear-avoidance beliefs. The FABQ contains 2 subscales: a 7-item work subscale (score range = 0-42) and a 4-item physical activity subscale (score range = 0-24). High levels of test-retest reliability for the physical activity subscale (ICC = .77) and the work subscale (ICC = .90) have been reported. (30) The FABQ work subscale has been associated with an increased likelihood of current and future disability and work loss in patients with chronic LBP (29,31-33) and acute LBP. (34)

After completion of the self-report measures, all subjects gave a history and underwent a physical examination. The history taking and physical examination were performed by the first physical therapist. The physical examination was then repeated during the same session by a second therapist in order to evaluate the interrater reliability of data obtained for the examination items. The second therapist performed the physical examination items only and was masked to the results of the first therapist's examination. The second therapist became the treating therapist. The results of the first physical examination were used for the data analysis.

Taking the history included recording the subject's age and sex from the medical record and interviewing the subject regarding the duration of current symptoms and mode of onset (gradual or sudden). Subjects were asked about a prior history of LBP. If a prior history was present, subjects were asked if they perceived the frequency of their LBP episodes to be increasing, decreasing, or remaining the same. Subjects were asked to rank sitting, standing, and walking as to which was the worst and best with respect to their symptoms. The physical examination included several components. The methods for conducting and grading of the physical examination are described in the Appendix. (35-49) The components of the physical examination are listed below:

* Waddell's nonorganic signs--The 5 nonorganic signs (regional disturbance, superficial/nonanatomic tenderness, simulation, distraction, and overreaction) were assessed as described by Waddell et al. (50) These signs are proposed to detect abnormal illness behavior, defined as "maladaptive overt illness related behavior which is out of proportion to the underlying physical disease and more readily attributable to associated cognitive and affective disturbances. (51(p210)) The number of positive signs was summed for a total score ranging between 0 and 5.

* Lumbar spine ROM--Active ROM of the lumbar spine was measured with a single inclinometer using the methods described by Waddell et al. (35) Total flexion, pelvic flexion, lumbar flexion, total extension, and left and right side bending were measured. The interrater reliability for these methods is excellent (ICC = .87-.95 in 60 subjects with chronic LBP). (35)

* Status change with lumbar ROM--Each lumbar ROM movement was judged as having 1 of 3 possible effects on the subject's status: centralization, peripheralization, or no change (ie, status quo). (36,52) Centralization occurs when, during a lumbar movement, the subject reports that paresthesia or pain is abolished or moves from the periphery toward the lumbar spine. Peripheralization occurs when a subject reports paresthesia is produced, or paresthesia or pain moves distally from the lumbar spine, during a movement. Movements that do not produce centralization or peripheralization are judged to be status quo. We have found good interrater reliability using these definitions for judging status change during lumbar active ROM (kappa = .79 for judgments made by 40 examiners on 12 patients with LBP). (53)

* Hip rotation ROM--Passive ROM of hip medial rotation (MR) and lateral rotation (LR) were measured with the subject in a prone position using an inclinometer and the methods described by Barbee-Ellison et al. (37) These authors reported excellent interrater reliability with these procedures (ICC = .95-.97 in 50 patients with LBP). (37) From the measurements of MR and LR for the right and left hips, we calculated the following values:

--Total rotation ROM--Total rotation ROM was calculated for each hip by adding the MR and LR ROM measurements.

--Average rotation ROM--Average rotation ROM values were calculated by adding the ROM of the left and right hips and dividing by 2. Average MR, LR, and total ROM were calculated.

--Rotation ROM Discrepancy--Discrepancy values were calculated by taking the absolute value of the right hip ROM minus the left hip ROM. Discrepancies in MR, LR, and total ROM were calculated.

* Straight-leg-raise ROM--The passive ROM of hip flexion with the knee maintained in extension was measured with an inclinometer using the method described by Waddell et al, (35) who reported excellent interrater reliability with this technique (ICC = .94-.96 in 60 subjects with chronic LBP).

* Lumbar segmental mobility and pain provocation--The mobility and pain provocation of each lumbar segment was assessed by applying posterior-to-anterior pressure over each spinous process with the subject in a prone position. For each spinal segment, the mobility was graded as normal, hypomobile, or hypermobile, and pain provocation was judged as being absent, local (ie, pain provoked is localized to the spinal segment), or distal (ie, pain provoked is not localized to the spinal segment). Some authors (38,54,55) have reported poor interrater reliability for judgments of spinal segmental mobility. These studies attempted to grade mobility on scales with 7 to 11 levels of judgments. Researchers who have used 3 to 5 levels of mobility judgments, similar to those used in this study, have reported better interrater reliability (kappa = .40-.68 in samples of 150 and 50 subjects). (56,57) Judgments of segmental pain provocation have generally been found to have higher interrater reliability than judgments of segmental mobility (kappa = .67-.71 in sample of 150 subjects). (57)

* Tests for sacroiliac (SI) region dysfunction--Numerous tests purported to indicate dysfunction in the SI region were assessed. The tests were grouped into the following 3 categories:

--Tests of bony landmark symmetry--These tests used palpation to assess the symmetry of the right and left sides for several bony landmarks. Tests were judged to be positive if asymmetry between the left and right landmarks was perceived by the examiner. The landmarks assessed were the anterior superior iliac spine (ASIS), iliac crest, posterior superior iliac spine (PSIS), and greater trochanter with the subject in a standing position; the PSIS with the subject in a sitting position; and the pubic tubercle with the subject in a supine position. The reliability of data obtained with tests of SI region bony landmark symmetry has been reported to be low in previous studies (percentages of agreement = 35%-51%). (58,59)

--Tests of motion symmetry--These tests assessed the symmetry of movement of bony landmarks. Tests were judged to be positive if movement was determined to be asymmetrical by the examiner. Motion symmetry tests included in this study were the Gillet, long-sitting, prone knee-bend, seated and standing flexion tests. The interrater reliability of data obtained with these motion symmetry tests has been reported to be low (kappa = .02-.22 in various samples of subjects with LBP). (39,40,58,60-62)

--Tests of pain provocation--These tests were judged to be positive when pain was provoked in the SI region. Provocation tests assessed in this study were the posterior shear, Patrick, Gaenslen, resisted hip abduction, sacral sulcus palpation, sacral thrust, and compression/distraction tests. Pain provocation tests have generally shown acceptable interrater reliability. Kappa coefficients of at least .62 have been reported for the Gaenslen, Patrick, posterior shear, and compression/distraction tests. (40,41,60) The interrater reliability of data obtained with the sacral sulcus palpation test has not been as high (kappa = .32-.41 for samples of 51 and 33 subjects with LBP). (41,63) The reliability of data obtained with the sacral thrust and resisted hip abduction tests has not been previously reported.

Intervention

All subjects received the same intervention protocol. After the initial examination was complete, the treating physical therapist performed a manipulation technique that has been found to be effective for reducing short-term disability in previous pilot studies. (64-65) The subject was positioned supine on the examining table. The therapist stood opposite the side to be manipulated and passively side-bent the subject's spine to the opposite side (ie, away from the therapist). The therapist then passively rotated the patient in the direction opposite to the side bending and delivered a thrust over the ASIS (Fig. 2). The side to be manipulated was determined with the following algorithm. First, if the standing flexion test was positive, the side found to be positive was manipulated; if the standing flexion test was negative, the side of greater tenderness during the sacral sulcus palpation test was manipulated. If both sides were tender, or if neither side was tender, the side reported by the subject to be more symptomatic was manipulated. If the subject could not identify a more symptomatic side, the therapist flipped a coin to determine the side to manipulate. This algorithm was designed to provide a consistent approach to determining the side of manipulation. Cibulka et al (42) found changes on both sides of the pelvis following the application of this manipulation technique; therefore, we believe it is likely that the technique affects both sides of the pelvis.

[FIGURE 2 OMITTED]

After the initial manipulation attempt, the physical therapist recorded whether a "pop" was heard or felt by either the therapist or the subject. If a pop was heard or felt, the therapist proceeded to the other treatment components. If no pop was heard or felt, the therapist attempted the manipulation again. If no pop was heard or felt on the second attempt, the therapist next attempted to manipulate the other side. A maximum of 2 attempts per side were permitted. If no pop was heard or felt after the fourth attempt, the therapist proceeded with the other treatment components.

The additional treatment components included during the initial session were (1) instruction to perform supine pelvic tilt exercises for ROM and (2) advice to remain as active as possible within the limits of the subject's symptoms. Subjects were instructed to lie supine with their hips and knees flexed and to alternately tilt the pelvis in an anterior and posterior direction in a pain-free ROM. Subjects were told to perform this activity for 10 repetitions, 3 to 4 times per day. Exercise for ROM was included along with manipulation in previous studies demonstrating the effectiveness of this manipulation technique. (64,65) The instruction to remain as active as possible is consistent with recommendations in clinical practice guidelines for the management of individuals with acute LBP. (4,6) Adherence to these instructions was not assessed.

The second treatment session occurred 2 to 4 days following the initial visit. Each subject completed the OSW prior to beginning the second session. The treating physical therapist calculated the percentage of improvement in OSW scores using the formula: (initial OSW score--follow-up OSW score)/(initial OSW score) x 100%. If the percentage of improvement was equal to or greater than 50%, the intervention was categorized as successful, and participation in the study was ended. If the percentage of improvement was less than 50%, the therapist repeated the physical examination and the manipulation protocol as outlined for the initial visit, and the subject returned for a third treatment session 2 to 4 days after the second session. During the third session, the subject again completed the OSW, and the percentage of improvement from the initial OSW score was calculated. If the percentage of improvement was 50% or greater, the intervention was categorized as successful, and study participation was ended. If the percentage of improvement was less than 50%, the intervention was categorized as not successful.

Subjects whose intervention was not successful with manipulation were further examined to identify the subgroup of subjects who not only did not reach the success threshold (50% improvement in OSW score), but did not show any improvement. Based on previous work (24,66) demonstrating the minimum clinically important difference in OSW scores to be 4 to 6 points, we categorized any subject who did not show greater than 5 points of improvement in the OSW score from initial treatment session to the third treatment session as having made no improvement. We believe that the inability to achieve even a minimum clinically important change in disability represents a treatment failure.

Data Analysis

Descriptive statistics (mean, standard deviation) were calculated for all baseline variables. Interrater reliability statistics were calculated for all physical examination variables using the results obtained by the first and second raters during the initial visit. Kappa coefficients and 95% confidence intervals (CIs) were calculated for variables graded as positive or negative. (67) Weighted kappa coefficients with 95% CIs were calculated for variables graded on ordinal scales (nonorganic signs, status change with movement testing, segmental mobility judgments, and SI region symmetry tests). (68) Equal weighting for each interval was used to incrementally penalize each level of disagreement. Intraclass correlation coefficients (model 2,1) with 95% CIs were calculated for each continuous baseline measure (lumbar, hip, and SLR ROM variables).

We classified the outcome of the manipulation as having shown improvement or having shown no improvement. Subjects categorized as showing no improvement were those subjects who demonstrated 5 points or less of improvement on the OSW by the time of the third treatment. All other subjects were categorized as having improved. This group included subjects reaching the threshold for success during the treatment based on a 50% or greater improvement on the OSW as well as those subjects who did not meet the threshold for success but did demonstrate 6 points or more of improvement on the OSW by the third treatment session. Because our study was a secondary analysis, we did not anticipate having adequate power to perform inferential and multivariate assessments of our data. We therefore did not adjust the significance level in our study, and we chose to calculate descriptive statistics to examine the relationship between individual baseline variables and treatment outcome. All baseline variables were analyzed for univariate significance with manipulation outcome. Pearson chi-square tests were used for nominal and ordinal baseline variables, and independent sample t tests were used for continuous baseline variables. A significance level of P<.05 was used for all comparisons.

The relative contributions of each variable with a univariate relationship to manipulation outcome were further explored using logistic regression analysis. All variables with a univariate relationship were entered into a logistic regression model with manipulation outcome as the dependent variable. Variables were entered in a standard manner (ie, all independent variables were entered at once). The goodness-of-fit of the final regression model was tested with the Hosmer-Lemeshow statistic. (69) The proportion of variance explained by the final model was determined using the Nagelkerke [R.sup.2] statistic. (70) For the final model, adjusted odds ratios (AORs) and 95% CIs were calculated for each independent variable to estimate the increase in odds of failure given a one-unit change in the independent variable with the other independent variables held constant. (69)

Results

The univariate relationships between the baseline variables and manipulation outcome are shown in Tables 1 through 4. Three variables from the self-reports and history were related to outcome: the duration of symptoms, the presence of LBP only, and having hypomobility in the lumbar spine with spring testing. Further examination of these variables showed that having a longer duration of symptoms, not having LBP only (ie, having buttock/lower-extremity symptoms), and not having some hypomobility in the lumbar spine were associated with lack of improvement with manipulation.

Several variables related to ROM of the hip had a relationship with manipulation outcome (Tab. 3). Two variables were related to MR (left hip MR ROM and discrepancy in MR ROM between the left and right hips), 2 variables were related to LR (left hip LR ROM and average LR ROM), and 2 variables were related to total hip rotation ROM (left hip total rotation ROM and average total hip rotation ROM). In order to reduce multicolinearity, the Pearson correlation coefficients among all the hip ROM variables were examined. Correlations existed between left hip LR and average LR (r=.94, P<.01), between left hip total rotation and average total hip rotation (r=.95, P<.01), and between hip MR discrepancy and left hip MR ROM (r=.25, P=.03). We eliminated the measures specific to the left hip from further consideration because we believed the average rotation and discrepancy values would be more generalizable. Average hip LR and average hip total rotation ROM also were highly correlated (r=.83, P<.01). Because average total hip rotation had a smaller probability value than average LR in its relationship with manipulation outcome, we eliminated average LR from further consideration. We therefore retained 2 hip ROM variables for further consideration: hip MR discrepancy and average total hip rotation ROM. Further examination of these 2 variables indicated that less discrepancy in MR ROM between the left and right hips and less average total hip rotation ROM were associated with not improving with manipulation.

Two tests for SI region dysfunction had a relationship with manipulation outcome: one test of symmetry (pubic tubercle asymmetry in a supine position) and one provocation test (Gaenslen sign) (Tab. 4). Further examination of these variables revealed that not improving with manipulation was associated with asymmetry of the pubic tubercles and with a positive Gaenslen sign. Because the assessment of pubic tubercle asymmetry had a negative kappa value, indicating that the level of agreement was less than that expected by chance, we excluded this variable from further analysis.

A total of 6 baseline variables had univariate relationships with manipulation outcome and were retained for further examination: duration of symptoms, having LBP only, some lumbar hypomobility, average total hip rotation ROM, MR ROM discrepancy, and Gaenslen sign. The 6 variables were entered into a logistic regression model. The final model fit the data (Hosmer-Lemeshow [chi square]=4.87, P=.77). The Nagelkerke [R.sup.2] value for the final model was .63, indicating that the 6 variables explained 63% of the variance in manipulation outcome. The AOR values for the 6 independent variables are given in Table 6. The only variable with an AOR that was not statistically significant in the final multivariate model was having LBP only.

Discussion

In our previous publication, we developed a clinical prediction rule that sought to identify patients with LBP who had a high likelihood of experiencing dramatic success with spinal manipulation (50% or greater reduction in disability). (23) The clinical prediction rule developed had 5 factors (Fig. 3). The presence of at least 4 of these 5 factors was associated with an increased likelihood of experiencing a dramatic decrease in disability (positive likelihood ratio=24.4). This result provided a first approximation of which variables may permit clinicians to identify patients with LBP who should be managed with manipulation because their probability of experiencing dramatic success is very high. The clinical response of subjects who did not experience dramatic success ranged from clinically important, but not dramatic, improvement to no improvement or even a worsening of disability. We believe spinal manipulation may be a viable intervention for patients who experience clinically important, but not dramatic, improvement. Further research, however, may identify other interventions as being more effective for this group. Patients who do not improve at all, or who actually experience a worsening of disability, in our opinion, should not be managed with manipulation. Therefore, our goal in this report was to characterize the clinical presentation of this small subset of patients. Although further examination of the importance of these variables is needed, we believe this information would be important to physical therapists because it would permit an a priori identification of patients with LBP who are unlikely to respond to manipulation.

Figure 3.
Variables composing the clinical prediction rule identified in
a previous study (23) for predicting dramatic success with spinal
manipulation. The presence of at least 4 of these findings was
associated with an increased likelihood of dramatic improvement
with spinal manipulation (positive likelihood ratio=24.4). (23)

1. Duration of symptoms (less than 16 days related to
     dramatic success)

2. Fear-Avoidance Beliefs Questionnaire work subscale (<18
     points related to dramatic success)

3. Symptom location (no symptoms distal to the knee related
     to dramatic success)

4. Hip medial rotation range of motion (>35[degrees] on at least one
     hip related to dramatic success)

5. Lumbar hypomobility (presence of some hypomobility related
     to dramatic success)

We examined only one manipulation technique in this study. The technique used has been described as being directed at the SI region (36,43); however, the validity of this claim has not been assessed. Some researchers have recently offered data questioning the specificity and mechanisms underlying spinal manipulation, (71,72) and it is currently unknown which tissues are affected by this or other manipulation techniques. Further research is needed to determine whether substantive differences exist among the multitude of different manipulation techniques available for the lumbosacral region, or whether the results obtained using one technique can be generalized to other techniques. We combined the manipulation with ROM exercises because we believe this is frequently done in clinical practice. The results reported here, therefore, are specific to manipulation with ROM exercise. Although these results are preliminary, we were able to identify several variables that appear to be associated with an increased likelihood of not improving with this manipulation technique.

From the subjects' medical history, the most important factors associated with inability to improve were a longer duration of symptoms and the presence of symptoms distal to the low back. These findings appear consistent with those of previous reports. The greater effectiveness of manipulation in patients with relatively acute symptoms than in those with longer-standing symptoms has been identified in subgroup analyses of previously published randomized trials. (11,73) Some authors (4,74) believe manipulation to be contraindicated for patients with sciatica. We excluded patients with signs of nerve root compression; however, we still found that patients with symptoms into the buttock or lower extremity were more likely not to improve with manipulation. Ninety percent of the subjects who did not improve had symptoms distal to the low back, and 40% had symptoms distal to the knee, compared with 61% and 20%, respectively, for subjects who improved.

Relatively few physical examination findings were associated with manipulation outcome. Most of the examination variables associated with treatment failure were related to hip rotation ROM. In general, we found that subjects who did not improve with manipulation had less MR and LR ROM. In particular, we found that subjects who did not improve had less discrepancy in MR ROM between the left and right hips and had less total rotation ROM. Cibulka et al (75) reported that patients with LBP suspected of having SI region dysfunction had greater LR than MR ROM on the painful side. The authors, however, did not link the ROM findings to the outcome of an intervention. We found that subjects who improved with manipulation had more ROM in both MR and LR and had greater side-to-side discrepancy in MR ROM. Other authors (76-78) have reported data suggesting an association between limited hip rotation ROM, particularly IR, and LBP. Some authors (37,75,79) have further speculated that a unilateral restriction in MR ROM in a patient with LBP may represent a unique pattern, indicating a specific intervention. Research is needed to explore the relationship between hip rotation ROM and manipulation outcome. Our data, however, suggest that this pattern may be well stilted to treatment with spinal manipulation, whereas symmetrical MR ROM may not respond well to this manipulation technique.

Only one test for the SI region, the Gaenslen test, was found to be associated with manipulation outcome. One reason for the lack of association for many of the tests may be the inability to obtain reliable measurements. With a few exceptions, our findings were similar to those of other researchers; interrater reliability was generally poor for the palpation of symmetry of bony landmarks and motion tests, whereas acceptable reliability was found for pain provocation test measurements. (40,80) We believe, however, that poor reliability may not be the sole explanation for the lack of association for the tests. Several tests that yielded data with acceptable reliability failed to show any relationship to the outcome of manipulation. Another explanation for the lack of utility of many of these tests may be related to faulty theories underlying their mechanisms. Most of these tests were developed based on theoretical assumptions that dysfunction in the spine, particularly the SI region, will result in bony misalignment and changes in movement patterns. (43,81,82) Evidence to support these theories has not been found. Even if the theories are valid, some studies have shown that the available movements in the SI region are extremely small, (83,84) raising doubts about the possibility, that manual assessment could detect the subtle alterations that might occur.

One examination variable that appears to be important in identifying patients who are unlikely to improve with manipulation was the lack of hypomobility in the lumbar spine. While this finding is intuitively attractive because it implies that patients without joint stiffness do not need manipulation, we believe the finding must be interpreted cautiously, given the reliability, of measurements obtained with this particular test. We examined the reliability of the therapist's judgment that some hypomobility was present in the lumbar spine or that no hypomobility was present, and we found a kappa value of .13. Kappa values for individual spinal levels ranged from .03 to .50. The percentage of agreement between the examiners was high (78%), and it is likely that the high prevalence of positive findings (85%) in our sample deflated the kappa coefficient somewhat. (85) Further work is needed to improve the interrater agreement on judgments of segmental mobility to make these results more generalizable.

The majority of our subjects (72%) showed meaningful clinical improvement with lumbar spinal manipulation, even without any attempt to identify subjects who actually needed the intervention. This result is not surprising given the favorable results of clinical trials that have studied the effectiveness of lumbar spinal manipulation in groups of patients with LBP and no signs of nerve root compression. (11,12,73) These findings also support the advice of clinical practice guidelines that advocate at least a trial of manipulation for all patients with a new onset of LBP who do not have signs of nerve root compression. (4,6,7) Clinical practice guidelines are designed to assist clinical decision making for a group of patients with a particular clinical condition such as LBP. (86) When examining large groups of patients undergoing intervention for LBP, therefore, manipulation should be observed to be in frequent use by evidence-based practitioners.

Practice guidelines, however, do not offer assistance in determining whether a specific patient being evaluated with LBP may be among the minority of patients unlikely to benefit from manipulation. The purpose of our study was to identify factors associated with an increased likelihood of not improving with lumbar spinal manipulation. We found that a longer symptom duration, the presence of symptoms distal to the low back, a lack of hypomobility in the lumbar spine, reduced hip rotation ROM, little discrepancy in hip MR ROM side-to-side, and a negative provocation test (Gaenslen test) were more common in subjects who did not improve with manipulation. If a patient exhibits several of these signs during an examination, the likelihood of improvement with manipulation may be minimal.

The results of our study are preliminary findings. We examined only one manipulation technique in one practice setting. We did not examine the long-term improvement with manipulation. Because this study was not a randomized trial, we cannot definitively connect the improvement, or lack of improvement, with the manipulation intervention. It is possible that the variables identified as associated with a lack of improvement with manipulation are generally poor prognostic factors for patients with LBP regardless of the intervention. Future studies using randomized subject assignment are needed to validate the relationship between the factors identified in this study and the outcome of management with a manipulation intervention.

Conclusion

Most subjects showed some improvement with the manipulation intervention. In the smaller subset of subjects who did not improve, several factors were identified as useful for explaining the difference in manipulation outcome. Further research is warranted to address these questions and to improve physical therapists' ability to identify patients who are likely to respond to a particular intervention, such as lumbar spinal manipulation.

Table 1.
Baseline Variables From the Self-Reports and Patient History

                                             All        Not Improved
Variable                                     Subjects   (n = 20)

Pain rating
  [bar]X                                      5.3        4.6
  SD                                          2.0        1.4
  Range                                      1-10       2-7
Modified Oswestry Low Back Pain Disability
    Questionnaire score
  [bar]X                                     42.4       40.0
  SD                                         11.7        9.4
  Range                                      30-86      30-64
Fear-Avoidance Beliefs Questionnaire score
  Work subscale
    [bar]X                                   12.9       15.0
    SD                                       10.3       11.2
    Range                                    0-34       0-32
  Physical activity subscale
    [bar]X                                   15.1       14.7
    SD                                        5.3        5.9
    Range                                    3-24       3-23
Distal-most extent of symptoms
  Low back pain only                         31%        10%
  Symptoms in the buttock/thigh              44%        50%
  Symptoms distal to the knee                25%        40%
Age (y)
  [bar]X                                     37.6       36.7
  SD                                         10.6        9.0
  Range                                      19-59      26-56
Sex (% female)                               41%        35%
Duration of symptoms (d)
  [bar]X                                     41.7       66.8
  SD                                         54.7       71.1
  Range                                      1-252      11-244
  Median                                     22         32
Mode of onset
  Gradual                                    31%        35%
  Sudden (minimal/no perturbation)           39%        30%
  Traumatic (eg, fall, lifting, pulling)     30%        35%
Prior history of low back pain               83%        85%
Episodes of low back pain becoming more
  frequent                                   35%        50%
Sitting ranked as best position              31%        35%
Standing ranked as best position             20%        10%
Walking ranked as best position              23%        30%
Sitting ranked as worst position             41%        30%
Standing ranked as worst position            27%        40%
Walking ranked as worst position             13%        15%

                                             Improved
Variable                                     (n = 51)   P

Pain rating
  [bar]X                                      5.4       .13
  SD                                          2.2
  Range                                      1-10
Modified Oswestry Low Back Pain Disability
    Questionnaire score                                 .28
  [bar]X                                     43.4
  SD                                         12.4
  Range                                      30-86
Fear-Avoidance Beliefs Questionnaire score
  Work subscale                                         .27
    [bar]X                                   12.0
    SD                                        9.9
    Range                                    0-34
  Physical activity subscale                            .39
    [bar]X                                   15.4
    SD                                        5.0
    Range                                    5-24
Distal-most extent of symptoms
  Low back pain only                         39%        .017 (a)
  Symptoms in the buttock/thigh              41%        .70 (a)
  Symptoms distal to the knee                20%        .076 (a)
Age (y)                                                 .65
  [bar]X                                     38.0
  SD                                         11.2
  Range                                      19-59
Sex (% female)                               43%        .53 (a)
Duration of symptoms (d)                                .002 (b)
  [bar]X                                     31.9
  SD                                         43.9
  Range                                      1-252
  Median                                     18
Mode of onset
  Gradual                                    29%        .65 (a)
  Sudden (minimal/no perturbation)           43%        .31 (a)
  Traumatic (eg, fall, lifting, pulling)     27%        .53 (a)
Prior history of low back pain               82%        .92 (a)
Episodes of low back pain becoming more
    frequent                                 29%        .10 (a)
Sitting ranked as best position              29%        .65 (a)
Standing ranked as best position             24%        .20 (a)
Walking ranked as best position              20%        .35 (a)
Sitting ranked as worst position             45%        .24 (a)
Standing ranked as worst position            22%        .12 (a)
Walking ranked as worst position             12%        .71 (a)

(a) Significance tested with chi-square test.

(b) Mann-Whitney test used due to nonnormal data distribution.

Table 2.
Variables From the Baseline Clinical Examination Used in This Study

                                         All        Not Improved
Variable                                 Subjects   (n = 20)

Nonorganic signs (b)
  [bar]X                                  0.70       0.75
  SD                                      1.0        1.2
  Range                                  0-4        0-4
  Median                                  0          0
Hypomobility (c) at one or more lumbar
  levels with spring testing             86%        70%
Pain at one or more lumbar levels with
  spring testing                         92%        85%
Lateral shift present                    13%        10%
Peripheralizes (d) with lumbar single
  movement testing                       25%        25%
Centralizes (e) with lumbar single
  movement testing                        6%         5%
Pain peripheralizes (d) at one or more
  levels with spring testing             41%        55%

                                         Improved
Variable                                 (n = 51)   P (a)

Nonorganic signs (b)                                .84 (b)
  [bar]X                                  0.58
  SD                                      0.86
  Range                                  0-3
  Median                                  0
Hypomobility (c) at one or more lumbar
    levels with spring testing           92%        .016
Pain at one or more lumbar levels with
    spring testing                       94%        .21
Lateral shift present                    14%        .67
Peripheralizes (d) with lumbar single
    movement testing                     26%        .97
Centralizes (e) with lumbar single
    movement testing                      6%        .89
Pain peripheralizes (d) at one or more
    levels with spring testing           35%        .13

(a) Significance values were calculated from chi-square tests
unless otherwise indicated.

(b) Mann-Whitney test used due to nonnormal data distribution.

(c)Hypomobility was determined by therapists' judgments during
segmental prone posterior-anterior mobility assessment.

(d) Peripheralization occurred when a test produced a report of
paresthesia or caused paresthesia or pain to move distally from
the lumbar spine.

(e) Centralization occurred when during a test produced a report
that paresthesia or pain was abolished or moved from the periphery
toward the lumbar spine.

Table 3.
Variables for Range of Motion (in Degrees) From the Baseline
Clinical Examination Used in This Study

                                     All Subjects (n = 71)

Variable                             [bar]X   SD     Range

Total flexion                        75.7     30.3   15-140
Pelvic flexion                       38.0     20.7    0-90
Lumbar flexion                       38.3     15.4    5-74
Total extension                      21.5      9.5    5-51
Left side bending                    27.5      7.8   10-43
Right side bending                   27.3      7.2   15-42
Average side bending                 27.7      6.5   13-40
Side-bending discrepancy (a)          5.3      4.7    0-25
Left straight leg raise              67.3     16.8   35-110
Right straight leg raise             67.5     16.7   40-125
Average straight leg raise           68.8     14.8   38-118
Straight-leg-raise discrepancy (a)    6.0      6.9    0-45
Left hip medial rotation             29.6      9.8    7-50
Right hip medial rotation            30.3     12.5    6-60
Average medial rotation              28.7      9.4    7-55
Medial rotation discrepancy (a)       6.4      5.7    0-30
Left hip lateral rotation            30.1     11.9    5-55
Right hip lateral rotation           30.3     12.5    8-57
Average lateral rotation             30.3     11.7    8-55
Lateral rotation discrepancy (a)      6.1      5.3    0-25
Left hip total rotation (b)          60.0     17.2   17-99
Right hip total rotation (b)         58.1     17.5   15-90
Average total rotation               59.0     16.5   16-93
Total rotation discrepancy (a,b)      7.8      7.0    0-44

                                     Not Improved (n = 20)

Variable                             [bar]X   SD     range

Total flexion                        84.3     27.5   40-140
Pelvic flexion                       44.0     23.7    5-90
Lumbar flexion                       40.4     13.3    6-55
Total extension                      22.5     11.8   10-51
Left side bending                    28.7      9.8   10-43
Right side bending                   30.0      6.7   15-42
Average side bending                 29.3      7.5   13-40
Side-bending discrepancy (a)          5.5      5.6    0-25
Left straight leg raise              73.0     19.4   35-110
Right straight leg raise             75.2     19.2   45-125
Average straight leg raise           74.1     18.4   40-118
Straight-leg-raise discrepancy (a)    6.8     10.0    0-45
Left hip medial rotation             25.6      8.0    7-40
Right hip medial rotation            28.4     11.2    6-40
Average medial rotation              29.9      9.6    7-40
Medial rotation discrepancy (a)       3.8      6.6    0-15
Left hip lateral rotation            25.4     11.0    5-45
Right hip lateral rotation           26.7     11.9    9-50
Average lateral rotation             26.0     10.0    8-47
Lateral rotation discrepancy (a)      5.3      4.4    0-17
Left hip total rotation (b)          50.9     13.9   17-80
Right hip total rotation (b)         52.2     16.0   15-80
Average total rotation               51.6     14.4   16-79
Total rotation discrepancy (a,b)      6.6      4.3    1-16

                                     Improved (n = 51)

Variable                             [bar]X   SD     Range    P

Total flexion                        76.2     28.5   15-131   .28
Pelvic flexion                       37.9     18.6    0-84    .24
Lumbar flexion                       38.5     16.0    5-74    .65
Total extension                      21.0      7.8    5-46    .53
Left side bending                    27.5      7.0   12-40    .59
Right side bending                   26.7      6.9   15-45    .076
Average side bending                 27.1      6.1   16-40    .20
Side-bending discrepancy (a)          5.2      4.4    0-15    .80
Left straight leg raise              67.1     13.7   35-90    .15
Right straight leg raise             66.5     13.3   40-100   .074
Average straight leg raise           66.8     12.9   38-95    .062
Straight-leg-raise discrepancy (a)    5.8      5.7    0-21    .62
Left hip medial rotation             31.5     10.1   10-50    .021
Right hip medial rotation            28.4     11.2    7-60    .32
Average medial rotation              29.9      9.6   10-55    .076
Medial rotation discrepancy (a)       7.4      6.2    0-30    .004
Left hip lateral rotation            32.1     12.1    6-55    .034
Right hip lateral rotation           32.0     12.8    8-57    .11
Average lateral rotation             32.3     11.6   12-55    .05
Lateral rotation discrepancy (a)      6.6      5.7    0-25    .35
Left hip total rotation (b)          63.6     17.1   20-99    .004
Right hip total rotation (b)         60.1     17.7   23-90    .077
Average total rotation               62.0     16.5   23-93    .016
Total rotation discrepancy (a,b)      8.3      7.7    0-44    .35

(a) Discrepancy values were calculated as the absolute value of the
left minus the right.

(b) Total hip rotation values were calculated by summing the medial
and lateral rotation values.

Table 4.
Tests for Sacroiliac Region Dysfunction Used in This Study (a)

                                      All Subjects   Not Improved
                                      (n = 71)       (n = 20)

Symmetry Tests

PSIS asymmetry in standing position   41%            40%
ASIS asymmetry in standing position   36%            30%
Iliac crest asymmetry in standing     33%            40%
PSIS asymmetry in sitting position    31%            30%
Pubic tubercle asymmetry in supine
  position                            49%            70%
Ischial tuberosity asymmetry in
  prone position                      41%            50%

Motion Tests

Standing flexion test                 55%            55%
Seated flexion test                   66%            65%
Long-sitting test                     46%            50%
Prone knee-bend test                  56%            70%
Gillet test                           60%            50%

Provocation Tests

Gaenslen test                         44%            25%
Posterior shear test                  56%            50%
Compression/distraction test          23%            25%
Patrick test-buttock pain             44%            40%
Patrick test-groin pain               14%            20%
Resisted hip abduction                36%            40%
Sacral sulcus test                    67%            45%
Sacral thrust test                    54%            65%

                                      Improved
                                      (n = 51)   P

Symmetry Tests

PSIS asymmetry in standing position   41%        .93
ASIS asymmetry in standing position   41%        .47
Iliac crest asymmetry in standing     31%        .49
PSIS asymmetry in sitting position    31%        .91
Pubic tubercle asymmetry in supine
  position                            41%        .029
Ischial tuberosity asymmetry in
  prone position                      39%        .43

Motion Tests
Standing flexion test                 55%        .99
Seated flexion test                   66%        .98
Long-sitting test                     43%        .65
Prone knee-bend test                  49%        .11
Gillet test                           65%        .25

Provocation Tests
Gaenslen test                         53%        .033
Posterior shear test                  59%        .50
Compression/distraction test          24%        .90
Patrick test-buttock pain             45%        .70
Patrick test-groin pain               12%        .37
Resisted hip abduction                35%        .71
Sacral sulcus test                    59%        .29
Sacral thrust test                    69%        .77

(a) Values represent the percentage of subjects with positive
tests. Significance values were determined with chi-square tests.
PSIS = posterior superior iliac spine, ASIS = anterior superior
iliac spine.

Table 5.
Interrater Reliability of Data Obtained for the Physical
Examination Items (a)

Categorical Examination Variables

                                      Prevalence    Prevalence
                                      of Positive   of Positive
                                      Responses     Responses
Variable                              (Rater 1)     (Rater 2)

Hypomobility at one or more lumbar
  levels with spring testing          87%           84%
Pain at one or more lumbar levels
  with spring testing                 91%           87%
Lateral shift present                 16%           16%
Peripheralizes with lumbar single
  movement testing                    18%           20%
Centralizes with lumbar single
  movement testing                     4%           11%
Pain peripheralizes at one or more
  levels with spring testing          45%           47%
Nonorganic signs                      0 signs-62%   0 signs-60%
                                      1 sign-31%    1 sign-36%
                                      2 signs-4%    2 signs-4%
                                      3 signs-4%    3 signs-0%
PSIS asymmetry in standing position   40%           45%
ASIS asymmetry in standing position   35%           45%
Iliac crest asymmetry in standing
  position                            35%           42%
PSIS asymmetry in sitting position    35%           40%
Pubic tubercle asymmetry in supine
  position                            55%           51%
Ischial tuberosity asymmetry in
  prone position                      40%           49%
Standing flexion test                 60%           60%
Seated flexion test                   35%           40%
Long-sitting test                     51%           42%
Prone knee-bend test                  58%           47%
Gillet test                           64%           42%
Gaenslen test                         32%           32%
Posterior shear test                  39%           41%
Compression/distraction test          31%           29%
Patrick test                          44%           41%
Resisted hip abduction                45%           39%
Sacral sulcus test                    42%           49%
Sacral thrust test                    26%           25%

                                      Percentage
                                      of           Kappa
Variable                              Agreement    (95% CI)

Hypomobility at one or more lumbar
  levels with spring testing          78%           .13 (-.31, .56)
Pain at one or more lumbar levels
  with spring testing                 89%           .44 (.02, .86)
Lateral shift present                 82%           .34 (-.04, .71)
Peripheralizes with lumbar single
  movement testing                    73%           .12 (-.26, .50)
Centralizes with lumbar single
  movement testing                    89%           .21 (-.39, .81)
Pain peripheralizes at one or more
  levels with spring testing          73%           .45 (.22, .69)
Nonorganic signs                      60%           .26 (b) (-.01, .53)
PSIS asymmetry in standing position   53%           .13 (b) (-.13, .40)
ASIS asymmetry in standing position   58%           .31 (b) (.09, .53)
Iliac crest asymmetry in standing
  position                            56%           .23 (b) (-.02, .47)
PSIS asymmetry in sitting position    58%           .23 (b) (-.03, .48)
Pubic tubercle asymmetry in supine
  position                            45%          -.04 (b) (-.27, .19)
Ischial tuberosity asymmetry in
  prone position                      42%           .03 (b) (-.10, .13)
Standing flexion test                 38%          -.08 (b) (-.25, .09)
Seated flexion test                   62%           .25 (b) (.04, .46)
Long-sitting test                     58%           .21 (b) (-.02, .44)
Prone knee-bend test                  41%           .21 (b) (.03, .38)
Gillet test                           67%           .59 (c) (.41, .76)
Gaenslen test                         79%           .54 (c) (.36, .73)
Posterior shear test                  85%           .70 (c) (.60, .71)
Compression/distraction test          69%           .26 (.11, .42)
Patrick test                          80%           .60 (c) (.50, .70)
Resisted hip abduction                77%           .41 (c) (.20, .61)
Sacral sulcus test                    82%           .64 (c) (.53, .74)
Sacral thrust test                    77%           .41 (c) (.24, .58)

Continuous Examination Variables

                             Rater 1                  Rater 2

Variable                     [bar]X   SD     Range    [bar]X

Total flexion                76.5     32.2   10-140   76.3
Pelvic flexion               39.0     22.3    0-90    40.4
Lumbar flexion               37.5     17.1    5-74    35.9
Total extension              21.2     10.0    5-51    19.6
Left side bending            28.0      8.1   10-43    28.1
Right side bending           26.9      7.6   10-45    27.1
Left straight leg raise      66.8     18.1   20-110   67.3
Right straight leg raise     67.4     17.5   20-125   68.2
Left hip medial rotation     28.5     10.0    7-50    29.7
Right hip medial rotation    26.9     10.8    6-60    28.4
Left hip lateral rotation    27.3     11.9    5-51    29.9
Right hip lateral rotation   27.3     12.2    8-57    30.3

                             Rater 2
                                             ICC
Variable                     SD     Range    (95% CI)

Total flexion                30.7    9-136   .93 (.88, .96)
Pelvic flexion               20.9    4-79    .94 (.90, .98)
Lumbar flexion               14.9    5-66    .69 (.52, .81)
Total extension               9.4    5-45    .72 (.56, .82)
Left side bending             8.3   12-50    .49 (.26, .67)
Right side bending            8.6   10-55    .49 (.26, .67)
Leh straight leg raise       16.1   10-100   .86 (.77, .91)
Right straight leg raise     13.4   27-100   .76 (.62, .85)
Left hip medial rotation     10.6    7-54    .73 (.58, .83)
Right hip medial rotation    10.8    9-55    .61 (.42, .76)
Left hip lateral rotation    13.5    5-65    .48 (.25, .66)
Right hip lateral rotation   12.4    4-65    .42 (.17, .61)

(a) CI = confidence interval, ICC = intraclass correlation
coefficient, PSIS = posterior superior iliac spine. ASIS = anterior
superior iliac spine. See Table 2 for definitions of centralization
and peripheralization. Reprinted with permission from Lippincott
Williams & Wilkins from: Flynn TW, Fritz JM, Whitman JM, et al.
A clinical prediction rule for classifying patients with low back
pain who demonstrate short term improvement with spinal
manipulation. Spine. 2000;27:2835-2843.

(b) Weighted kappa.

(c) Right and left sides were judged separately, and scores were
combined to determine the overall reliability for the test.

Table 6.
Logistic Regression Analysis With Manipulation Outcome as the
Dependent Variable and Variables With Univariate Relationships
With Outcome as the Independent Variable (a)

                                   Adjusted
Variable                          Odds Ratio     95% CI       P

Average total hip rotation ROM
  (less ROM associated with
  failure)                        0.95         0.90, 1.00    .038
Duration of symptoms (longer
  duration associated with
  failure)                        1.03         1.01, 1.06    .007
Low back pain only (not having
  LBP only associated with
  failure)                        0.14         0.014, 1.46   .10
Gaenslen sign (negative result
  associated with failure)        0.11         0.019, 0.68   .017
Any hypomobility in the lumbar
  spine with spring testing
  (absence of hypomobility
  associated with failure)        0.092        0.010, 0.84   .035
Hip medial rotation ROM
  discrepancy (less discrepancy
  associated with failure)        0.68         0.51, 0.90    .007

(a) CI = confidence interval, ROM = range of motion,
LBP = low back pain.

Appendix
Operational Definitions for Performance and Grading of the
Physical Examination (a)

Lumbar Spine ROM                      Procedure

Flexion ROM (35)                      The patient stands erect. The
                                        inclinometer is held at
                                        T12-L1, and the patient is
                                        asked to reach down as far as
                                        possible toward the toes while
                                        keeping the knees straight.
Extension ROM (35)                    The patient stands erect. The
                                        inclinometer is held at
                                        T12-L1. The patient is asked
                                        to arch backward as far as
                                        possible.
Right and left side-bending           The patient stands erect with
  ROM (35)                              the inclinometer aligned and
                                        vertically in line with the
                                        spinous processes of T9 and
                                        T12. The patient is asked to
                                        lean straight over to one side
                                        as far as possible with the
                                        fingertips reaching down the
                                        side of the thigh.

Hip Rotation ROM                      Procedure

Right and left hip medial             The patient is positioned prone,
  rotation (37)                         and the hip on the side to be
                                        tested is placed in 0[degrees]
                                        of abduction with the knee
                                        flexed to 90[degrees]. The
                                        contralateral hip is placed
                                        in about 30[degrees] of
                                        abduction with the knee
                                        extended. The inclinometer
                                        is placed just proximal to the
                                        medial malleolus. The hip is
                                        moved passively into medial
                                        rotation until movement of
                                        the pelvis is felt or
                                        observed.
Right and left hip lateral            The patient setup and inclino-
  rotation (37)                         meter placement are the same
                                        as described for medial
                                        rotation. The hip is moved
                                        passively into lateral rota-
                                        tion until movement of the
                                        pelvis is felt or observed.

Straight-leg-Raise ROM                Procedure

Right and left straight-leg-raise     The patient is positioned
  ROM (35)                              supine. The inclinometer is
                                        positioned on the tibial crest
                                        just below the tibial
                                        tubercle. The leg is raised
                                        passively by the examiner,
                                        whose other hand maintains the
                                        knee in extension. The leg is
                                        raised slowly to the maximum
                                        tolerated straight leg raise
                                        (not the onset of pain).

Lumbar Segmental Testing              Procedure

Lumbar segmental testing for pain     The patient is positioned prone.
  provocation (38)                      The 12th rib and T12 vertebra
                                        are identified. L1 is identi-
                                        fied as the first spinous
                                        process distal to T12. The L1
                                        spinous process is contacted
                                        with the examiner's thenar
                                        eminence, and an anteriorly
                                        directed force is applied.
                                        The procedure is repeated at
                                        each lumbar level.
Lumbar segmental testing for          The procedure is the same as
  mobility (38)                         described for mobility
                                        testing.

Tests for Symmetry                    Procedure

PSIS symmetry-standing (36)           Palpation of right and left
                                        PSISs with the patient in a
                                        standing position.
ASIS symmetry-standing (36)           Palpation of right and left
                                        ASISs with the patient in a
                                        standing position.
Iliac crest symmetry-standing (36)    Palpation of right and left
                                        iliac crests with the patient
                                        in a standing position.
PSIS symmetry-sitting (44)            Palpation of the right and left
                                        PSISs with the patient in
                                        a sitting position.
Pubic tubercle symmetry-supine (43)   Palpation of the right and left
                                        tubercles with the patient in
                                        a supine position.
Ischial tuberosity                    Palpation of the right and left
  symmetry-prone (43)                   ischial tuberosities with the
                                        patient in a prone position.

Motion Tests                          Procedure

Standing flexion test (44,45)         The patient is positioned stan-
                                        ding, and the relative heights
                                        of the PSIS are assessed. The
                                        patient is asked to flex
                                        forward as far as possible,
                                        with the examiner continuing
                                        to palpate the PSIS.
Seated flexion tests (36,44)          The patient is seated, and the
                                        relative heights of the PSIS
                                        are judged. The patient is
                                        asked to bend forward as far
                                        as possible, with the examiner
                                        continuing to palpate the
                                        PSIS.
Long-sitting test (44,46)             The patient is positioned supine
                                        with hips and knees extended.
                                        The examiner grasps around
                                        each ankle with the thumbs
                                        below the medial malleoli.
                                        A visual estimation of leg
                                        length is made. The patient is
                                        assisted to a long-sitting
                                        position, and the examiner
                                        re-examines the relative leg
                                        lengths.
Prone knee-bend tests (36,42)         The patient is positioned prone.
                                        The relative leg lengths are
                                        assessed by looking at the
                                        heels. The examiner passively
                                        flexes the patient's knees to
                                        approximately 90[degrees].
                                        The relative leg lengths are
                                        assessed again in this
                                        position.
Gillet test (39)                      The patient is positioned
                                        standing. The examiner places
                                        one thumb under the PSIS on
                                        the side being tested, with
                                        the other thumb over the S2
                                        spinous process. The patient
                                        is instructed to stand on one
                                        leg and flex the other hip and
                                        knee, bringing the leg toward
                                        the chest.

Provocation Tests                     Procedure

Goenslen test (39,41,47,48)           The patient is positioned supine
                                        with both legs extended. The
                                        leg being tested is passively
                                        brought into full hip and knee
                                        flexion, while the opposite
                                        hip is maintained in an
                                        extended position. Over-
                                        pressure is applied to the
                                        flexed extremity.
Posterior shear test (39,48,49)       The patient is positioned
                                        supine. The hip is flexed to
                                        90[degrees] and adducted. The
                                        examiner applies an axial
                                        force through the femur at
                                        different angles of hip
                                        adduction/abduction.
Compression/distraction test          The patient is positioned
  (41,48)                               supine. Pressure is applied
                                        first in a posterior and
                                        lateral direction
                                        (compression) on the ASIS
                                        simultaneously. Pressure
                                        is then applied in an anterior
                                        and medial direction on the
                                        ASIS (distraction).
Patrick test-buttock pain (41)        The patient's hip is flexed,
                                        abducted, and laterally
                                        rotated by placing the lateral
                                        malleolus on the knee of the
                                        opposite leg. Overpressure is
                                        applied to the medial knee
                                        while the pelvis is
                                        stabilized.
Resisted hip abduction (39)           The patient is positioned supine
                                        with the hip in about 30
                                        [degrees] of abduction. The
                                        examiner pushes the leg
                                        medially to cause an isometric
                                        contraction of the
                                        hip abductors.
Sacral sulcus test (40)               The patient is positioned prone.
                                        The examiner palpates with
                                        firm pressure in the region
                                        directly medial to the PSIS.
Sacral thrust tests (41)              The patient is positioned prone.
                                        The examiner delivers an
                                        anteriorly directed thrust
                                        directly over the sacrum.

Lumbar Segmental Testing               Judgment

Lumbar segmental testing for pain     Pain provocation is judged as
  provocation (38)                      absent, local (pain provoked
                                        is localized to the spinal
                                        segment), or distal (pain
                                        provoked is not
                                        localized to the spinal
                                        segment).
Lumbar segmental testing for          Mobility is judged as normal,
  mobility (38)                         hypomobile, or hypermobile.
Tests for Symmetry                     Criteria for Positive
PSIS symmetry-standing (36)           One PSIS judged to be higher
                                        than the other.
ASIS symmetry-standing (36)           One ASIS judged to be higher
                                        than the other.
Iliac crest symmetry-standing (36)    One iliac crest judged to be
                                        higher than the other.
PSIS symmetry-sitting (44)            One PSIS judged to be higher
                                        than the other.
Pubic tubercle symmetry-supine (43)   One pubic tubercle judged to be
                                        higher than the other.
Ischial tuberosity                    One ischial tuberosity judged to
  symmetry-pronen (43)                  be higher than the other.
Motion Tests                          Criteria for Positive
Standing flexion test (44,45)         A change in the relative
                                        relationship of the PSIS is
                                        found in the fully flexed
                                        position.
Seated flexion tests (36,44)          A change in the relative
                                        relationship of the PSIS is
                                        found in the fully flexed
                                        position.
Long-sitting test (44,46)             A change in the relative
                                        position of medial malleoli
                                        occurs.
Prone knee-bend tests (36,42)         A change in relative lengths
                                        occurs between the 2
                                        positions.
Gillet test (39)                      The PSIS fails to move posterior
                                        and inferior with respect
                                        to S2.
Provocation Tests                     Criteria for Positive
Goenslen test (39,41,47,48)           Pain is reproduced in either SI
                                        joint region with performance
                                        of the test.
Posterior shear test (39,48,49)       Buttock pain is produced.
Compression/distraction test          Pain is reproduced in the SI
  (41,48)                               joint region with either
                                        maneuver.
Patrick test-buttock pain (41)        Buttock or low back pain is
                                        produced.
Resisted hip abduction (39)           Buttock pain is produced.
Sacral sulcus test (40)               Pain is reproduced in the
                                        SI region.
Sacral thrust tests (41)              Pain is reproduced in the SI
                                        region.

(a) ROM = range of motion, PSIS = posterior superior iliac spine,
ASIS = anterior superior iliac spine, SI = sacroiliac. Reprinted
with permission from Lippincott Williams & Wilkins from: Flynn
TW, Fritz JM, et al. A clinical prediction rule for classifying
patients with low back pain who demonstrate short-term improvement
with spinal manipulation. Spine. 2002;27:2835-2843.

References

(1) van Tulder MW, Koes BW, Bouter LM. Conservative treatment of acute and chronic nonspecific low back pain: a systematic review of randomized controlled trials of the most common interventions. Spine. 1997;22:2128-2156.

(2) van Tulder MW, Malmivaara A, Esmail R, Koes BW. Exercise therapy for low back pain: a systematic review within the framework of the Cochrane Collaboration Back Review Group. Spine. 2000;25: 2784-2796.

(3) Philadelphia Panel Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions for Low Back Pain. Phys Ther. 2001;81:1641-1674.

(4) Bigos S, Bowyer O, Braen G, et al. Acute Low Back Problems in Adults. AHCPR publication 95-0642. Rockville, Md: Agency for Health Care Policy and Research, Public Health Service, US Dept of Health and Human Services; 1994.

(5) Faas A. Exercises: which ones are worth trying, for which patients, and when? Spine. 1996;21:2874-2878.

(6) Waddell G, McIntosh A, Hutchinson A, et al. Low Back Pain Evidence Review. London, United Kingdom: Royal College of General Practitioners; 1999.

(7) New Zealand Acute Low Back Pain Guide. Wellington, New Zealand: ACC and National Health Committee; 1997.

(8) Guide to Physical Therapist Practice. 2nd ed. Alexandria, Va: American Physical Therapy Association; 2001.

(9) Koes BW, Assendelft WJ, van den Heijden GJ, Bouter LM. Spinal manipulation for low back pain: an updated systematic review of randomized clinical trials. Spine 1996;21:2860-2871.

(10) Li LC, Bombardier C. Physical therapy management of low back pain: an exploratory survey of therapist approaches. Phys Ther. 2001; 81:1018-1028.

(11) Ottenbacher K, Di Fabio RP. Efficacy of spinal manipulation/ mobilization therapy: a meta-analysis. Spine. 1985;10:833-837.

(12) Hadler NM, Curtis P, Gillings DB, Stinnett S. A benefit of spinal manipulation as an adjunctive therapy for acute low-back pain: a stratified controlled study. Spine. 1987;12:703-705.

(13) Meade TW, Dyer S, Browne W, et al. Low back pain of mechanical origin: randomised comparison of chiropractic and hospital outpatient treatment. BMJ. 1990;300:1431-1437.

(14) Jette DU, Jette AM. Professional uncertainty and treatment choices by physical therapists. Arch Phys Med Rehabil. 1997;78:1346-1351.

(15) Gracey JH, McDonough SM, Baxter GD. Physiotherapy management of low back pain. Spine. 2002;27:406-411.

(16) Adams G, Sim J. A survey of UK manual therapists' practice of and attitudes towards manipulation and its complications. Physiother Res Int. 1998;3:206-227.

(17) Di Fabio RP. Manipulation of the cervical spine: risks and benefits. Phys Ther. 1999;79:50-65.

(18) Hurwitz EL, Aker PD, Adams AH, et al. Manipulation and mobilization of the cervical spine: a systematic review of the literature. Spine. 1996;21:1746-1759.

(19) Haldeman S, Rubentein SM. Cauda equina syndrome in patients undergoing manipulation of the lumbar spine. Spine. 1992;17: 1469-1473.

(20) Shekelle PG, Adams AH, Chassin MR, et al. Spinal manipulation for low-back pain. Ann Intern Med. 1992;117:590-598.

(21) Senstad O, Leboeuf-Yde C, Borchgrevink C. Frequency and characteristics of side effects of spinal manipulative therapy. Spine. 1997;22: 435-441.

(22) Leboeuf-Yde C, Hennius B, Rudberg E, et al. Side effects of chiropractic treatment: a prospective study. J Manipulative Physiol Ther. 1997;20:511-515.

(23) Flynn TW, Fritz JM, Whitman JM, et al. A clinical prediction rule for classifying patients with low back pain who demonstrate short-term improvement with spinal manipulation. Spine. 2002;27:2835-2843.

(24) Fritz JM, Irrgang JJ. A comparison of a modified Oswestry Low Back Pain Disability Questionnaire and the Quebec Back Pain Disability Scale. Phys Ther. 2001;81:776-788.

(25) Jensen MP, Turner JA, Romano JM. What is the maximum number of levels needed in pain intensity measurement? Pain. 1994;58: 387-392.

(26) Mann NH, Brown MD, Hertz DB, et al. Initial-impression diagnosis using low-back pain patient pain drawings. Spine. 1993;18:41-53.

(27) Werneke M, Hart DL, Cook D. A descriptive study of the centralization phenomenon: a prospective analysis. Spine. 1999;24:676-683.

(28) Fairbank JC, Couper J, Davies JB, O'Brien JP. The Oswestry Low Back Pain Disability Questionnaire. Phsiotherapy. 1980;66:271-273.

(29) Waddell G, Newton M, Henderson I, et al. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain. 1993;52:157-168.

(30) Jacob T, Baras M, Zeev A, Epstein L. Low back pain: reliability of a set of pain measurement tools. Arch Phys Med Rehabil. 2001;82:735-742.

(31) Crombez G, Vlaeyen JW, Heuts PH, Lysens R. Pain-related fear is more disabling than fear itself: evidence on the role of pain-related fear in chronic back pain disability. Pain. 1999;80:329-339.

(32) Klenerman L, Slade PD, Stanley IM, et al. The prediction of chronicity in patients with an acute attack of low back pain in a general practice setting. Spine. 1995;20:478-484.

(33) Hadijistavropoulos HD, Craig KD. Acute and chronic low back pain: cognitive, effective, and behavioral dimensions. J Consult Clin Psychol. 1994;62:341-349.

(34) Fritz JM, George SZ, Delitto A. The role of fear avoidance beliefs in acute low back pain: relationships with current and future disability and work status. Pain. 2001;94:7-15.

(35) Waddell G, Somerville D, Henderson I, Newton M. Objective clinical evaluation of physical impairment in chronic low back pain. Spine. 1992;17:617-628.

(36) Delitto A, Erhard RE, Bowling RW. A treatment-based classification approach to low back syndrome: identifying and staging patients for conservative management. Phys Ther. 1995;75:470-489.

(37) Barbee-EIlison JB, Rose SJ, Sahrmann SA. Patterns of hip rotation range of motion: comparisons between healthy subjects and patients with low back pain. Phys Ther. 1990;70:537-541.

(38) Maher C, Adams R. Reliability of pain and stiffness assessments in clinical manual lumbar spine examination. Phys Ther. 1994;74: 801-811.

(39) Meijne W, van Neerbos K, Aufdemkampe G, van der Wurff P. Intraexaminer and interexaminer reliability of the Gillet test. J Manipulative Physiol Ther. 1999;22:4-9.

(40) Dreyfuss P, Michaelsen M, Pauza K, et al. The value of medical history and physical examination in diagnosing sacroiliac joint pain. Spine. 1996;21:2594-2602.

(41) Laslett M, Williams M. The reliability of selected provocation tests for sacroiliac joint pathology. Spine. 1994;19:1245-1249.

(42) Cibulka MT, Delitto A, Koldehoff R. Changes in innominate tilt after manipulation of the sacroiliac joint in patients with low back pain: an experimental study. Phys Ther. 1988;68:1359-1363.

(43) Greenman P. Osteopathic manipulation of the lumbar spine and pelvis. In: White A, Anderson A, eds. Conservative Care of Low Back Pain. Baltimore, Md: Williams & Wilkins; 1991:210-215.

(44) Cibulka MT, Koldehoff R. Clinical usefulness of a cluster of sacroiliac joint tests in patients with and without low back pain. J Orthop Sports Phys Ther. 1999;29:83-92.

(45) Sutton SE. Postural imbalance: evaluation and treatment using flexion tests. J Am Osteopath Assoc. 1978;77:456-465.

(46) Bemis T, Daniel M. Validation of tile long sitting test on subjects with iliosacral dysfunction. J Orthop Sports Phys Ther. 1987;8:336-345.

(47) Maigne J-Y, Aivaliklis A, Pfefer F. Results of sacroiliac joint double block and value of sacroiliac pain provocation tests in 54 patients with low back pain. Spine. 1996;21:1889-1892.

(48) Gibson T, Grahame R, Harkness J, et al. Controlled comparison of short-wave diathermy treatment with osteopathic treatment in nonspecific low-back pain. Lancet. 1985;8440:1258-1260.

(49) Broadhurst NA, Bond MJ. Pain provocation for the assessment of sacroiliac joint dysfunction. J Spinal Disord. 1998;11:341-345.

(50) Waddell G, McCulloch JA, Kummel E, Venner RM. Nonorganic signs in low-back pain. Spine. 1980;5:117-125.

(51) Waddell G, Main CJ, Morris EW, et al. Chronic low-back pain, psychologic distress, and illness behavior. Spine. 1984;9:209-213.

(52) McKenzie RA. The Lumbar Spine: Mechanical Diagnosis and Therapy. Waikanae, New Zealand: Spinal Publications Ltd; 1989.

(53) Fritz JM, Delitto A, Vignovic M, Busse RG. Inter-rater reliability of judgments of the centralization phenomenon and status change during movement testing in patients with low back pain. Arch Phys Med Rehabil. 2000;81:57-61.

(54) Binkley J, Stratford PW, Gill C. Interrater reliability of lumbar accessory motion mobility testing. Phys Ther. 1995;75:786-795.

(55) Gonella C, Paris SV, Kutner M. Reliability in evaluating passive intervertebral motion. Phys Ther. 1982;62:436-444.

(56) Strender LE, Sjoblom A, Sundell K, et al. Interexaminer reliability in physical examination of patients with low back pain. Spine. 1997;22: 814-820.

(57) Lundberg G, Gerdle B. The relationships between spinal sagittal configuration, joint mobility, general low back mobility and segmental mobility in female homecare personnel. Scand J Rehabil Med. 1999;31: 197-206.

(58) Potter NA, Rothstein JM. Intertester reliability of selected clinical tests of the sacroiliac joint. Phys Ther. 1985;65:1671-1675.

(59) O'Haire C, Gibbons P. Inter-examiner and intra-examiner agreement for assessing sacroiliac anatomical landmarks using palpation and observation: pilot study. Manual Therapy. 2000;5:13-20.

(60) Albert H, Godskesen M, Westergaard J. Evaluation of clinical tests used in classification procedures in pregnancy-related pelvic joint pain. Eur Spine J. 2000;9:161-166.

(61) Carmicheal JP. Inter- and intra-examiner reliability of palpation for sacroiliac joint dysfunction. J Manipulative Physiol Ther. 1987;12: 155-158.

(62) Vincent-Smith B, Gibbons P. Inter-examiner and intra-examiner reliability of the standing flexion test. Manual Therapy. 1999;4:87-93.

(63) McCombe PF, Fairbank JC, Cockersole BC, Pynsent PB. 1989 Volvo Award in clinical sciences: reproducibility of physical signs in low-back pain. Spine. 1989;14:908-918.

(64) Delitto A, Cibulka MT, Erhard RE, et al. Evidence for use of an extension-mobilization category in acute low back syndrome: a prescriptive validation pilot study. Phys Ther. 1993;73:216-28.

(65) Erhard RE, Delitto A, Cibulka MT. Relative effectiveness of an extension program and a combined program of manipulation and flexion and extension exercises in patients with acute low back syndrome. Phys Ther. 1994;74:1093-1100.

(66) Beurskens AJHM, de Vet HC, Koke AJA. Responsiveness of functional status in low back pain: a comparison of different instruments. Pain. 1996;65:71-76.

(67) Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20:37-46.

(68) Landis RJ, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159-174.

(69) Sharma S. Applied Multivariate Techniques. New York, NY: John Wiley & Sons Inc; 1996.

(70) Nagelkerke NJD. A note on a general definition of the coefficient of determination. Biometrika. 1991;78:691-692.

(71) Herzog W, Scheele D, Conway PJ. Electromyographic responses of back and limb muscles associated with spinal manipulative therapy. Spine. 1999;24:146-152.

(72) Herzog W, Symons B. The effective forces transmitted by highspeed, low-amplitude thoracic manipulation. Spine. 2001;26: 2105-2111.

(73) MacDonald RS, Bell CMJ. An open controlled assessment of osteopathic manipulation in nonspecific low-back pain. Spine. 1990;15: 364-370.

(74) Eck JC, Circolone NJ. The use of spinal manipulation in the treatment of low back pain: a review of goals, patient selection, techniques, and risks. J Orthop Sci. 2000;5:411-417.

(75) Cibulka MT, Sinacore DR, Cromer GS, Delitto A. Unilateral hip rotation range of motion asymmetry in patients with sacroiliac joint regional pain. Spine. 1998;23:1009-1015.

(76) Chesworth BM, Padfield BJ, Helewa A, et al. A comparison of hip mobility in patients with nonspecific low back pain. Physiotherapy Canada. 1994;46:267-274.

(77) Mellin G. Correlation of hip mobility with degree of back pain and lumbar spinal mobility in chronic low back pain patients. Spine. 1988;13:668-670.

(78) Fairbank JC, Pysant PB, Van Poortvliet JA, et al. Influence of anthropometric factors and joint laxity in the incidence of adolescent back pain. Spine. 1984;9:461-464.

(79) Cibulka MT. The treatment of the sacroiliac component to low back pain: a case study. Phys Ther. 1992;72:917-922.

(80) Freburger JK, Riddle DL. Using published evidence to guide the examination of the sacroiliac joint region, Phys Ther. 2001;81: 1135-1143.

(81) Bourdillon JF, Day EA, Bookout MR. Spinal Manipulation. Oxford, United Kingdom: Butterworth-Heinemann; 1992.

(82) Lee D. The Pelvic Girdle: An Approach to the Examination and Treatment of the Lumbo-Pelvic-Hip Region. Edinburgh, Scotland: Churchill Livingstone, 1999.

(83) Stueresson B, Selvik G, Uden A. Movements of the sacroiliac joint: a roentgen stereophotogrammetric analysis. Spine. 1989;25:214-217.

(84) Tullberg T, Blomberg S, Branth B, Johnsson R. Manipulation does not alter the position of the sacroiliac joint: a roentgen stereophotogrammetric analysis. Spine. 1998;23:1124-1128.

(85) Spitznagel EL, Helzer JE. A proposed solution to the base rate problem in the kappa statistic. Arch Gen Psychiatry. 1985;42:725-728.

(86) Hayward RS, Wilson MC, Tunis SR, et al. Users' guide to the medical literature, VIII: how to use clinical practice guidelines, A: are the recommendations valid? JAMA. 1995;274:570-574.

JM Fritz, PT, PhD, ATC, is Assistant Professor, Department of Physical Therapy, University of Pittsburgh, 6035 Forbes Tower, Pittsburgh, PA 15260 (USA) ([e-mail]). Address all correspondence to Dr Fritz.

JM Whitman, PT, DSc, OCS, FAAOMPT, is Element Chief, Physical Therapy Element, Kirtland Air Force Base, Albuquerque, NM.

TW Flynn, PT, PhD, OCS, FAAOMPT, is Associate Professor, Department of Physical Therapy, Regis University, Denver, Colo.

RS Wainner, PT, PhD, OCS, ECS, FAAOMPT, is Assistant Professor, US Army-Baylor University Graduate Program in Physical Therapy.

JD Childs, PT, PhD, MBA, OCS, CSCS, FAAOMPT, is Senior Physical Therapist and Director of Research, Department of Physical Therapy, Wilford Hall Medical Center, San Antonio, Tex.

Dr Fritz, Dr Whitman, Dr Flynn, and Dr Wainner provided concept/idea/research design. Dr Fritz, Dr Flynn, Dr Wainner, and Dr Childs provided writing. Dr Whitman, Dr Flynn, and Dr Wainner provided data collection. Dr Fritz, Dr Whitman, and Dr Childs provided data analysis. Dr Flynn provided project management, and Dr Fritz provided fund procurement. Dr Fritz, Dr Flynn, and Dr Wainner provided subjects and facilities/equipment. Special thanks to Jake Magel, PT, DSc, OCS, FAAOMPT, Dan Rendeiro, PT, DSc, OCS, FAAOMPT, Matthew Garber, PT, DSc, OCS, FAAOMPT, and Barb Butler, PT, OCS, for their assistance in data collection. Special thanks to Stephen Allison, PT, PhD, ECS, for his assistance in the study design and statistical support.

This work was supported by a Research Grant from the Foundation for Physical Therapy.

This study was approved by the institutional review boards at Brooke Army Medical Center and Wilford Hall Air Force Medical Center.

The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense.

This article was received June 17, 2003, and was accepted August 18, 2003.