Interventions That Increase or Decrease the Likelihood of a Meaningful Improvement in Physical Health in Patients With Sciatica

Dianne V Jewell, Daniel L Riddle


Background and Purpose. The purpose of our study was to determine whether physical therapy interventions predicted meaningful short-term improvement in physical health for patients diagnosed with sciatica. Subjects. We examined data from 1,804 patients (age: X̄=52.1 years, SD=15.6 years; 65.7% female, 34.3% male) who had been diagnosed with sciatica and who had completed an episode of outpatient physical therapy. Methods. Principal components factor analysis was used to define intervention categories from specific treatments applied during the plan of care. A nested-model logistic regression analysis identified intervention categories that predicted meaningful improvement in physical health. Meaningful improvement was defined as a change of 14 or more points on the Physical Component Scale-12 (PCS-12) summary score. Results. Twenty-six percent (n=473) of patients had a meaningful improvement in physical health. Improvement was more likely in patients receiving joint mobility interventions (odds ratio [OR]=2.5, 95% confidence interval [CI]=1.5–4.4) or general exercise (OR=1.5, 95% CI=1.2–2.0). Patients who received spasm reduction interventions were less likely to improve (OR=0.77, 95% CI=0.60–0.98). Discussion and Conclusion. Physical therapists should emphasize the use of joint mobility interventions and exercise when treating patients with sciatica, whereas interventions for spasm reduction should be avoided.

Sciatica, a symptom often attributed to lumbosacral spine pathology, is frequently discussed both in the lay press and in the medical literature. Sciatica is defined as “pain radiating down the leg(s) below the knee along the distribution of the sciatic nerve, usually related to mechanical pressure and/or inflammation of lumbosacral nerve roots.1(p118) Results from multiple epidemiologic studies have suggested that the lifetime incidence of sciatica is between 2% and 40%.2 Risk factors for sciatica are previous trauma to the lower back, taller height, smoking, pregnancy, and occupational and work-related postures or movements.36 Societal costs resulting from disability attributable to sciatica, in combination with other types of low back pain, have been estimated at between $20 and $50 billion annually.1(p5) Despite the substantial individual and societal burdens associated with sciatica, relatively little is known about which nonsurgical interventions increase or decrease the likelihood of a successful outcome.

Management options for sciatica include analgesic medications, manipulation, surgical decompression, bed rest, and a variety of interventions provided by physical therapists.7,8 Physical therapy interventions include therapeutic exercise; functional training; manual therapy techniques, including mobilization and manipulation; electrotherapeutic modalities; mechanical modalities; and physical agents.9 We found 2 studies that examined physical therapist treatment preferences for patients with sciatica. Li and Bombardier10 used patient vignettes in a survey of Canadian physical therapists (N=274) and found that 80% to 99% of the therapists preferred patient education about back care as well as exercise at home and in the clinic for the vignette describing a patient with acute low back pain and sciatica. Battié et al11 used the same sciatica vignette as that described by Li and Bombardier10 and reported that 86% of physical therapists surveyed (N=186) in Washington state preferred applying ice, whereas 71% preferred body mechanics education. Only 45% of the therapists in the study by Battié et al11 indicated that they would use stretching exercises for the sciatica vignette. The findings of these studies suggested that, for patients with sciatica, there is a large amount of variation in physical therapists' treatment preferences.

Randomized clinical trials (RCTs) of physical therapy interventions for patients with sciatica are less common than trials conducted on patients with only low back pain7 or trials conducted on heterogeneous groups of patients with and without sciatica.12,13 A systematic review of RCTs conducted on patients with sciatica suggested that staying active (generally continuing the patient's daily routine and staying active at home) is no better at improving functional status or reducing sick leave than is bed rest or inactivity.14 Van Tulder and colleagues15 conducted a systematic review to determine whether exercise is effective in patients with low back pain. In most cases, physical therapists instructed patients in the exercises being studied. Van Tulder et al15 reported that only 4 of the 39 RCTs studied a homogeneous sample of patients with sciatica. In the other trials,7,1214 the authors either did not report whether sciatica was present or studied heterogeneous groups of patients. Van Tulder et al15 concluded that the effect of exercise on patients with sciatica could not be determined from the small number of studies.

Vroomen et al7 examined the effect of 4 conservative interventions in their systematic review of 19 RCTs conducted on patients diagnosed with sciatica. The majority of these trials examined pharmacologic approaches, such as nonsteroidal anti-inflammatory medications or epidural injections. Traction was the only physical therapy intervention examined. These authors found that the odds of improvement when traction was provided to patients with sciatica were essentially no better than chance (odds ratio [OR]=1.2, 95% confidence interval [CI]=0.7–2.0]).

Assendelft and colleagues16 conducted a meta-analysis of the effectiveness of spinal manipulative therapy for patients with various low back pain disorders, including sciatica. A total of 39 articles met the inclusion criteria for the analysis. The authors reported that the data from patients with sciatica were insufficient to draw conclusions about effectiveness. The systematic reviews7,14,15 and the meta-analysis16 suggested that there is no substantive randomized controlled trial evidence supporting the use of interventions such as those applied by physical therapists for patients with sciatica.

Several studies also used large databases of patient cohorts to examine the outcomes of interventions applied by physical therapists for patients with low back pain, some of whom presumably were diagnosed with sciatica.1720 None of these studies examined a homogeneous group of patients with low back pain. Rather, patients with a variety of low back pain disorders were pooled for analysis. For example, Jette and Jette18 examined the relationship between categories of interventions and health outcomes on the basis of data collected from 1,097 patients with various spinal disorders. They reported that provision of endurance exercises predicted better outcomes. We found no large-sample cohort studies of patients who were diagnosed with sciatica and who were treated by physical therapists. Given the lack of convincing evidence from systematic reviews, more research is needed to determine whether various interventions applied by physical therapists have either a positive or a negative impact on outcome.

Studies of physical therapy treatment preferences for patients with sciatica provide an indication of therapists' hypothetical choices but do not provide actual patient data with which to confirm these behaviors. Randomized clinical trials of physical therapy interventions conducted on patients with sciatica have provided no substantial evidence of treatment effectiveness, and outcome studies have not examined homogeneous groups of patients. The purpose of our study therefore was to determine whether physical therapy interventions, either in isolation or in combination, predicted meaningful short-term changes in physical health status for a large cohort of patients diagnosed with sciatica.


Focus On Therapeutic Outcomes Inc (FOTO),* a private rehabilitation outcomes management company, provided data for this study. Clinics contracting with FOTO use standardized questionnaires21 comprising several demographic and disorder-related questions as well as the Physical Component Summary-12 (PCS-12), a generic physical health measure.22 Consenting patients completed the questionnaires at initial intake and discharge from physical therapy. Therapists were not informed of their patients' initial or discharge PCS-12 scores.

The PCS-12 is a 12-item survey instrument derived from the Physical Component Scale (PCS-36) of the Medical Outcomes Study 36-Item Short-Form Health Survey questionnaire (SF-36). Scores on the PCS-12 range from 13 to 69, with 50 representing the mean score for the population of adults in the United States. Higher scores indicate less physical disability. The PCS-12 scores were shown to be highly correlated (Pearson r=.95) with scores on the PCS-36 for a sample of 2,333 people from the general US population.22

Several authors23,24 have assessed the utility of the PCS-12 in patients with low back pain. Luo and colleagues24 studied 2,520 patients with low back pain and found the PCS-12 scores to be moderately correlated (Pearson r=−.63) with scores on the Oswestry Low Back Pain Disability Index. The authors also evaluated the instrument's responsiveness in 506 patients and found the PCS-12 to be responsive to changes in pain intensity after 3 to 6 months. These results suggest that the PCS-12 is likely to be sufficiently sensitive to detect meaningful changes in the physical health of people with sciatica.

The initial sample in this study consisted of 3,690 patients who received outpatient physical therapy services for sciatica (ICD-9 code 724.3)25 between January 1, 1998, and December 31, 2000. We excluded patients who were not at least 18 years of age (n=5), did not complete their episode of care (n=895), had surgery for their condition (n=440), or had additional ICD-9 codes documented by the treating therapist (n=385). Because we were unable to distinguish between interventions delivered for sciatica and interventions delivered for other reasons, we studied only patients with an isolated diagnosis of sciatica. Also excluded were patients for whom categorical or ordinal variables were miscoded (ie, coded with scores not included among the survey options) or for whom interval or ratio variables had values that were greater than 2 standard deviations above their means (n=161).26

All FOTO data were masked to prevent recognition of individual patients, clinicians, or clinics. The FOTO research oversight committee approved the use of these data after review of the study proposal.


We studied 1,804 patients who were diagnosed with sciatica and who completed an episode of outpatient physical therapy during the specified time frame. An episode of care was defined as outpatient physical therapy services provided from a specific admission date to a specific discharge date within the study time frame. The majority of patients were female (65.7%), with a mean age of 52.1 years (SD=15.6 years). Information regarding patient race or ethnicity was not available in the database. Almost one half of the patients (44%) were employed full-time. Approximately one third of the patients (35.6%) indicated that the time since the onset of their condition was greater than 90 days. Most of the patients (62.6%) took medication for their symptoms, but the specific pharmacologic regimen was not recorded. The mean number of visits per episode of care was 8.35 (SD=5.1, range=1–72). The mean PCS-12 composite score at admission was 34.9 (SD=9.6), which was 15 points lower than the mean score of 50 for the general US population.27

Differences among included and excluded patients were noted in several characteristics. For example, patients included in the analysis were slightly older, less frequently employed, and less frequently taking medications for their back pain than those excluded according to the study criteria (Tab. 1).

Table 1.

Comparison Between Patients Included in and Excluded From Analyses

Therapists and Participating Clinics

A total of 910 physical therapists with an average of 8.6 years of clinical experience participated in the study. Of the 910 participating therapists, 4% reported that they were clinical specialists in orthopedic physical therapy, as defined by the American Board of Physical Therapy Specialties, or specialists in manual therapy, as defined by the American Academy of Orthopedic Manual Physical Therapists. Participating clinics were evenly divided between for-profit and not-for-profit ownership status and were predominantly located in the Midwest (34.8%) and South (33.1%) US Census Bureau regions. Physical therapist and clinic characteristics are summarized in Table 2.

Table 2.

Characteristics of Participating Physical Therapists and Clinics

Intervention Category Development

When a patient was discharged, the physical therapist completed a form that indicated the interventions provided over the episode of care. The therapist checked relevant items from a list of 60 treatment options. Definitions for each treatment are provided in the FOTO training manual.28 In our study, therapists used an average of 8.26 (SD=3.17) different interventions over an episode of care. Given that therapists are likely to use a variety of techniques as an episode of care evolves,17 we used a statistical method to group interventions to more accurately reflect the multidimensional nature of the plan of care.

We first assessed intervention frequency in order to evaluate whether all possible treatment options in the FOTO database should be used to derive treatment categories for patients with sciatica. An intervention was included when it was provided to at least 5% of patients in the sample. We suspected that when an intervention was used for 5% of patients or fewer, the intervention likely would not be represented frequently enough in the data set to assess its effect with adequate precision. We also reasoned that interventions used for fewer than 5% of patients were used by a small minority of therapists and therefore were not as relevant for study as more commonly used interventions. Thirty-seven of the 60 interventions, including spinal manipulation, work hardening, aquatic therapy techniques, and craniosacral therapy, did not meet our 5% criterion and were excluded. Table 3 summarizes the remaining interventions included in the study.

Table 3.

Percentages of Patients Receiving Individual Treatment Options

We conducted a principal components factor analysis with varimax rotation29 in order to determine whether the remaining interventions were provided to the sample patients in consistent combinations. Several researchers1820 have hypothesized that groups of interventions with common properties (eg, the thermal effects of ultrasound and moist heat) are commonly implemented by therapists treating patients with low back pain. We also anticipated that therapists would select from groups of interventions to address different aspects of a patient's problem. For example, we suspected that therapists would address severe pain by choosing interventions designed primarily to address this impairment (eg, moist heat, transcutaneous electrical nerve stimulation). We used an eigenvalue of >1 as the a priori criterion for the identification of treatment groups or factors.30 We then examined the factor loading score for each intervention to evaluate the extent to which each treatment option correlated with a specific factor. A minimum factor loading score of .500 was used for factor interpretation, as recommended by Sharma.30

Eight factors explaining 52% of the common variance were extracted in the factor analysis. Factors 1 to 7 each contained at least 2 interventions that met the criterion of ≥.500. We applied the following category labels on the basis of the definitions of the interventions (Tab. 4) in each of the groups: exercise, postural correction, joint mobility, functional activities, pain reduction, spasm reduction, and alternative therapies. Factor 8 was defined by only 1 intervention (ice) and was labeled accordingly. These treatment groups are similar to those previously reported1820 and appear to describe combinations of interventions with clinically sensible bases. Four of the 23 interventions in the analysis did not meet the criterion of ≥.500 and therefore were excluded from the analysis: endurance exercises (.463), home exercises (.415), pain modulation exercises (.403), and McKenzie exercises (.333). Table 5 summarizes the intervention categories and relevant factor loading scores obtained in the analysis.

Table 4.

FOTO Training Manual Definitions of Physical Therapy Interventions28,a

Table 5.

Physical Therapy Intervention Categories for Patients With Sciatica

Dependent and Independent Variables

For our dependent variable, we defined clinically meaningful improvement as an increase of 14 points or greater in the PCS-12 score. This threshold was selected based on the findings of an RCT examining outcomes of patients with acute low back pain, including patients with sciatica.31 In the study of Fritz and colleagues,31 patients were randomly assigned to either a classification-based treatment approach or an intervention group on the basis of guidelines published by the Agency for Health Care Policy and Research (AHCPR).1 These authors found that the use of a classification-based approach resulted in significantly improved physical health, a 14-point average improvement in PCS-36 scores after 4 weeks, compared with therapy guided by AHCPR guidelines. Based on the study by Fritz and colleagues31 and given the very high correlation between PCS-12 and SF-36 scores, we chose a 14-point change in the PCS-12 as a cutoff for differentiating between patients who did and patients who did not show a clinically important amount of change.

Because Fritz and colleagues31 examined patients with acute low back pain and only 18% had sciatica, we acknowledge that a cutoff of 14 points for clinically important change may be too conservative for patients in our study. Patients with acute low back pain and patients without sciatica tend to improve to a greater extent than do those with chronic low back pain or sciatica.32,33 Some of the patients in our study may have changed by less than 14 points but still may have had clinically meaningful improvement. Given the findings of Fritz and colleagues,31 however, we are reasonably confident that patients in our study who improved by more than 14 points had clinically meaningful short-term improvement. Data in our study were coded as 0 when an improvement of less than 14 points was found and 1 when an improvement of greater than 14 points was found.

The primary independent variables in our study were the 8 groups (factors) of interventions derived from the factor analysis. For each of the 8 factors described in Table 5, we scored a factor as present (coded as 1 in the data set) when a patient received 50% or more of the interventions represented by the factor. A factor was scored as absent (coded as 0 in the data set) when a patient received fewer than 50% of the interventions represented by the factor. We reasoned that, when at least one half of the interventions for a factor were provided to a patient, that factor was adequately represented. For example, the intervention category of postural correction was coded as present when the patient received any 2 of the 3 interventions listed for that category in Table 5. When the patient received none or only 1 of these interventions, the postural correction category was coded as absent for the patient. Finally, patient median age (>51 years), length of time since the onset of the condition (>90 days), and PCS-12 score at admission to physical therapy were included as potential confounding variables.

Data Analysis

Potential collinearity among the demographic variables was assessed with the Kendall tau or Spearman rho correlation analysis. We used multiple logistic regression to identify the most parsimonious set of variables that independently predicted clinically meaningful improvement for patients with sciatica. In order to account for variance from multiple levels of data, we used a model that nested patient data within therapists and therapist data within clinics. To protect against violation of assumptions of the model (non-zero intracluster correlation among clinics and compound symmetry), we used a robust variance estimation method.34 The calculated intracluster correlation for this study was .05, a value that represents the proportion of total variation in meaningful improvement in physical health status that can be attributed to differences between clinics.35 The levels of probability for entry of the primary independent variables into and removal from the model were set at .05 and .10, respectively.26 The secondary independent variables were added to the model to adjust for potential confounding. A P value of ≤.05 for the Wald test was used to determine the statistical significance of the predictor variables. A second model that included all possible interaction terms between the retained intervention categories also was tested.

Initial descriptive and correlational statistical analyses were conducted with SPSS 12.0 software for Windows. Logistic regression analysis was generated with SUDAAN 8.0,§ a statistical software program designed to analyze cluster-correlated data such as that used in this study.


Twenty-six percent (n=473) of the patients demonstrated a meaningful clinical improvement in their PCS-12 scores. Patient demographic characteristics were not significantly correlated with one another, with values ranging from .007 to .151.

Three intervention categories, exercise, joint mobility, and spasm reduction, remained in the model (P<.05) after adjustment for patient age, length of time since condition onset, and PCS-12 score at admission. Odds ratios indicated that joint mobility interventions and exercise increased the likelihood of meaningful improvement (Tabs. 6 and 7). For example, a PCS-12 score change of 14 points or more was 2.5 times more likely to occur than not to occur in patients who received joint mobility interventions. The use of spasm reduction interventions, however, was found to decrease the likelihood of a successful outcome (OR=0.77, 95% CI=0.60–0.98). That is, the odds of achieving an increase of 14 points or greater in PCS-12 scores were reduced 23% in patients receiving spasm reduction interventions.

Table 6.

Logistic Regression Model Predicting Meaningful Clinical Improvementa

Table 7.

Effect of Interaction Between Joint Mobility Interventions and Exercise Interventions on Odds for Clinically Meaningful Improvementa

We also found a statistically significant interaction between the exercise and joint mobility intervention groups. In this analysis, the reference group consisted of patients who received neither exercise interventions nor joint mobility interventions. The reference group was compared with 3 other groups: those who received exercise interventions alone, those who received joint mobility interventions alone, and those who received both joint mobility and exercise interventions. Odds ratios (Tab. 7) for the interaction indicated a statistically significant effect for joint mobility alone (P=.001), joint mobility combined with exercise (P=.004), and exercise alone (P=.004) in comparison with the reference group. Between-group comparisons for the 3 treatment categories further revealed that joint mobility alone had a statistically larger effect than exercise alone (P=.05); the remaining contrasts, however, were not statistically significant (exercise alone versus exercise and joint mobility combined: P=.45; joint mobility alone versus exercise and joint mobility combined: P=.22).


The results of our study suggest that joint mobility interventions, in the absence of exercise, increase the likelihood of meaningful clinical improvement in physical health in patients with sciatica (OR=2.5). This intervention category comprises manual joint mobilization techniques as well as activities that the patient performs to improve joint mobility. Assendelft et al16 noted in their systematic review that the data were insufficient to draw conclusions about the effectiveness of spinal manipulation for individuals with sciatica. However, our findings are consistent with individual studies examining the effectiveness of interventions focusing on improving joint movement in this patient population. Mathews et al13 demonstrated in an RCT of 513 patients that patients who had acute sciatica and who were receiving physical therapy showed improvement with manipulation alone compared to control patients who received infrared heat. Bronfort et al2 reported improvement in leg pain, back pain, and the Roland-Morris Disability Scale score in patients receiving spinal manipulation in combination with massage or traction from a chiropractor. Manipulation, defined as a “low-amplitude, high-velocity thrust at the limit of the patient's joint range,”36(pp616–617) was not included as an intervention in our factor analysis because fewer than 5% of the physical therapists reported using it for the sample patients. Mobilization was provided to 30% of our overall sample and to 35% of patients with meaningful clinical improvement (Tab. 3). Mobilization involves “repetitive low-velocity passive movements usually within or at the limit”36(p616) of the patient's available range of motion. Because they are related techniques, mobilization and manipulation may provide similar mechanisms for improvement in patients with sciatica.

One limitation of our data set is that we cannot determine the specific interventions that were provided as part of the joint mobility intervention. FOTO defines joint mobility28 as self-explanatory, implying to us that therapists had to judge for themselves whether the interventions that they applied were designed to affect joint mobility. We suspect that joint mobilization interventions are reasonably similar among therapists; however, in the absence of professional consensus on the operational definitions of all physical therapy interventions, joint mobility interventions may vary considerably depending on the therapist. Details from actual patient records would be required in order to identify the specific techniques attributed to these labels. The possibility that patients with sciatica may be able to improve their conditions with self-administered treatments focusing on improving joint mobility is an implication that should be considered.

The results of this study indicate that exercise interventions alone improve the likelihood of meaningful clinical improvement (OR=1.5) to a lesser degree than do joint mobility interventions alone. Recent studies examining the effectiveness of exercise in patients with low back pain varied in their results. Van Tulder et al15 reviewed 39 RCTs and concluded that exercise was not more effective than other options, such as manipulation or physical agents provided by physical therapists to patients with acute low back pain. Liddle and colleagues,37 however, found that exercise resulted in improvement in measures of pain and function in 16 studies of patients with chronic (>3 months) low back pain. Exercise was broadly defined in both reviews and included general strengthening, flexibility, and aerobic programs as well as specific approaches, such as McKenzie extension and Williams flexion exercises. The exercise intervention category in this study comprised strengthening, stretching, and flexibility activities. This study also focused on patients with sciatica and adjusted for the duration of symptoms, whereas the reviews included studies with heterogeneous samples of patients defined as having either acute or chronic symptoms. Our findings suggest that exercise may be beneficial for patients with sciatica but that this intervention alone may not increase the likelihood of clinically meaningful improvement as well as joint mobility.

The odds of meaningful clinical improvement were not statistically different when exercise and joint mobility interventions were provided in combination compared with being provided individually. Hofstee et al12 also reported no meaningful change in patients who had sciatica and who were receiving physical therapy comprising joint mobilization and exercise techniques in combination. This study and the study of Hofstee et al,12 however, are limited by the lack of detail regarding the exercise and joint mobilization techniques provided to the patients studied. The higher OR for the combined-intervention group (OR=1.8) than for the exercise-only group (OR=1.5) in this study suggests the need for additional research evaluating the relative effectiveness of specific combinations of exercise and mobility techniques.

The nature of the interaction between joint mobility and exercise also is important, given the intervention preferences described in this study. Twice as many patients in the sample received stretching, strengthening, or flexibility exercise interventions than received joint mobilization; 3 times as many received these interventions than received joint mobility interventions (Tab. 3). These values serve as an indicator of physical therapists' commitment to the use of traditional therapeutic exercise techniques. The preference for these exercise techniques may result from provider training or habit; however, their use may be indicated on the basis of the results of the patient examination. In either case, physical therapists would benefit from an understanding of which specific combinations of exercise and joint mobility, if any, would enhance patient outcome.

Our findings indicate that the application of passive techniques to address muscle spasms decreases the odds of a successful outcome (OR=0.77). This treatment category comprised ultrasound and massage, interventions provided to 46% and 25% of the sample, respectively. Li and Bombardier10 reported a strong preference for physical modalities despite evidence indicating a lack of their effectiveness. This study extends the implications of the prevailing evidence1,18 by suggesting that the use of passive physical modalities may actually hinder patient recovery.

This study also provides a description of the actual physical therapy treatment of patients with sciatica that differed from the choices indicated when US therapists were queried about hypothetical cases. Battié et al11 reported that physical therapists whom they surveyed overwhelmingly favored ice, followed by body mechanics training and stretching. All 3 interventions were used differently in the study by Battié et al11 than in the present study (86% versus 18%, 71% versus 44%, and 46% versus 64%, respectively). These differences may reflect a regional bias, as the physical therapists surveyed in the study by Battié et al11 were from 1 state in the Pacific Northwest, whereas our data were collected in clinics located throughout the United States. The time between studies also may explain the different results, as Battié et al11 conducted their survey in 1990, before the publication of clinical practice guidelines that recommended exercise for low back pain management.1,8

The results obtained by Li and Bombardier10 from surveys conducted in 1998 suggest an evolution of physical therapist practice that is more consistent with the findings of the present study. For example, 91% of their survey respondents indicated a preference for a home exercise program for patients with sciatica, and 88% of the patients in our sample received a home exercise program. Similar proportions also occurred for spinal mobilization (34% and 31%). Exercise other than at home was an aggregated term in the work by Li and Bombardier10; therefore, proportions cannot be compared directly with those for the individual techniques (ie, stretching, strengthening, and stabilization) used here. However, exercise was preferred or used a majority of the time in both studies. Of interest to Li and Bombardier10 was a strong preference for physical modalities (80%) despite the lack of evidence for their effectiveness indicated in the previously published AHCPR Clinical Practice Guideline.1 The use of specific physical agents, such as ultrasound (46%), moist heat (45%), and electrical stimulation for pain relief (34%), in our study also suggests that physical therapists still value this treatment approach despite the lack of supporting evidence.

Research evaluating the use of numerous physical therapy interventions for low back pain primarily has relied upon clinical judgment, rather than statistical methods, in order to create smaller numbers of treatment groups. These a priori classifications are based on a tacit understanding within the profession as to what constitutes a group of related techniques.10,11,1720 For example, Jette and Jette18 created 8 treatment categories for their assessment of health outcomes in patients receiving physical therapy for cervical or lumbar spinal impairments. These treatment categories apparently were based on the similarities of properties among the clustered interventions, as suggested by labels such as cold modalities, strength exercises, endurance exercises, and manipulation or mobilization. The advantage of an a priori classification approach is that it uses clinical theory and experience as its foundation; however, assumptions about relationships among techniques are not verified in an objective fashion. The statistically created intervention categories in this study confirm several of the previously used a priori groups, with the exception of modalities and exercise. Rather than clustering according to similar properties (ie, hot versus cold), physical agents appear to have been grouped according to their intended effects (ie, pain reduction and spasm reduction). However, exercises were not segregated on the basis of type (ie, strengthening versus stretching). In light of these results, we recommend the use of statistical techniques as an objective method by which to categorize multiple interventions in order to assess their effects on patient outcomes.


The ability to generalize the findings from this study is constrained by sample selection and composition. First, the patients were chosen in a nonrandomized fashion from a single outcomes management company that does not have a uniform distribution of clients across the United States. Selection bias within the FOTO data set also is suggested by the statistically significant differences demonstrated in age, employment status, and use of medications for symptoms between patients included in the data set and patients excluded from the data set (Tab. 1). Despite these differences, both groups had similar PCS-12 scores, suggesting that these attributes had limited influence on the impairments and functional limitations of these patients. Commitment to job or adequate symptom control with medications may explain why so many patients in the excluded group did not complete their episodes of care.

Second, the sample was trimmed to include patients with sciatica but without coexisting conditions. We used this strategy because so few studies focus on this patient population and because we could not identify which, if any, physical therapy interventions were provided for other impairments occurring during the episodes of care. Lumbago (n=41), back sprain (n=19), lumbosacral joint or ligament sprain (n=10), and arthralgia of the pelvis (n=10) were the most frequently listed secondary diagnoses in the sample. Therefore, our findings cannot be applied to people with sciatica and these coexisting conditions. We were unable to confirm the diagnoses recorded by the physical therapists or the methods by which the physical therapists selected individual diagnostic labels; therefore, the potential lack of reliability and validity of the ICD-9 codes themselves also should be considered.38

Coding and interpretation of treatment options posed other challenges in this study. Details from patient medical records were not available to confirm or clarify specific interventions beyond the general labels listed in the FOTO instrument. For example, techniques intended to increase lumbar range of motion may be classified by some physical therapists as flexibility, whereas other therapists may have selected stretching from the survey list. Functional training and joint mobility interventions are additional treatment options whose broad definitions may have allowed for different interpretations by survey respondents. As a result, we were unable to identify specific interventions within the categories that increased the likelihood of a meaningful clinical change in the patients. Professional consensus regarding operational definitions for specific intervention labels may address this challenge in the future.

A related issue is our factor analysis, which explained only 52% of the variance in treatment categories used by physical therapists in this study. The substantial portion of unexplained variance may be attributable to the interventions that we excluded because of weak factor loading or infrequent representation in the data set. Future research should address the impact of the interventions that were excluded from our analysis.

Finally, we were unable to assess the timing or sequence of interventions applied over the plan of care. Physical therapists identified all interventions provided during the episode of care upon patient discharge; therefore, we were unable to identify what role, if any, the timing or sequence of interventions may have played in patient outcomes. Jette and Delitto17 demonstrated that intervention choices evolve over time, presumably in response to changes in a patient's signs and symptoms. A deeper understanding of the interaction effect between joint mobility and exercise interventions might have been possible had we been able to evaluate the entire episode of care rather than its end point.


Physical therapists should consider increasing their use of joint mobility and exercise interventions, as defined in this study, and reducing or discontinuing their use of the other interventions that we examined for people with sciatica. Spasm reduction interventions should not be used to improve outcome in these individuals. Future research is needed to clarify the relationship between joint mobility and exercise interventions in people with sciatica so that appropriate combinations of interventions may be identified. Several interventions (eg, McKenzie exercise, manipulation) were not represented in our study and warrant further examination.


  • Both authors provided concept/idea/research design, writing, and data collection and analysis. The authors acknowledge the assistance of the following individuals on this project: Robert E Johnson, PhD, and Misook Park, MS, Department of Biostatistics, Virginia Commonwealth University, and Dennis L Hart, PT, PhD, Director of Research and Consulting Services, Focus On Therapeutic Outcomes Inc.

    The Virginia Commonwealth University Institutional Review Board approved this study according to expedited review criteria.

    This research, in part, was presented as a poster at the Combined Sections Meeting of the American Physical Therapy Association; February 4–8, 2004; Nashville, Tenn.

  • * Focus On Therapeutic Outcomes Inc, PO Box 11444, Knoxville, TN 37939-1444.

  • SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

  • Microsoft Corp, One Microsoft Way, Redmond, WA 98052-6399.

  • § Research Triangle Institute, 3040 Cornwallis Rd, Research Triangle Park, NC 27709-2194.

  • Received August 5, 2004.
  • Accepted April 11, 2005.


View Abstract