Predictors of short-term outcome in people with a clinical diagnosis of cervical radiculopathy.

Cervical radiculopathy is a common condition with a reported annual incidence of approximately 83 per 100,000 and an increased prevalence in the fifth decade of life (203 per 100,000). (1-7) The condition may result in neck pain; however, the primary symptoms reported in this population are often upper-extremity pain, numbness, and weakness, which often result in significant functional limitations and disability. (8-11)

Several intervention strategies are commonly used in the management of cervical radiculopathy and range from conservative approaches, including physical therapy, to surgical intervention. The long-term benefits of surgical intervention in the management of cervical radiculopathy are questionable, as more than 25% of people who undergo surgery continue to experience debilitating pain at a 12-month follow-up. (1) Additionally, other studies (9,11) have shown that people treated with conservative management approaches may experience outcomes superior to those achieved with surgical intervention.

Several physical therapy interventions are commonly used in the management of cervical radiculopathy, yet it is unclear whether any physical therapy approach is effective. Although intermittent cervical traction is commonly used in the management of cervical radiculopathy, (12-16) only one clinical trial to date has attempted to isolate the effects of cervical traction. Joghataei et al (12) randomly assigned 30 people to take part in a treatment program consisting of ultrasound and exercise either with or without intermittent cervical traction for 10 sessions. The group receiving intermittent cervical traction exhibited greater improvements in grip strength (force-generating capacity), the primary outcome measure after 5 sessions; however, no statistically significant difference between groups existed at the time of discharge from physical therapy (10 treatment sessions at 3 sessions per week). (12) Unfortunately, the authors did not measure any outcomes associated with pain, function, or disability; therefore, the true benefits of cervical traction could not be ascertained.

The majority of the studies reporting outcomes in people referred for physical therapy have not used well-defined outcome measures or an eligibility criterion that ensures a homogeneous group of people with cervical radiculopathy. However, in 2 recently published case series, (14,15) a test item cluster (17) was used to identify the presence of cervical radiculopathy. In both case series, study participants were treated with a multimodal treatment approach including manual therapy, strengthening exercises, and cervical traction; the vast majority of participants exhibited improved outcomes. A more recent prospective cohort design (16) described the clinical outcomes of study participants receiving a combination of cervical thrust manipulation or muscle energy techniques, neural mobilization, end-range exercises to promote the centralization of symptoms, cervical stabilization, and "over-the-door" traction. The results demonstrated that, at the time of discharge (mean = 11 visits), 24 of 31 participants (77.4%) surpassed the minimal clinically important difference for changes in the Bournemouth Disability Questionnaire. This value increased to 93% (25 of 27 participants) at the time of a long-term follow-up (mean=8.2 months). (16)

There appears to be relative consistency regarding the clinical prognosis for people with cervical radiculopathy. (7,9) For example, Radhakrishnan and colleagues (7) reported that nearly 90% of people with cervical radiculopathy had only mild symptoms at a median follow-up of 4.9 years. However, Honet and Puri (9) found that only 70% of people with cervical radiculopathy exhibited good or excellent outcomes after a 2-year follow-up. Although the outcomes for the people in the aforementioned studies (7,9) appeared to be favorable and suggested that 70% to 90% of this population can experience dramatic improvement without surgical intervention, clinical variables for determining which people are most likely to experience successful outcomes have not been identified. The purpose of this prospective cohort study was to identify whether variables from the baseline examination or particular physical therapy interventions received could predict the clinical course in a group of people who had a clinical diagnosis of cervical radiculopathy and who were treated with physical therapy management programs.

Method

In this prospective cohort study, consecutive patients referred to any of 3 physical therapy clinics by their primary care physicians with a diagnosis of either cervical radiculopathy or neck and arm pain (symptoms extending distal to the shoulder) were screened for eligibility criteria. To be eligible, the patients had to satisfy all 4 of the criteria in the test item cluster for identifying the presence of cervical radiculopathy as defined by Wainner et al. (17) The 4 criteria included cervical rotation toward the symptomatic side of less than 60 degrees and positive findings on the Spurling test, (18) the upper-limb tension test, (19) and the cervical distraction test. Specific operational definitions for each procedure are provided in Appendix 1. Wainner et al (17) reported that, in comparison with a reference standard of neuro-diagnostic testing, the presence of these 4 findings was associated with a positive likelihood ratio (+LR) of 30.3 (sensitivity=.24; 95% confidence interval [CI] = .005, 0.43; specificity=.99; 95% CI=.97, 1.0) for detecting cervical radiculopathy. Because this was a descriptive outcome study, no exclusion criteria were established.

Between February 2004 and September 2006, 141 consecutive patients were screened for possible eligibility criteria. A total of 101 patients met the eligibility criteria; 5 of them did not return for a reexamination visit. As a result, there were no follow-up measurements for these patients; therefore, they were excluded from the analysis. Ninety-six patients, with a mean age of 50.8 years (SD=9.5, range=27-76) (64% women), satisfied the eligibility criteria and were included in the analysis. The mean number of visits for all patients was 6.4 (SD=1.7, range=4-13), with a mean duration of 28 days (SD=9, range=l0-56) between the initial examination and the final follow-up treatment session.

To be included in the present study, the patients were required to complete serf-report measures at the initial examination and follow-up self-report measures at either a reexamination visit or discharge. In addition, the patients had to read and understand English well enough to complete the self-report measures.

This study qualified for exempt status from the institutional review boards at Concord Hospital, SOAR Physical Therapy, and Franklin Pierce College. All subjects signed informed consent forms approved by the appropriate aforementioned institutional review boards.

Therapists

A total of 10 physical therapists (mean age=38.3 years, SD=7.5, range=27-50) participated in recruiting subjects. All therapists were provided a comprehensive manual of standard procedures with the operational definitions of each examination procedure used in this study. The purpose of the manual of standard procedures was to standardize the examination procedure only; thus, no details regarding the selection of treatment interventions were included. All therapists underwent a 30-minute training session provided by the principal investigator for the examination techniques used in this study. Interventions were selected solely at the discretion of the treating physical therapists.

Participating therapists had a mean of 11.6 years (SD=8.7, range=l-25) of clinical experience. Nine of the 10 participating therapists treated patients on a full-time basis, and 2 therapists were board certified in orthopedics by the American Board of Physical Therapy Specialties. When asked about their experience with manual therapy techniques, 1 therapist reported less than 1 year of experience, 2 reported 1 to 3 years of experience, 3 reported 3 to 5 years of experience, 1 reported between 3 and 10 years of experience, and 3 reported more than 10 years of experience. None of the therapists had undergone formal residency or fellowship training in manual therapy techniques prior to data collection. However, all therapists reported that they had participated in at least 20 hours of continuing education training in manual therapy techniques directed at the spine.

Examination

All subjects completed several self-report measures before a detailed history taking and physical examination were performed by a physical therapist. The self-report measures included the Neck Disability Index (NDI), the Patient-Specific Functional Scale (PSFS), and the Numeric Pain Rating Scale (NPRS). The NDI contains 10 items, 7 related to activities of daily living, 2 related to pain, and 1 related to concentration. (20) Each item is scored from 0 to 5, and the total score is expressed as a percentage, with higher scores corresponding to greater disability. The NDI has demonstrated moderate test-retest reliability (intraclass correlation coefficient [ICC]=.68; 95% CI=.03, .90) and has been shown to be a valid health outcome measure for a population of people with cervical radiculopathy. (21)

The PSFS requires people to list 3 activities that are difficult to perform as a result of their symptoms, injury, or disorder. (22) The people rate each activity on a scale of 0 to 10, with 0 representing the inability to perform the activity and 10 representing the ability to perform the activity as well as they could prior to the onset of symptoms. (23) The final PSFS score is determined by averaging the 3 activity scores. The PSFS was developed by Stratford et al (23) in an attempt to devise a standardized measure for recording a patient's perceived level of disability across a variety of conditions. The PSFS has been shown to exhibit validity as well as test-retest reliability (ICC=.82; 95% CI=.54, .93) for a population of people with cervical radiculopathy. (21)

The NPRS is used to measure a person's level of pain. People are asked to indicate the intensities of current, best, and worst levels of pain over the preceding 24 hours by using an 11-point scale ranging from 0 ("no pain") to 10 ("worst pain imaginable"). (24) The average of the 3 ratings is used to represent the person's level of pain over the preceding 24 hours. This method of measuring a person's pain levels was recently used by Cleland et al (21) for a group of people with cervical radiculopathy and exhibited test-retest reliability (ICC=.63; 95% CI=.28, .96). The subjects in the present cohort study were not asked to separately identify the magnitude of pain in the neck and upper extremity; thus, only one value was calculated for the NPRS for each subject, as in the proceedings described by Cleland et al. (21) The NPRS has been shown to be significantly correlated (P<.001) with the PSFS for a population of people with cervical radiculopathy. (21)

The standardized history consisted of age, sex, past medical history, location of symptoms (with the use of a body diagram), duration and nature of symptoms, relieving or aggravating activities, medication usage, and prior episodes of neck and arm pain. All data were collected with standardized data collection forms to determine whether the findings differed between subjects who experienced short-term successful outcomes and those who did not.

The physical examination consisted of the previously described tests and measures, which are part of the test item cluster for identifying the presence of cervical radiculopathy (Appendix 1). (17) In addition, all physical therapists performed a postural assessment, (25) a neurological assessment (myotomes, dermatomes, and reflexes) (26) (Appendix 2), an assessment of cervical and thoracic active range of motion and symptom responses, (17) segmental mobility testing of the cervical spine (mobility of the segment in a transverse plane), spring testing of the cervical and thoracic (T1-T9) spine in a postero-anterior direction, (27) and examination of deep neck flexor (28) and scapulothoracic (25) muscle strength.

The operational definitions for the postural assessment used in this study were as follows: the subject was identified as having a forward head if the subject's external auditory meatus was anteriorly deviated (anterior to the lumbar spine), (25) and the shoulders were identified as protracted if the acromion was noted to be anteriorly deviated (anterior to the lumbar spine). (25) The examiners were instructed to identify the contour of the spine for the following groups of segments: C7 to T2 (cervicothoracic junction), T3 to T5, and T6 to T10. Each group of segments was recorded as being normal (no deviation) or having excessive kyphosis or diminished kyphosis. (29) Excessive kyphosis was defined as an increase in the convexity of the thoracic spine, and diminished kyphosis was defined as a flattening of the convexity of the thoracic spine (for each group of segments). (29)

A reliability analysis for the aforementioned tests and measures was not performed in this study. However, clinicians were instructed to perform the physical examination in a manner identical to that in a previous study in which the reliability data for these tests and measures were reported to range from poor to substantial. (30)

Interventions

Because no evidence exists to suggest the best practice for this population, no specific guidance was provided to clinicians with regard to what interventions were most appropriate for the subjects. Therapists were instructed to select interventions on the basis of their own clinical decision making, and their intervention choices were recorded on a standardized treatment data collection form at each visit (Supplemental Appendix, available online only at www.ptjournal.org).

Reexamination and Follow-up

At reexamination visits (every 5 or 6 visits) and at discharge, subjects again completed the self-report measures. The last set of self-report measures collected from each subject was used for data analysis. For example, when a subject completed the measures at a reexamination visit and then again at discharge, only the measures completed at discharge were used for data analysis. Conversely, when a subject completed the measures at a reexamination visit, attended 3 more physical therapy sessions, and then did not return for the final treatment session, the measures completed at that reexamination visit were used for data analysis. Additionally, interventions performed for that subject after the reexamination visit were not entered into the database. This system was used in an attempt to include all subjects attending physical therapy sessions for the management of cervical radiculopathy rather than omitting those who did not return for the final treatment session.

Categorizing Interventions

Because the most effective interventions and dose responses for these interventions in a population of people with cervical radiculopathy are unknown, subjects were categorized as having received a particular intervention if they received that intervention in at least 50% of the sessions that they attended. The 50% criterion was used in a recent study investigating the validity of a classification system for the management of neck pain. (31) Additionally, on the basis of interventions used in a few recent case studies, (14,15) we elected to determine whether subjects receiving a multimodal treatment approach including cervical traction (manual or mechanical), manual therapy, and deep neck flexor muscle strengthening exhibited a better prognosis for a successful outcome than those who did not receive this treatment strategy. Manual therapy for the analysis carried out in this cohort study could have included muscle energy techniques or thrust/ nonthrust mobilization/manipulation techniques directed at the cervical or thoracic spine.

Classifying Subjects as Experiencing Short-Term Successful Outcomes

Subjects were classified as having experienced successful outcomes if, at the time of discharge from physical therapy or at their last reexamination, they surpassed the minimal clinically important change (MCIC) for all outcome measures, including the NDI, PSFS, NPRS, and the Global Rating of Change (GROC). It has been reported that the MCIC for the NDI in a population of people with cervical radiculopathy is 7 points, whereas that for the PSFS in the same population has been shown to be 2 points. (21) The MCIC has not been specifically identified for people with cervical radiculopathy; hence, we decided to use the reported MCIC of 2 points derived from a population of people with spinal pain. (32)

The GROC is a 15-point global rating scale ranging from -7 ("a very great deal worse") to 0 ("about the same") to + 7 ("a very great deal better"). (33) Intermittent descriptors of worsening and improving are assigned values from -1 to -6 and from +1 to +6, respectively. (34,35) It has been reported that scores of +4 and +5 are indicative of moderate changes in a person's status and that scores of +6 and +7 indicate large changes in a person's status. (33) It was determined that subjects who rated their perceived recovery on the GROC as "a very great deal better," "a great deal better," or "quite a bit better" (ie, a score of + 5 or greater) at a reexamination visit would be considered to have experienced short-term successful outcomes.

Data Analysis

Subjects were dichotomized as "successful" and "unsuccessful" on the basis of achieving the MCIC for the NDI, NPRS, PSFS, and GROC at the time of a reexamination visit. Individual variables from self-report measures, the history, and the physical examination as well as the number of physical therapy visits, duration of physical therapy, and interventions received were tested for univariate relationships with the reference criteria by use of independent sample t tests for continuous variables and chi-square tests for categorical variables. Additionally, the multimodal treatment approach including cervical traction (manual or mechanical), manual therapy, and deep neck flexor muscle strengthening was tested for univariate relationships on the basis of previous case studies (14,15) that offered preliminary evidence for the effectiveness of this type of approach. Variables with a significance level of less than .10 were retained as potential predictor variables. (36) For continuous variables with significant univariate relationships, sensitivity and specificity values for identifying subjects who experienced successful outcomes were calculated for all possible cutoff points and then plotted as a receiver operating characteristic curve. (37) The point on the curve nearest the upper left-hand corner represented the value with the best diagnostic accuracy, and this point was selected as the cutoff value defining a positive test, indicating that a subject was likely to experience a short-term successful outcome. (37) Sensitivity, specificity, +LR, and negative likelihood ratio values were calculated for potential predictor variables.

Additionally, all potential predictor variables were entered into a stepwise logistic regression model to determine the most accurate set of variables for predicting success. A significance level of greater than .10 was required for removal from the equation to minimize the likelihood of excluding potentially helpful variables. (36) For prognostic purposes, variables retained in the regression model were used as the most optimal cluster of variables for predicting optimal outcomes for subjects in physical therapy management programs for cervical radiculopathy. Odds ratios (ORs) and corresponding 95% CIs were calculated for all variables that were retained in the model. Additionally, change scores between groups (successful versus unsuccessful) for the NDI, PSFS, NPRS, and GROC were calculated and analyzed with independent t tests. Data analysis was performed with the SPSS version 14.0 statistical software package. *

Results

Subject demographics and initial baseline variables from the subject history and self-report measures for the entire sample as well as for the successful and unsuccessful groups are shown in Table 1. The baseline measurements for the neurological examination (myotomes, dermatomes, and reflexes) for the entire sample as well as for both groups are shown in Table 2. Ninety-one subjects (95%) exhibited positive findings for at least one dermatome, myotome, or reflex at one segmental level (spinal nerve). Forty-seven subjects (49%) had at least 2 positive findings at the same nerve root level, and 7 subjects (7%) had 3 positive findings at the same nerve root level. Fifty subjects surpassed the MCIC for the NDI, PSFS, NPRS, and GROC at re-examination and were categorized as having achieved successful outcomes. Forty-six subjects did not meet this threshold and were categorized as unsuccessful. The mean numbers of days between the initial examination and reexamination were 28.8 (SD = 10) and 27.3 (SD=7.9) (P=.46) for the successful and unsuccessful groups, respectively. The percentage of subjects in each group receiving interventions is shown in Table 3. Baseline, reexamination, and change scores for the NDI, PSFS, NPRS, and GROC are shown in Table 4.

The 8 potential predictor variables (for predicting success) (Tab. 5) that exhibited a significance level of less than. 10 were entered into the logistic regression model. The cutoff values determined from the receiver operating characteristic curves for the continuous variables are shown in Table 5. The accuracies (and 95% CIs) of all 8 variables are shown in Table 5. The + LR values ranged from 1.3 to 2.2. The following 4 variables were retained in the final regression model for predicting short-term success: age of less than 54 years (OR=5.3; 95% CI=1.6, 16.9); dominant arm is not affected (OR=4.1; 95% CI=1.3, 12.5); looking down does not worsen symptoms (OR=3.3; 95% CI=1.1, 10.0); and multimodal treatment including manual therapy, cervical traction, and deep neck flexor muscle strengthening for at least 50% of visits (OR=5.4; 95% CI= 1.9, 15.2). The cumulative adjusted coefficient of determination (Nagelkerke [R.sup.2]) for the final model was .45 (P<.001). These 4 variables comprised the most parsimonious combination of predictors for identifying subjects who had cervical radiculopathy, who were in physical therapy management programs, and who were most likely to achieve short-term success.

Accuracies were calculated for the variables that were present (Tab. 6). The pretest probability for the likelihood of short-term success was 53%. When 3 of the 4 variables were present, the +LR was 5.2 (95% CI = 2.4, 11.3) and the posttest probability of success increased to 85%. When all 4 variables in the final model were present, the +LR was 8.3 (95% CI=1.9, 63.9) and the posttest probability of success was 90%.

Discussion

To the best of our knowledge, the present study provides data on the largest group of people referred for physical therapy for the management of cervical radiculopathy. We have presented the clinical course for this group of subjects as well as the individual physical therapy interventions received by the subjects in this cohort. The most common intervention used was active cervical range of motion followed by mechanical traction. Clinicians were not instructed to use specific interventions because evidence to support the effectiveness of physical therapy interventions for the treatment of cervical radiculopathy remains preliminary and was primarily derived from case studies (14,15) and a cohort design. (16) However, it is interesting that not all of the subjects received traction even when an inclusion criterion was a positive finding on the cervical distraction test, suggesting that traction might be a beneficial intervention. Although there are preliminary data supporting the use of mechanical traction, (12,38) perhaps clinicians elected to not use traction on the basis of recent guidelines suggesting that the effectiveness of this approach is unclear. (39)

The study revealed 8 predictor variables that identified people who were referred for physical therapy for cervical radiculopathy and who were likely to exhibit short-term success. Three of these variables were obtained from the subject history (age of <54 years, dominant arm is not affected, and looking down does not worsen symptoms), 1 was obtained from the clinical examination (>30[degrees] of cervical flexion), and 4 were associated with the interventions received (receiving mechanical traction, receiving thoracic spine thrust manipulation, not receiving soft-tissue mobilization, and a multimodal approach). As in the present study, age (women over 50 years of age) was previously shown to be a predictor of poor outcomes in people with cervical radiculopathy. (40) It is also possible that people who exhibited greater cervical flexion that did not exacerbate their symptoms had a less severe case of cervical radiculopathy, as it is believed that forward flexion might cause a cervical disk to bulge posteriorly and compress a nerve root. (6,41)

The aforementioned variables individually exhibited likelihood ratios between 1.3 and 2.2, which resulted in a relatively small shift in probability that a subject would be likely to experience a successful outcome. (42) However, the logistic regression analysis identified a parsimonious set of predictor variables that maximized the ability to identify a subject likely to experience a successful outcome: age of less than 54 years; dominant arm is not affected; looking down does not worsen symptoms; and multimodal treatment including manual therapy, cervical traction, and deep neck flexor muscle strengthening for at least 50% of visits. When all 4 variables were present, the +LR was 8.3 and the posttest probability of success was 90%. However, the 95% CI for this +LR was wide (1.9, 63.9). With a smaller 95% CI, a clinician might achieve better accuracy in predicting whether a subject would be likely to have a short-term successful outcome if only 3 of the 4 variables were present (posttest probability=85%). That is, if 3 of the 4 previously mentioned variables were present, a clinician could predict successful outcomes in 85% of subjects referred for physical therapy for cervical radiculopathy.

Interestingly, none of the interventions in isolation was maintained in the model as a predictor of outcome, yet a multimodal approach including cervical traction, manual therapy, and deep neck flexor muscle exercises was included in the final subset of predictors. These findings coincide with those of 2 recent case series (14,15) and a prospective cohort design, (16) which reported dramatic reductions in disability following a management strategy that included both cervical traction and manual therapy interventions. Although it appears that a multimodal treatment approach including cervical traction, manual therapy, and deep neck flexor muscle exercises may be associated with positive outcomes, an optimal dose response cannot be determined from the present study because no control was exerted over the interventions used by the treating therapists. For example, clinicians selected different forces and durations of mechanical traction for the subjects, and the numbers of manual therapy techniques used for the subjects varied. Additionally, we did not measure the length of time for which some interventions were used. Therefore, inferences regarding the effectiveness of individual versus multimodal treatment approaches cannot be made from the present study. Future randomized clinical trials should be performed to further investigate the effectiveness of such approaches.

Persson and colleagues (43) reported that outcomes associated with physical therapy were as good as those associated with surgical intervention in the management of cervical radiculopathy. It seems sensible that people would select the most conservative approach to treatment prior to agreeing to surgical intervention, which often has mixed and disappointing results. (9,11)

In the present prospective cohort study, 53% of subjects referred for physical therapy for cervical radiculopathy exhibited successful outcomes. This value is lower than the values reported by Radhakrishnan et al (7) and Honet and Puri, (9) although the percentages of patients achieving positive outcomes with conservative care were recorded at a long-term folow-up in each of those studies (4.9 and 2 years, respectively). The short-term follow-up in our study is a limitation because the subjects might have improved simply over the passage of time rather than because of the influence of any other variable, and we cannot be certain that the same outcomes would have existed at a long-term follow-up. Honet and Puri (9) reported that the percentages of patients experiencing positive outcomes diminished between the completion of conservative care and the 2-year follow-up, suggesting that spontaneous recovery might not be the reason why they found higher percentages of positive outcomes than we did.

Another reason for the differences in outcomes reported in the present study and those of Radhakrishnan et al (7) and Honet and Puri (9) may be directly related to the operational definitions used to identify positive outcomes. Radhakrishnan et al (7) arbitrarily categorized patients as having positive outcomes if they were normal or mildly incapacitated (symptoms producing slight incapacity but not preventing ordinary everyday activities), as determined through medical record reviews. Honet and Puri (9) also identified positive outcomes through medical record reviews and classified patients as having good outcomes when the quantitative findings decreased, yet the patients did not need to show a full obliteration of qualitative symptoms.

We prospectively collected data on subjects, identified positive outcomes by using reliable and valid methods, and set the cutoff rather high by requiring subjects to have surpassed the MCIC for 4 outcome tools: NDI, NPRS, PSFS, and GROC.

Additionally, the successful group experienced change scores that were significantly higher than those experienced by the unsuccessful group. All subjects in the present study exceeded a change in outcome measures that surpassed the MCIC for the NDI, PSFS, and NPRS. If we had reduced the criteria for success, then the results of the present study might have been similar to those reported by Radhakrishnan et al (7) and Honet and Puri. (9)

Besides the short-term follow-up, there are other limitations of the present study. The use of the test item cluster described by Wainner et al (17) for the inclusion of cervical radiculopathy may in itself be a limitation. The study by Wainner et al (17) was the first step in the development of a decision rule, and follow-up studies are required to validate the findings. (44) Wainner et al (17) also recruited a cohort of patients who were referred for electrodiagnostic testing, perhaps suggesting that this cluster captured patients with more serious cases of cervical radiculopathy. It also is possible that using the test item cluster established by Wainner et al (17) as an inclusion criterion resulted in the exclusion of subjects with radiculopathy associated with the eighth cervical nerve root. The consecutive prospective design used by Wainner et al (17) included a measurement for only one patient with expected C8 radiculopathy; hence, we cannot be confident about the diagnostic utility of the test item cluster for people with C8 nerve root involvement. Although we might have missed some subjects with C8 radiculopathy, the most common level of nerve root involvement is C7, followed by C6 (7); this level of involvement appears to be representative of our sample on the basis of the neurological examination. Additionally, findings from the neurological examination also indicated that we did obtain measurements for at least some subjects referred for physical therapy for expected C8 radiculopathy (Tab. 2).

It is possible that some unmeasured factors also might have predicted improved outcomes achieved by subjects; these include psychosocial variables that were not measured at baseline, subjects' motivation, and subjects' adherence to home exercise programs. Numerous interventions are used by physical therapists in the management of cervical radiculopathy; however, we are not able to ascertain treatment effectiveness from the present study. We recommend that further scientific investigations in the form of clinical trials be performed to determine whether a cause and effect relationship exists between individual and multimodal treatment approaches and outcomes in people with cervical radiculopathy.

Conclusion

The results suggested that a subset of predictor variables from the baseline examination and interventions received was able to accurately identify people with cervical radiculopathy likely to experience short-term successful outcomes. The study design did not allow for the identification of a cause-and-effect relationship, but it appears that intermittent cervical traction, manual therapy, and deep neck flexor muscle strengthening may be beneficial in the management of cervical radiculopathy. Future studies should investigate the effects of a multimodal intervention strategy on the outcomes of people with cervical radiculopathy.

Appendix 1.

Operational Definitions for Tests and Measures
Used to Identify Cervical Radiculopathy

Test             Reliability (95%
                 Confidence Interval)

Spurling         Kappa = .60
test (17)        (.32,.87)

Cevical          Kappa = .88
distraction      (.64, 1.0)
test (17)

Upper-limb       Kappa = .76
tension          (.51, 1.0)
test A (17)

Neck rotation
  Right          Intraclass
                   correlation
                   coefficient = .78
                   (.55, .90) (17)

  Left           Intraclass
                   correlation
                   coefficient = .77
                   (.52, .90) (17)

Test               Technique

Spurling         The subject is seated, and the neck
test (17)          is passively bent toward the
                   symptomatic side. The examiner
                   applies approximately 7 kg of
                   force through the subject's head
                   with a caudally directed force.

Cevical          The subject is supine, and the
distraction        examiner grasps under the
test (17)          subject's chin and occiput. The
                   examiner flexes the neck to the
                   subject's comfort and then
                   applies a distraction force of
                   approximately 14 kg.

Upper-limb       The subject is supine, and the
tension            examiner places the subject's
test A (17)        upper extremity into scapular
                   depression, shoulder abduction,
                   forearm supination and wrist and
                   finger extension, shoulder
                   external rotation, elbow
                   extension, and contralateral and
                   then ipsilateral cervical lateral
                   flexion.

Neck rotation    Rotation is measured with a
                   universal goniometer. The
                   subject is seated, looking directly
                   forward, with the neck in a
  Right            neutral position. The fulcrum of
                   the goniometer is placed over
                   the top of the head with the
                   stationary arm aligned with the
                   acromion and the movable arm
                   bisecting the subject's nose. The
  Left             subject is asked to rotate as far
                   as possible in each direction.

Test               Criteria for Positive
                   Test

Spurling         Reproduction of the
test (17)          subject's upper-
                   extremity symptoms.

Cevical          Reduction or resolution
distraction        of the subject's
test (17)          upper-extremity
                   symptoms.

Upper-limb       Any of the following:
tension            * Symptom
test A (17)          reproduction

                   * Greater than 10[degrees]
                     difference in elbow
                     extension from side to
                     side
                   * An increase in
                     symptoms with
                     contralateral cervical
                     side bending or a
                     decrease in symptoms
                     with ipsilateral side
Neck rotation        bending.

  Right

  Left

Appendix 2.

Criteria Provided to Clinicians for Performing the Neurological
Assessment and Reliability Values for Measures (17, 26)

                                             Reliability
                                             (9S% Confidence
Segment   Key Muscles for Myotome Testing    Interval)

C5        Deltoid: shoulder in 90[degrees]   Kappa = .62 (.28,.96)
            of abduction, resistance
            against lateral upper arm
            into adduction
C6        Biceps brachii: elbow at           Kappa = .69 (.36, 1.0)
            90[degrees] of flexion with
            forearm supinated, resistance
            against lower forearm
            into extension

          Extensor carpi radialis            Kappa = .63 (.26, 1.0)
            longus/brevis: wrist
            extended/radially deviated
            with forearm pronated,
            resistance against dorsum
            of hand into flexion/ulnar
            deviation

C7        Triceps: arm placed overhead       Kappa = .29 (0, .79)
            with elbow slightly flexed,
            resistance against forearm
            into flexion

          Flexor carpi radialis: wrist       Kappa = .23 (0, .69)
            flexed/radially deviated with
            forearm supinated, resistance
            against thenar eminence into
            extension/ulnar deviation

C8        Abductor pollicis: thumb placed    Kappa = .39 (0-80)
            in abduction, resistance
            against phalanx into adduction

T1        First dorsal interosseus: index    Kappa = .37 (0,.80)
            and middle finger separated,
            resistance against medial
            aspect of proximal phalanx of
            index finger into adduction

Segment   Dermatomal Area                    Reliability
                                             (95% Confidence
                                             Interval)

C5        Mid-deltoid                        Kappa = .67
                                             (.33, 1.0)

C6        Radial aspect                      Kappa = .28
          of second                          (0, .58)
          metacarpal/digit

C7        Dorsum of                          Kappa = .40
          third finger                       (.06, .74)

C8        Medial aspect                      Kappa = .16
          of fifth finger                    (0,.50)

T1        Medial forearm                     Kappa = .62
                                             (.28,.96)

Segment   Key Muscle for                     Reliability
          Reflex Testing                     (9S% Confidence
                                             Interval)

C5        Biceps                             Kappa = .73
          brachii                            (.38, 1.0)

C6        Brachioradialis                    Kappa = not
                                             reported

C7        Triceps                            Kappa = not
                                             reported

C8        Not applicable

T1        Not applicable

This article was received September 23, 2006, and was accepted July 26, 2007.

DOI: 10.2522/ptj.20060287

References

(1) Heckmann J, Lang J, Zobelein I, et al. Herniated cervical intervertebral discs with radiculopathy: an outcome study of conservatively or surgically treated patients. J Spinal Disord. 1999;12:396-401.

(2) Wainner R, Gill H. Diagnosis and nonoperative management of cervical radiculopathy. J Orthop Sports Phys Ther. 2000; 12:728-744.

(3) Muhle C, Bischoff L, Weinert D, et al. Exacerbated pain in cervical radiculopathy at axial rotation, flexion, extension, and coupled motions of the cervical spine. Invest Radiol. 1998;5:279-288.

(4) Ellenberg M, Honet J, Treanor W. Cervical radiculopathy. Arch Plays Med Rehabil. 1994;75:342-352.

(5) Tanaka N, Fujimoto Y, An H, et al. The anatomic relation among the nerve roots, intervertebral foramina, and intervertebral discs of the cervical spine. Spine. 2000;25:286-291.

(6) Farmer J, Wisneski R. Cervical spine nerve root compression: an analysis of neuroforaminal pressures with varying head and arm positions. Spine. 1994;19:1850-1855.

(7) Radhakrishnan K, Litchy WJ, O'Fallon WM, et al. Epidemiology of cervical radiculopathy: a population-based study from Rochester, Minnesota, 1976 through 1990. Brain. 1994;117:325-335.

(8) Benini A. Clinical features of cervical root compression C5-C8 and their variations. Neuro-Orthopedics. 1987;4:74-88.

(9) Honet J, Puri K. Cervical radiculitis: treatment and results in 82 patients. Arch Phys Med Rehabil. 1976;57:12-16.

(10) Caplan L. Management of cervical radiculopathy. Eur Neurol. 1995;35:309-320.

(11) Sampath P, Bendebba M, Davis JD, et al. Outcome in patients with cervical radiculopathy: prospective, multicenter study with independent clinical review. Spine. 1999;24:591-597.

(12) Joghataei MT, Arab AM, Khaksar H. The effect of cervical traction combined with conventional therapy on grip strength on patients with cervical radiculopathy. Clin Rehabil. 2004;18:879-887.

(13) Moeti P, Marchetti G. Clinical outcome from mechanical intermittent cervical traction for the treatment of cervical radiculopathy: a case series. J Orthop Sports Phys Ther. 2001;31:207-213.

(14) Cleland JA, Whitman JM, Fritz JM, et al. Manual physical therapy, cervical traction and strengthening exercises in patients with cervical radiculopathy: a case series. J Orthop Sports Phys Ther. 2005;35: 802-811.

(15) Waldrop MA. Diagnosis and treatment of cervical radiculopathy using a clinical prediction rule and a multimodal intervention approach: a case series. J Orthop Sports Phys Ther. 2006;36:152-159.

(16) Murphy DR, Hurwitz EL, Gregory A, et al. A nonsurgical approach to the management of patients with cervical radiculopathy: a prospective observational cohort study. J Manipulative Physiol Ther. 2006;29:279-287.

(17) Wainner R, Fritz J, Irrgang J, et al. Reliability and diagnostic accuracy of the clinical examination and patient self-report measures for cervical radiculopathy. Spine. 2003;28:52-62.

(18) Spurling RG, Scoville WB. Lateral rupture of the cervical intervertebral discs: a common cause of shoulder and arm pain. Surg Gynecol Obstet. 1944;78:350-358.

(19) Elvey RL. The investigation of arm pain: signs of adverse responses to the physical examination of the brachial plexus and related tissues. In: Boyling JD, Palastanga N, eds. Grieve's Modern Manual Therapy. 2nd ed. New York, NY: Churchill Livingstone; 1994:577-585.

(20) Vernon H, Mior S. The Neck Disability Index: a study of reliability and validity. J Manipulative Physiol Ther. 1991;14: 409-415.

(21) Cleland JA, Fritz JM, Whitman JM, et al. The reliability and construct validity of the Neck Disability Index and Patient Specific Functional Scale in patients with cervical radiculopathy. Spine. 2006;31:598-602.

(22) Chatman AB, Hyams SP, Neel JM, et al. The Patient-Specific Functional Scale: measurement properties in patients with knee dysfunction. Phys Ther. 1997;77:820-829.

(23) Stratford P, Gill C, Westaway M, et al. Assessing disability and change on Individual patients: a report of a patient-specific measure. Physiother Can. 1995;47:258-263.

(24) Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain. 1986; 27:117-126.

(25) Kendall FP, McCreary EK, Provance PG. Muscles: Testing and Function. 4th ed. Baltimore, Md: Williams & Wilkins; 1993.

(26) Magee D. Orthopedic Physical Assessmerit. 4th ed. Philadelphia, Pa: Saunders; 2002.

(27) Childs JD, Whitman JM, Fritz JM, et al. Physical Therapy for the Cervical Spine and Temporomandibular Joint. La Crosse, Wis: Orthopaedic Section of the American Physical Therapy Association Inc; 2003:8-63.

(28) Harris KD, Heer DM, Roy TC, et al. Reliability of a measurement of neck flexor muscle endurance. Phys Ther. 2005;85: 1349-1355.

(29) Griegel-Morris P, Larson K, Mueller-Klaus K, et al. Incidence of common postural abnormalities in the cervical, shoulder, and thoracic regions and their association with pain in two age groups of healthy subjects. Phys Ther. 1992;72:425-431.

(30) Cleland JA, Childs JD, Fritz JM, et al. Interrater reliability of the history and physical examination in patients with mechanical neck pain. Arch Phys Med Rehabil. 2006;87:1388-1395.

(31) Fritz JM, Brennan GP. Preliminary examination of a proposed treatment-based classification system for patients receiving physical therapy interventions for neck pain. Phys Ther. 2007;87:513-524.

(32) Childs JD, Piva SR, Fritz JM. Responsiveness of the Numeric Pain Rating Scale in patients with low back pain. Spine. 2005;30:1331-1334.

(33) Jaeschke R, Singer J, Guyatt G. Measurement of health status: ascertaining the minimal clinically important difference. Control Clin Trials. 1989;10:407-415.

(34) Koes BW, Bouter LM, van Mameren H, et al. The effectiveness of manual therapy, physiotherapy, and treatment by the general practitioner for nonspecific back and neck complaints: a randomized clinical triM. Spine. 1992;17:28-35.

(35) Koes BW, Bouter LM, van Mameren H, et al. Randomised clinical trial of manipulative therapy and physiotherapy for persistent back and neck complaints: results of one year follow up. BMJ. 1992;304: 601-605.

(36) Freedman DA. A note on screening regression equations. Am Stat. 1983;37: 152-155.

(37) Deyo RA, Centor RM. Assessing the responsiveness of functional scales to clinical change: an analogy to diagnostic test performance. J Chronic Dis. 1986;39: 897-906.

(38) Constantoyannis C, Konstantinou D, Kourtopoulos H, et al. Intermittent cervical traction for cervical radiculopathy caused by large-volume herniated disks. J Manipulative Physiol Ther. 2002;25:188-192.

(39) Brosseau L, Tugwell P, Wells GA. Philadelphia Panel Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions for Neck Pain. Phys Ther. 2001;81:1701-1717.

(40) British Association of Physical Medicine. Pain in the neck and arm: a muiticentre trial of the effects of physiotherapy. Br Med J. 1966;1:253-258.

(41) McKenzie RA. Cervical and Thoracic Spine: Mechanical Diagnosis and Therapy. Minneapolis, Minn: Orthopaedic Physical Therapy Products; 1990.

(42) Jaeschke R, Guyatt GH, Sackett DL. Users' guides to the medical literature, m: how to use an article about a diagnostic test, B: What are the results and will they help me in caring for my patients? The EvidenceBased Medicine Working Group. JAMA. 1994;271:703-707.

(43) Persson L, Moritz U, Brantdt L, et al. Cervical radiculopathy: pain, muscle weakness and sensory loss in patients with cervical radiculopathy treated with surgery, physiotherapy or cervical collar. Eur Spine J. 1997;6:256-266.

(44) Childs JD, Cleland JA. Development and application of clinical prediction rules to improve decision making in physical therapist practice. Phys Ther. 2006;86: 122-131.

Dr Cleland, Dr Fritz, and Dr Whitman provided concept/idea/research design, writing, data collection, and project management. Dr Cleland provided data analysis. Dr Cleland and Ms Heath provided subjects and facilities/equipment. All authors provided consultation (including review of manuscript before submission). The authors thank Britt Smith, PT, for his efforts on this project as well as all of the data collection therapists.

JA Cleland, PT, PhD, OCS, FAAOMPT, is Associate Professor, Department of Physical Therapy, Franklin Pierce College, 5 Chenell Dr, Concord, NH 03301 (USA); Research Coordinator, Rehabilitation Services, Concord Hospital, Concord, NH; and Faculty, Manual Physical Therapy Fellowship Program, Regis University, Denver, Colo. Address all correspondence to Dr Cleland at: [e-mail].

JM Fritz, PT, PhD, ATC, is Associate Professor, Division of Physical Therapy, University of Utah, Salt Lake City, Utah, and Clinical Outcomes Research Scientist, Intermountain Health Care, Salt Lake City, Utah.

JM Whitman, PT, DSc, OCS, FAAOMPT, is Assistant Faculty, Department of Physical Therapy, Regis University, and Faculty, Manual Physical Therapy Fellowship Program, Regis University.

R Heath, PT, is Physical Therapist, Rehabilitation Services, Concord Hospital.

[Cleland JA, Fritz JM, Whitman JM, Heath R. Predictors of short-term outcome in people with a clinical diagnosis of cervical radiculopathy. Phys Ther. 2007;87:1619-1632.]

Table 1.
Demographics, Baseline Self-report Variables,
and Baseline Characteristics of Subjects

Variable (a)                     All         Successful
                               Subjects       Subjects
                                (N=96)         (n=50)

Age, y, [bar.X] (SD)         50.8 (9.5)     48.2 (7.8)
Sex, no. (%)
  Women                        61 (64)         34 (68)
  Men                          35 (36)         16 (32)
Duration of
  symptoms, d, [bar.X] (SD)  60.2 (43.5)     58.0 (48.7)
NDI score, [bar.X] (SD)      28.6 (9.2)     29.2 (9.0)
PSFS score, [bar.X] (SD)      4.20 (1.3)      4.0 (1.2)
NPRS score, [bar.X] (SD)      6.5 (2.4)       6.8 (1.6)
Prior episode
  of symptoms, no. (%)
  of subjects                  27 (28)         12 (24)
Dominant arm is affected,
  no. (%) of subjects          35 (36)         13 (26)
No. of physical
  therapy visits, [bar.X]
  (SD)                        6.4 (1.7)       6.4 (2.0)
Duration of physical
  therapy, d, [bar.X] (SD)   28.1 (9.1)     28.8 (10.1)
Symptom location,
    no. (%) of subjects
  Neck                         64 (67)         32 (64)
  Shoulder                     56 (58)         25 (50)
  Upper arm                    72 (75)         35 (70)
  Forearm                      65 (68)         35 (70)
  Hand                         40 (42)         20 (40)
Aggravating factor,
    no. (%) of subjects
  Looking up                   69 (72)         39 (78)
  Looking down                 40 (42)         16 (32)
  Looking right                52 (54)         28 (56)
  Looking left                 58 (60)         32 (64)
  Driving                      66 (69)         32 (64)
  Reading                      46 (48)         25 (50)
  Using a computer             40 (42)         20 (40)

Variable (a)                 Unsuccessful         P
                               Subjects
                                (n=46)

Age, y, [bar.X] (SD)         53.7 (10.3)      .004 (b)
Sex, no. (%)                                   .40 (c)
  Women                        27 (59)
  Men                          19 (41)
Duration of
  symptoms, d, [bar.X]
  (SD)                       62.5 (37.4)       .61 (b)
NDI score, [bar.X] (SD)      28.1 (9-5)       .57 (b)
PSFS score, [bar.X] (SD)      4-5 (1.3)       0.18 (b)
NPRS score, [bar.X] (SD)      6.3 (3-0)        .29 (b)
Prior episode
  of symptoms, no. (%)
  of subjects                  15 (33)         .37 (c)
Dominant arm is affected,
  no. (%) of subjects          22 (48)         .03 (c)
No. of physical
  therapy visits, [bar.X]
  (SD)                         6.5 (1-3)       0.94 (b)
Duration of physical
  therapy, d, [bar.X] (SD)    27.3 (7.9)       .46 (b)
Symptom location,
    no. (%) of subjects
  Neck                         32 (70)         .40 (c)
  Shoulder                     31 (67)         .12 (c)
  Upper arm                    37 (80)         .38 (c)
  Forearm                      30 (65)         .83 (c)
  Hand                         20 (43)         .68 (c)
Aggravating factor,
    no. (%) of subjects
  Looking up                   30 (65)         .18 (c)
  Looking down                 24 (52)         .06 (c)
  Looking right                24 (52)         .84 (c)
  Looking left                 26 (57)         .53 (c)
  Driving                      34 (74)         .40 (c)
  Reading                      21 (46)         .69 (c)
  Using a computer             20 (43)         .68 (c)

(a) NDI = Neck Disability Index, PSFS = Patient-Specific
Functional Scale, NPRS = Numeric Pain Rating Scale.

(b) As determined by independent sample t tests.

(c) As determined by chi-square tests.

Table 2.
Findings From Neurological Examination at Baseline
for All Subjects Included in the Analysis

                           No. (%) of Subjects

Variable                   All        Successful
                         (N = 96)      (n = 5O)

Dermatomes
  C5                     17 (18)       12 (24)
  C6                     30 (31)       15 (30)
  C7                     40 (42)       21 (42)
  C8                     11 (11)        4 (8)
  TI                     11 (11)        3 (6)
  At least 1
    dermatome
    affected             74 (77)       40 (80)
Myotomes
  C5                     9 (9.3)       7 (14)
  C6                     22 (23)        9 (18)
  C7 (b)                 27 (280       15 (30)
  C8 (c)                 15 (16)        8 (16)
  T1                     17 (18)        7 (14)
  At least 1
    myotome affected     52 (54)       30 (60)
Muscle stretch
reflexes
  C5                     20 (21)       10 (20)
  C6                     26 (27)       12 (24)
  C7                     33 (34)       20 (40)
  At least 1
    reflex affected      54 (56)       34 (68)

                           No. (%) of Subjects

Variable               Unsuccessful
                         (n = 46)     [p.sub.a]

Dermatomes
  C5                      5 (11)         .15
  C6                     15 (33)         .83
  C7                     19 (41)         1.0
  C8                      7 (15)         .34
  TI                      8 (17)         .12
  At least 1
    dermatome
    affected             34 (74)         .14
Myotomes
  C5                      2 (4)          .65
  C6                     13 (28)         .33
  C7 (b)                 12 (26)         .82
  C8 (c)                 7 (15)          .78
  T1                     10 (22)         .42
  At least 1
    myotome affected     22 (48)         .89
Muscle stretch
reflexes
  C5                     10 (22)         1.0
  C6                     14 (30)         .50
  C7                     13 (28)         .20
  At least 1
    reflex affected      20 (43)         .31

(a) As determined by chi-square tests.

(b) Deficit in either biceps brachii or extensor carpi
radialis longus/brevis muscle groups.

(c) Deficit in either triceps or flexor carpi
radialis muscle groups.

Table 3.
Specific Treatments Received by Each Group of Subjects

Treatment                       No. (%) of Subjects

                                All                   Successful
                                Subjects              (n=50)
                                Receiving
                                Treatment
                                (N=96)

Cryotherapy                     30 (31)               13 (26)
Moist heat                      33 (34)               14 (28)
Ultrasound                      16 (17)               6 (12)
Electrical stimulation          20 (21)               8 (16)
Active range of motion
  (including cervical
  retraction)                   69 (72)               36 (72)
Stretching                      52 (54)               26 (52)
Strengthening
  (scapula/upper extremity)     47 (49)               25 (50)
Strengthening
  (deep neck flexor muscles)    38 (40)               23 (30)
Aerobic activity                30 (31)               15 (30)
Functional activity             29 (30)               15 (40)
Traction, manual                34 (35)               20 (40)
Traction, mechanical            53 (55)               33 (66)
Total duration in
  min, [bar.X] (SD, range)      17.7 (2.4, 15-20)     17.8 (2.3, 15-20)
Pounds,[bar.X] (SD, range)      24.6 (2.9, 20-28)     24.3 (3.2, 20-28)
Muscle energy techniques        28 (29)               13 (26)
Nonthrust manipulation
  of the cervical spine         47 (49)               27 (54)
Nonthrust manipulation
  of the thoracic spine         40 (42)               23 (46)
Thrust manipulation
  of the cervical spine         0 (0)                 0 (0)
Thrust manipulation
  of the thoracic spine         27 (28)               18 (36)
Soft-tissue mobilization        35 (36)               14 (28)
Neural mobilization             23 (24)               13 (26)
Multimodai treatment
  including manual therapy,
  cervicaltraction, and
  deep neck flexor muscle
  strengthening for at
  least 50% of visits           43 (45)               31 (62)

                                No. (%) of Subjects

                                Unsuccessful              [p.sup.a]
                                (n=46)

Cryotherapy                     17 (37)                      .28
Moist heat                      19 (41)                      .20
Ultrasound                      10 (22)                      .27
Electrical stimulation          12 (26)                      .32
Active range of motion
  (including cervical
  retraction)                   33 (72)                      1.0
Stretching                      26 (57)                      .90
Strengthening
  (scapula/upper extremity)     22 (48)                      .84
Strengthening
  (deep neck flexor muscles)    15 (33)                      .21
Aerobic activity                15 (33)                      .83
Functional activity             14 (30)                      1.0
Traction, manual                14 (30)                      .39
Traction, mechanical            20 (43)                      .03
Total duration in
  min, [bar.X] (SD, range)      17.5 (2.6, 15-20)
Pounds,[bar.X] (SD, range)      25.2 (2.5, 20-28)
Muscle energy techniques        15 (33)                      .51
Nonthrust manipulation
  of the cervical spine         20 (43)                      .32
Nonthrust manipulation
  of the thoracic spine         17 (37)                      .41
Thrust manipulation
  of the cervical spine         0 (0)                         -
Thrust manipulation
  of the thoracic spine         9 (20)                       .04
Soft-tissue mobilization        23 (50)                      .06
Neural mobilization             10 (22)                      .64
Multimodai treatment
  including manual therapy,
  cervicaltraction, and
  deep neck flexor muscle
  strengthening for at
  least 50% of visits           12 (26)                   .001 (b)

(a) As determined by chi-square tests. -- = unable to calculate.

(b) As determined by independent t tests.

Table 4.
Reexamination and Change Scores for the Neck Disability Index,
the Patient-Specific Functional Scale, the Numeric Pain Rating
Scale, and the Global Rating of Change

                             [bar.X] (SD)

                             All            Successful    Unsuccessful
                             Subjects       Subjects      Subjects
Scale or Rating              (N=96)         (n=50)        (n=46)

Neck Disability Index
  Initial examination        28.6 (9.2)     29.2 (9.0)    28.1 (9.5)
  Reexamination              15 (10.0)      10.9 (8.5)    19.5 (9.6)
  Change score               13.9 (10.5)    18.9 (8.5)    8.4 (9.7)

Patient-Specific
Functional Scale
  Initial examination        4.2 (1.3)      4.0 (1.2)     4.5 (1.3)
  Reexamination              7.5 (2.0)      8.3 (1.1)     6.6 (2.2)
  Change score               3.4 (1.9)      4.4 (1.2)     2.4 (2.0)

Numeric Pain Rating Scale
  Initial examination        6.5 (2.4)      6.8 (1.6)     6.3 (3.0)
  Reexamination              2.8 (2.5)      2.2 (1.2)     3.5 (3.4)
  Change score               3.8 (2.1)      4.5 (1.9)     2.9 (2.3)

Global Rating of Change
  Reexamination              3.8 (1.9)      4.6 (1.3)     2.9 (2.2)

                             Difference in            P
                             Change Scores
                             (95% Confidence
Scale or Rating              Interval)

Neck Disability Index
  Initial examination
  Reexamination
  Change score               10.5 (6.8, 14.2)         < .001

Patient-Specific
Functional Scale
  Initial examination
  Reexamination
  Change score               2.0 (1.3, 2.6)           < .001

Numeric Pain Rating Scale
  Initial examination
  Reexamination
  Change score               1.6 (0.80, 2.4)          < .001

Global Rating of Change
  Reexamination              1.7 (0.96, 2.4)          < .001

Table 5.
Accuracy and 95% Confidence Intervals (CI) of Individual
Predictor Variables for Identifying Short-Term Successful
Outcomes (a)

                             Sensitivity         Specificity
Variable                      (95% CI)            (95% CI)

Age (<54 y)               0.76 (0.64, 0.89)   0.52 (0.38, 0.67)

Dominant arm is           0.74 (0.62, 0.86)   0.52 (0.38, 0.67)
not affected

Looking down              0.68 (0.55, 0.81)   0.48 (0.34, 0.62)
does not worsen
symptoms

>30[degrees] of           0.56 (0.42, 0.70)   0.59 (0.44, 0.73)
cervical flexion

Traction,                 0.66 (0.53, 0.79)   0.57 (0.42, 0.71)
mechanical

Thrust                    0.36 (0.24, 0.49)   0.80 (0.69, 0.92)
manipulation
of the
thoracic spine

No soft-tissue            0.76 (0.69, 0.83)   0.50 (0.36, 0.64)
mobilization

Multimodal treatment
  including manual
  therapy, cervical
  traction, and deep
  neck flexormuscle
  strengthening for
  at least 50%
  of visits               0.62 (0.49, 0.75)   0.72 (0.59, 0.85)

                              Positive           % Posttest
Successful Outcomes (a)      Likelihood          Probability
                                Ratio                of
Variable                      (95% CI)             Success

Age (<54 y)                1.5 (1.2, 2.1)           62.9

Dominant arm is            1.5 (1.1, 2.2)           62.9
not affected

Looking down               1.3 (0.93, 1.8)          59.5
does not worsen
symptoms

>30[degrees] of            1.4 (0.89, 2.1)          61.0
cervical flexion

Traction,                  1.5 (1.0, 2.2)           62.9
mechanical

Thrust                     1.8 (0.92, 3.7)          71.3
manipulation
of the
thoracic spine

No soft-tissue             1.5 (1.1, 2.1)           62.9
mobilization

Multimodal treatment
  including manual
  therapy, cervical
  traction, and deep
  neck flexormuscle
  strengthening for
  at least 50%
  of visits                2.2 (1-3, 3.7)           71.3

The probability of success was
calculated as positive likelihood
ratios and assumed a pretest
probability of 53%.

Table 6.
Combination of Predictor Variables Identified in the Logistic
Regression Analysis and Associated Accuracy and 95% Confidence
Intervals (CI) of Individual Predictor Variables for Identifying
Success (a)

No. of
Predeictor
Variables       Sensitivity         Specificity
Present          (95% CI)            (95% CI)

4+           0.18 (0.07, 0.29)   0.98 (0.94, 1.0)
3+           0.68 (0-55,0.81)    0.87 (0.77, 0.97)
2+           0.94 (0.87, 1.0)    0.37 (0.23, 0.51)
1+            1.0 (1.0, 1.0)     0.08 (0.01, 0.2)

No. of           Positive
Predeictor       Likelihoo               %
Variables          Ratio            Probability
Present          (95% CI)           of Success

4+            8.3 (1.9,63-9)           90.4
3+            5.2 (2.4,11-3)           85.4
2+            1.5 (1.2, 1.9)           62.9
1+            1.1 (1.0, 2.0)           55.4

(a) The probability of success was calculated as positive
likelihood ratios and assumed a pretest probability of 53%.
The 4 predictor variables were age (<54 y); dominant arm
is not affected; looking down does not worsen symptoms;
and multimodal treatment including manual therapy, cervical
traction, and deep neck flexor muscle strengthening for at
least 50% of visits.