Phys Ther. 2007 (Jan); 87 (1): 9–23 ~ FULL TEXT
Cynthia K Peterson, Jennifer Bolton, B. Kim Humphreys
University of Zürich and Orthopaedic University Hospital Balgrist,
8008 Zürich, Switzerland
BACKGROUND AND PURPOSE: To date, no studies have investigated the predictive validity of variables from the initial examination to identify patients with neck pain who are likely to benefit from thoracic spine thrust manipulation. The purpose of this study was to develop a clinical prediction rule (CPR) to identify patients with neck pain who are likely to experience early success from thoracic spine thrust manipulation.
SUBJECTS: This was a prospective, cohort study of patients with mechanical neck pain who were referred for physical therapy.
METHODS: Subjects underwent a standardized examination and then a series of thoracic spine thrust manipulation techniques. They were classified as having experienced a successful outcome at the second and third sessions based on their perceived recovery. Potential predictor variables were entered into a stepwise logistic regression model to determine the most accurate set of variables for prediction of treatment success.
RESULTS: Data for 78 subjects were included in the data analysis, of which 42 had a successful outcome. A CPR with 6 variables was identified. If 3 of the 6 variables (positive likelihood ratio=5.5) were present, the chance of experiencing a successful outcome improved from 54% to 86%.
DISCUSSION AND CONCLUSION: The clinical prediction rule (CPR) provides the ability to a priori identify patients with neck pain who are likely to experience early success with thoracic spine thrust manipulation. However, future studies are necessary to validate the rule.
From the FULL TEXT Article:
Neck pain is a common occurrence with a lifetime incidence ranging from 22% to 70%. [1, 2] Over a third of patients will develop chronic symptoms lasting more than 6 months in duration,  representing a serious health concern.  Over 50% of patients with neck pain are referred for physical therapy and comprise approximately 25% of all patients seeking physical therapy services. [5, 6] Although cervical spine thrust manipulation has been advocated as an intervention appropriate for the care of patients with neck disorders, clinicians must consider the benefits relative to the potential risks, especially vertebral artery insult. [7, 8] The lack of evidence for premanipulative screening to identify which patients may be at risk has caused some authors to suggest that serious complications, although rare, are unpredictable and that the potential benefits of cervical spine thrust manipulation do not outweigh the inherent risks. [8, 9]
Clinical experience and preliminary evidence suggest that thoracic spine thrust manipulation may be useful in the management of patients with neck pain.  The biomechanical link between the cervical spine and the thoracic spine suggest that disturbances in joint mobility in the thoracic spine may serve as an underlying contributor to the development of neck disorders. In addition, it has been demonstrated that a significant association exists between decreased mobility of the thoracic spine and the presence of patient-reported complaints associated with neck pain.  With inherently lower risk of serious complications, thoracic spine thrust manipulation might be a suitable alternative, or supplement, to cervical spine thrust manipulation. Perhaps this accounts for why some clinicians perform thoracic spine thrust manipulation rather than cervical spine thrust manipulation at much higher rates in patients with neck pain. 
Although widely used in patients with neck pain, there are currently no decision-making strategies to identify individual patients with neck pain who are most likely to benefit from thoracic spine thrust manipulation. [10, 13, 14] Classification provides a means of breaking down a larger entity into more homogenous subgroups of patients based on examination data. [15, 16] Moreover, classification is most helpful for physical therapists when it is based on signs and symptoms that match interventions to the subgroup of patients most likely to benefit from them (ie, treatment-based classification). 
Clinical prediction rules (CPRs) consist of combinations of variables obtained from self-report measures and the historical and clinical examinations and assist with subgrouping patients into specific classifications. Recently, CPRs have been shown to be useful in classifying patients with low back pain (LBP) who are likely to benefit from a particular treatment approach. [18–20] Although a treatment-based classification system for the management of neck pain has recently been proposed,  no studies have investigated the predictive validity of variables from the initial examination to identify patients with neck pain who are likely to benefit from thoracic spine thrust manipulation. Therefore, the purpose of this study was to develop a CPR to identify patients with neck pain who are likely to benefit from thoracic spine thrust manipulation based on a reference standard of patient-reported improvement.
Materials and Methods
We conducted a prospective cohort study of consecutive patients with mechanical neck pain who were referred for physical therapy at one clinical site (Rehabilitation Services, Concord, Hospital, Concord, NH). Inclusion criteria required subjects to be between the ages of 18 and 60 years, with a primary complaint of neck pain with or without unilateral upper-extremity symptoms and a baseline Neck Disability Index (NDI) score of 10% or greater. Exclusion criteria were as follows: identification of any medical “red flags” suggestive of a nonmusculoskeletal etiology of symptoms, history of a whiplash injury within 6 weeks of the examination, a diagnosis of cervical spinal stenosis, evidence of any central nervous system involvement, or signs consistent with nerve root compression (at least 2 of the following had to be diminished to be considered nerve root involvement: myotomal strength, sensation, or reflexes). All subjects reviewed and signed a consent form approved by the Institutional Review Board at Concord Hospital, Concord, NH.
Four physical therapists participated in the examination and treatment of subjects in this study. All therapists underwent a standardized training regimen, which included studying a manual of standard procedures with the operational definitions and video clips demonstrating each examination and treatment procedure used in this study. All participating therapists then underwent a 1-hour training session in which they practiced the examination and treatment techniques to ensure that all study procedures were performed in a standardized fashion. Prior to participating in data collection, therapists were visually observed by the principal investigator as being able to successfully perform all examination and treatment procedures on a patient with neck pain. Participating therapists had a mean of 12.3 years (SD=10.0, range=3–23) of clinical experience.
Subjects provided demographic information and completed a variety of self-report measures, followed by a standardized history and physical examination at baseline. Self-report measures included a body diagram to assess the distribution of symptoms,  a numeric pain rating scale (NPRS),  the NDI,  and the Fear-Avoidance Beliefs Questionnaire (FABQ). Subjects recorded the location of their symptoms on the body diagram to determine the most distal extent of their symptoms. 
The FABQ was used to quantify the subjects’ beliefs about the influence of work and activity on their neck pain.25 The FABQ consists of a work (FABQW) subscale and a physical activity (FABQPA) subscale, both of which have been shown to exhibit a high level of test-retest reliability.  The FABQW subscale has been shown to exhibit predictive validity in the identification of patients with LBP who are likely to respond to spinal manipulation, [19, 20] but the predictive validity for patients with neck pain is unknown. For this study, the FABQ was modified to replace the word “back” with “neck.”  Finally, the NDI was used to capture the subjects’ perceived level of disability as a result of their neck pain. 
The historical examination included questions regarding the mode of onset, nature and location of symptoms, aggravating and relieving factors, and prior history of neck pain. The physical examination began with a neurological screen  followed by postural assessment.  The operational definitions for postural assessment used in this study were as follows: a subject was identified as having a forward head if the subject’s external auditory meatus was anteriorly deviated (anterior to the lumbar spine),  and the shoulders were identified as protracted if the acromion was noted to be anteriorly deviated (anterior to the lumbar spine).29 The examiners were instructed to identify the contour of the spine for the following groups of segments: C7 through T2 (cervicothoracic junction), T3 through T5, and T6 through T10. Each group was recorded as normal (no deviation), as having excessive kyphosis, or as having diminished kyphosis. ,Excessive kyphosis was defined as an increase in the convexity, and diminished kyphosis was defined as a flattening of the convexity of the thoracic spine (at each segmental group). 
The clinician next measured cervical range of motion and symptom response  and assessed the length  and strength (force-generating capacity)  of the muscles of the upper quarter and endurance of the deep neck flexor muscles.  The amount of motion and symptom response were recorded for both segmental mobility testing  of the cervical spine and spring testing33 of the cervical spine and thoracic spine (C2–T9).
The physical examination culminated with a number of special tests typically performed in the examination of patients with neck pain, including the Spurling test,  Roos test,  Neck Distraction Test,  and Upper Limb Neurodynamic Test.  Specific operational definitions for each test and criteria defining a positive test are presented in the Appendix.
Of the 80 subjects who were enrolled in the study, 22 underwent a second examination by an additional therapist who was blind to the findings of the first clinician. The 22 subjects who underwent a second evaluation were selected based on the availability of a second clinician to perform the examination. The reliability analysis was performed to evaluate the reliability of the identified potential predictor variables.
All subjects received a standardized treatment regimen, regardless of the results of the clinical examination, because treatment outcome served as the reference criterion.  Each subject received 3 different thrust manipulation techniques directed at the thoracic spine during each session: a seated “distraction” manipulation, a supine upper thoracic spine manipulation, and a middle thoracic spine manipulation. The first manipulation performed was the “distraction” manipulation. The subject was seated, and the therapist placed his or her upper chest at the level of the subject’s middle thoracic spine and grasped the subject’s elbows. A high-velocity distraction thrust was performed in an upward direction (Figure 1).
The upper thoracic spine manipulation was performed with the subject positioned supine and clasping his or her hands across the base of the neck. The therapist used his or her manipulative hand to stabilize the inferior vertebra of the motion segment (the therapist was instructed to target between T1 and T4 with this technique) and used his or her body to push down through the subject’s arms to perform a high-velocity, low-amplitude thrust (Figure 2).
The middle thoracic spine manipulation
was performed in the identical
fashion as the upper thoracic technique,
except the subject grasped
the opposite shoulder with his or
her hands and the therapist was
instructed to target between T5 and
T8 with the thrust (Figure 3). Immediately
after performing a manipulation,
the treating therapist recorded
whether a “pop” was heard. Regardless
of the presence of a “pop,” the
therapist again performed the identical
manipulation technique. Therefore,
each subject received 6 manipulations
per treatment session
Following the manipulation techniques,
all subjects were instructed
in a cervical-range-of-motion (CROM)
exercise (10 repetitions performed
3– 4 times daily)  (Figure 4) and were
advised to maintain their usual activity
within the limits of pain. The
CROM exercise consisted of the subject
placing his or her fingers over
the manubrium and placing his or
her chin on the fingers. The subject
was instructed to rotate to one side
as far as possible and return to neutral.
This was performed alternately
to both sides within pain tolerance.
The first treatment session was
always performed on the day of the
initial examination, and the subject
was scheduled for a follow-up visit
within 2 to 4 days.
The global rating of change (GROC)
served as the reference criterion for
establishing a successful outcome.
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”). 
Intermittent descriptors of worsening
or improving are assigned values
from 1 to 7 and 1 to 7,
respectively. [41, 42] It has been
reported that scores of 4 and 5
are indicative of moderate changes
in patient status and scores of 6
and 7 indicate large changes in
patient status.  It was determined a
priori 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 the
second session were categorized as
having a successful outcome, and
their participation in the study was
A high threshold for determining a
successful outcome was established
to maximize the likelihood that the
clinical outcome was attributable to
meaningful improvements in symptoms
due to the intervention as
opposed to the passage of time. Subjects whose scores on the GROC did
not exceed the 5 cutoff at the second
session again received the thrust
manipulations as in the first treatment
and were scheduled for a
follow-up within 2 to 4 days. At the
start of the third session, subjects
again completed the GROC and
were judged to have a successful
outcome based on the previously
described criterion. If the subjects
still did not meet the threshold for
success, they were categorized as
having a nonsuccessful outcome. At
this point, their participation in the
study was complete, and further
treatment was administered at the
discretion of their therapist.
In contrast to other studies identifying
predictor variables for treatment
success in patients with LBP, [18, 19] we
elected to use perceived recovery
rather than a perceived level of disability
to determine success as the
GROC. This decision is based on the
fact that the GROC is considered to
be a valid reference standard for
identifying clinically important
change. [43–45] Perceived recovery also
was used as the reference criterion
because the NDI has been criticized
for not adequately capturing low levels
of disability and for not being
responsive to small, but clinically
important, changes in patients with
low levels of initial disability.  In
addition, a measure of success rate
based on patient’s perceived recovery
has previously been used in trials
of patients with neck pain and has
been shown to be responsive to
changes with physical therapy management
programs. [42, 46]
Subjects were dichotomized as having
a successful outcome or as having
a nonsuccessful outcome based
on the treatment response, as indicated
on the GROC. The mean NDI
and NPRS change scores (and 95%
confidence intervals [CIs]) were
calculated for the both groups and
analyzed using an independent t
test to determine whether a difference
existed between groups. Individual
variables from self-report
measures, the history, and the
physical examination were tested
for univariate relationship with the
GROC reference criterion using
independent-samples t tests for
continuous variables and chisquare
tests for categorical variables.
Variables with a significance
level of P<.10 were retained as
potential prediction variables. 
This significance level was selected
to increase the likelihood that no
potential predictor variables would
For continuous variables with a significant
univariate relationship, sensitivity
and specificity values were
calculated for all possible cutoff
points and then plotted as a receiver
operating characteristic (ROC)
curve.  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 defining a positive
test.  Sensitivity, specificity,
and positive likelihood ratios (LRs)
were calculated for potential predictor
variables. Potential predictor
variables were entered into a stepwise
logistic regression model to
determine the most accurate set of
variables for prediction of treatment
success. A significance level of .10
was required for removal from the
equation to minimize the likelihood
of excluding potentially helpful variables. 
Variables retained in the
regression model were obtained as
the CPR for classifying patients with
neck pain who are likely to benefit
from thoracic spine thrust manipulation,
exercise, and patient education
for this sample of subjects.
We further analyzed the data to determine
whether weighting individual
predictors according to the relative
size of the beta coefficients increases
the prognostic accuracy of the model.
Weights were calculated by taking the
beta coefficient for each variable in
the final model and dividing it by the
lowest beta coefficient and then
rounding to the nearest integer. 
Once the weight was formulated, an
ROC curve was used to identify the
cutoff value that represented the best
diagnostic accuracy for the pointbased
system.  Sensitivity, specificity,
and positive LRs as well as corresponding
95% confidence intervals
were calculated for the cutoff point
that maximized the diagnostic utility
of the weighting system.
The Cohen kappa ()50 was used to
calculate the interrater reliability of
categorical data with only 2 possible
response options from the patient
history and clinical examination. A
weighted kappa  was used to calculate
the reliability of categorical data
with 3 response options such as
intersegmental mobility assessment
techniques as well as the symptom
response (increased pain, decreased
pain, no change). Intraclass correlation
coefficients (ICC[2,1]) and the
95% CIs were calculated to determine
the interrater reliability for
continuous variables. 
Therapists were characterized by
years of experience to determine the
effect of experience on patient outcomes.
Therapists were dichotomized
as having 3 or fewer years of
experience or more than 3 years of
experience. Only one treating clinician
had less than 3 years of experience.
The percentage of successful
outcomes for each group (≤3 years
of experience or >3 years of experience)
was calculated and compared
using a chi-square test of independence.
The NDI change scores
also were calculated and were compared
between groups using independent
Between March 2004 and September
2005, 80 subjects were recruited for
the study. The total number of subjects
screened, reasons for ineligibility,
and dropouts are shown in Figure 5. Two subjects failed to return
for the second treatment session,
and their data were excluded from
the analysis. Subject demographics
and initial baseline variables from
the patient history and self-report
measures for the entire sample as
well as for the successful outcome
and nonsuccessful outcome groups
are presented in Table 1. Baseline
clinical examination variables for the
entire sample and for the successful
outcome and nonsuccessful outcome
groups are shown in Table 2
for categorical data and in Table 3
for continuous data. Forty-two subjects
were categorized as having
achieved a successful outcome, and
36 subjects were categorized as having
achieved a nonsuccessful outcome.
Twenty-three subjects (55%)
were classified as having a successful
outcome after the initial treatment,
and 19 subjects (45%) were classified as having a successful outcome
after 2 sessions.
The mean number
of days between visit 1 and visit 2
was 2.3 (SD=0.7) and 2.3 (SD=0.6)
(P=.53) for the successful outcome
and nonsuccessful outcome groups,
respectively. The mean number of
days between visit 1 and visit 3 was
6.3 (SD=1.2) and 6.2 (SD=1.2)
(P=.99) for the successful outcome
and nonsuccessful outcome groups
respectively. Analysis of NPRS and
NDI change scores revealed that the
successful outcome group experienced
significantly greater improvements
(P<.001) in pain (NPRS
change score=2.2, 95% CI=1.4 –
2.9) and disability (NDI change
score=18.6%, 95% CI=13.3–25.0)
over the nonsuccessful outcome
The 10 potential predictor variables
(Tab. 4) that exhibited a significance
level of less than .10 were entered
into the logistic regression. The cutoff
values determined by the ROC
curves were 11.5 for the FABQPA
subscale, 9.5 for the FABQW subscale,
30 days since the onset of
symptoms, and 30 degrees of cervical
extension. In addition, the number
of prior episodes of neck pain
was dichotomized into <3 episodes
or ≥3 episodes. Accuracy statistics
for all 10 variables (and 95% CIs) are
shown in Table 4. The positive LRs
ranged from 1.1 to 6.4, with the
strongest predictor being symptom
duration of <30 days.
The following 6 variables were
retained in the final regression model:
symptom duration of <30 days,
no symptoms distal to the shoulder,
subject reports that looking up does
not aggravate symptoms, FABQPA
score of <12, diminished upper thoracic
spine kyphosis (T3–T5), and
cervical extension of <30 degrees
(χ2 =55.0, df=6, P<.001, Nagelkerke
R2=.682). These 6 variables
were used to form the most parsimonious
combination of predictors for
identifying patients with neck pain
who are likely to benefit from thoracic
spine thrust manipulation. Reliability
data for these variables are
shown in Table 4. The reliability values
for the remainder of the patient
history and clinical examination are
reported elsewhere. 
Fourteen out of 15 subjects who
were positive on at least 4 of the
criteria and 32 of 37 subjects who
were positive on at least 3 criteria
were in the successful outcome
group. Of the 41 subjects with 2 or
fewer variables, 31 were in the
nonsuccessful outcome group
(Table 5). Accuracy statistics were
calculated for the numbers of variables
present (Table 6). The pretest
probability for the likelihood of
success with thoracic spine thrust
manipulation for this study was
54% (42 out of 78 subjects). If a
subject exhibited 4 out of the 6
variables, the positive LR was 12.0
(95% CI=2.3–70.8) and the posttest
probability of success
increased to 93%. If a subject was
positive on 3 out of the 6 variables,
the positive LR was 5.5 (95%
CI=2.7–12.0) and the posttest
probability of success was 86%. If
only 2 of the 6 variables were
present, the positive LR decreased
to 2.1 (95% CI=1.5–2.5) and the
posttest probability of success was
The analysis of the point-based system
revealed a possible total of 10
points (for the 6 variables). The cutoff
that maximized the diagnostic
accuracy of the point-based system
was 3.5 points. This resulted in a
sensitivity of .83 (95% CI=.69 –.92),
a specificity of .86 (95% CI=.71–
.94), a positive LR of 5.9 (95%
CI=2.6 –13.0), and a posttest probability
There was no significant difference
in outcomes among therapists with
varying levels of experience for
either the percentage of successful
outcomes or NDI change scores
(P>.05). The group with 3 years of
experience achieved a success rate
of 16/30 (53%), and the group that
had ≤3 years of experience demonstrated
a success rate of 26/48 (54%).
The NDI change scores were 12.8
(SD=15.7) for the group with ≤3
years of experience and 14.8
(SD=14.6) for the group with >3
years of experience.
The LR is the statistic often used to determine the usefulness of a CPR.  We selected to report the positive LR because the purpose of this study was to determine the change in probability that patients are likely to experience a successful outcome when they satisfy the criteria of the CPR. Based on the pretest probability in this study (54%) that a subject would respond positively to thoracic spine thrust manipulation, if the subjects exhibited 4 of the 6 criteria (positive LR=12), the posttest probability of success increased dramatically to 93%. However, based on the wide CI associated with positive findings on 4 out of 6 tests (95% CI=2.28–70.8), clinicians can have greater accuracy when determining the likelihood that a patient with neck pain will exhibit a rapid response to thoracic spine thrust manipulation when using 3 out of 6 variables (positive LR=5.5, 95% CI=2.72–12.0) to guide decision making (posttest probability=86%).
In some circumstances, assigning a weight to individual predictors based on the beta coefficients increases the accuracy of prognostic models.  However, in some instances, it is possible that translating a prognostic model to a point-based scoring system can decrease the discriminatory power of the index.  The cutoff point for the point-based system that maximized the diagnostic accuracy resulted in a positive LR of 5.9 and a posttest probability of 87%, which only exceeded the posttest probability of the equal scoring system of the CPR by 1%. We therefore refrained from using the point-based system because it does not add to the predictive accuracy of the rule and would increase the complexity of the CPR, likely further detracting from the implementation of the rule in clinical practice. 
The ability to a priori identify patients with neck pain who are likely to experience an early success with thoracic spine thrust manipulation while avoiding the potential risk associated with cervical spine thrust manipulation is useful for guiding clinical decision making for individual patients. The CPR also is useful for identifying patients with neck pain who should perhaps receive other forms of treatment rather than thoracic spine thrust manipulation. In our study, for example, if subjects exhibited only one of the variables, the positive LR was only 1.2, suggesting that the posttest probability of these subjects achieving a successful outcome is not much larger than chance, corresponding to a negligible increase of the posttest probability to 58% (Table 6).
Six predictor variables were retained in the logistic regression analysis as maximizing the accuracy of predicting patients with neck pain who are likely to respond to thoracic spine thrust manipulation. Although the duration of the current episode was the strongest individual predictor, we used a higher threshold for defining success on the GROC than what has been recommended  to provide a greater degree of distinction between subjects who improved dramatically with manipulation and those who were improving over time simply due to natural history of the disorder. In addition, the magnitude of the difference in change scores for both the NPRS and NDI further substantiates that an important clinical change occurred in the group that was identified as having experienced a successful outcome.
The duration of the current episode was identified as the strongest predictor in a CPR for identifying patients with LBP who are likely to experience a rapid and dramatic response to spinal manipulation (positive LR=4.39).  The validation of the CPR also demonstrated that a shorter duration of symptoms was predictive for identifying patients who would respond to manipulation (positive LR=4.4).  However, duration of symptoms was not predictive of the outcomes associated with the comparison group who received an exercise program (positive LR=1.0), suggesting that a shorter duration is predictive of response to manipulation and not the natural history of the disorder.  Further validation studies are needed to determine whether this is also the case with the current CPR.
The FABQ was a predictor variable for identifying patients with LBP who are likely to respond to either spinal manipulation (FABQW) [19, 20] or spinal stabilization (FABQPA).  In contrast to patients with LBP who are likely to benefit from spinal stabilization who exhibited elevated FABQPA scores (>8),  our study identified lower FAPQPA scores (<12) as a predictor of a successful outcome. A correlation between disability and the FABQPA was identified by George et al  and Nederhand et al,  suggesting that fear-avoidance beliefs exhibit predictive validity in identifying patients with neck pain who may be at risk for chronic disability. Further research is necessary to clarify the role of fear-avoidance beliefs in patients with neck pain.
One common flaw in the development of CPRs is that researchers often do not investigate the reliability of the measures used in their study and thus cannot determine whether predictor variables provide adequate reproducibility to be included in the rule.  We investigated the reliability of potential predictor variables and, according to the descriptive criteria provided by Landis and Koch,  all variables in the CPR exhibited fair to substantial reliability. We consider these reliability coefficients acceptable to guide clinical decision making in the management of patients with neck pain.
The predictor variables of a decreased upper thoracic spine kyphosis from T3 through T5 and decreased cervical extension may be associated with the biomechanical link between the thoracic spine and the cervical spine. Recent literature identified a correlation between mobility at the cervicothoracic junction and thoracic spine with neck-shoulder pain. [11, 60, 61] It is also possible that impaired mobility in the thoracic spine may be a contributor to mechanical neck pain. [62–64] Patient reports of “looking up does not aggravate the symptoms” and “no symptoms distal to the shoulder,” as recorded on a body diagram, also were identified as predictor variables in the CPR. In contrast, the population that has pain distal to the shoulder that is aggravated by looking up could potentially be a subgroup of patients with cervical radiculopathy rather than solely mechanical neck pain. [65, 66] Although symptoms extending into the arm and radicular signs are not associated with a worse prognosis,  it has been suggested that patients with more distal symptoms may be more responsive to a different treatment approach such as cervical traction and other distraction-oriented interventions. 
We successfully achieved the purpose of developing a CPR that identifies patients with neck pain who are likely to exhibit early success after thoracic spine thrust manipulation. However, this is only the first step in the process of developing and testing a CPR.  Although no difference in outcomes occurred among clinicians with varying levels of experience, it should be recognized that data were collected at only one clinical site by 4 physical therapists. Future studies are necessary to validate our results and determine whether similar findings occur in a broader patient population with different treating clinicians. Additionally, a validation study should include a long-term follow-up and a comparison group to further investigate the predictive value of the variables in the CPR. If the rule is validated, an impact analysis of implementation of the rule on clinical practice patterns, outcomes, and costs of care should be investigated.
Cote P, Cassidy J, Carroll L. The factors
associated with neck pain and its related
disability in the Saskatchewan population.
Palmer KT, Walker-Bone K, Griffin MJ,
et al. Prevalence and occupational associations
of neck pain in the British population.
Scand J Work Environ Health. 2001;
Cote P, Cassidy JD, Carroll LJ, Kristman V.
The annual incidence and course of neck
pain in the general population: a population-
based cohort study. Pain. 2004;112:
Wright A, Mayer T, Gatchel R. Outcomes
of disabling cervical spine disorders in
compensation injuries: a prospective comparison
to tertiary rehabilitation response
for chronic lumbar disorders. Spine. 1999;
Jette AM, Smith K, Haley SM, Davis KD.
Physical therapy episodes of care for
patients with low back pain. Phys Ther.
Borghouts J, Janssen H, Koes B, et al. The
management of chronic neck pain in general
practice: a retrospective study. Scand
J Prim Health Care. 1999;17:215–220.
Haldeman S, Kohlbeck FJ, McGregor M.
Unpredictability of cerebrovascular ischemia
associated with cervical spine manipulation
therapy: a review of sixty-four
cases after cervical spine manipulation.
Spine. 2002;27:49 –55.
Di Fabio RP. Manipulation of the cervical
spine: risks and benefits. Phys Ther. 1999;
Haldeman S, Kohlbeck FJ, McGregor M.
Stroke, cerebral artery dissection, and cervical
spine manipulation therapy.
J Neurol. 2002;249:1098–1104.
Cleland JA, Childs JD, McRae M, et al.
Immediate effects of thoracic manipulation
in patients with neck pain: a randomized
clinical trial. Man Ther. 2005;10:
Norlander S, Nordgren B. Clinical symptoms
related to musculoskeletal neckshoulder
pain and mobility in the cervicothoracic
spine. Scand J Rehabil Med.
Adams G, Sim J. A survey of UK manual
therapists’ practice of and attitudes
towards manipulation and its complications.
Physiother Res Int. 1998;3:
Fernandez-de-las-Penas C, Fernandez-Carnero
J, Fernandez AP, et al. Dorsal manipulation
in whiplash injury treatment: a randomized
controlled trial. Journal of
Whiplash and Related Disorders. 2004;3:
Savolainen A, Ahlberg J, Nummila H, Nissinen
M. Active or passive treatment for
neck-shoulder pain in occupational health
care? A randomized controlled trial. Occup
Med (Lond). 2004;54:422–424.
Leboeuf-Yde C, Lauritsen JM, Lauritzen T.
Why has the search for causes of low back
pain largely been nonconclusive? Spine.
Petren-Mallmin M, Linder J. MRI cervical
spine findings in asymptomatic fighter
pilots. Aviat Space Environ Med. 1999;70:
Rose S. Physical therapy diagnosis: role
and function. Phys Ther. 1989;69:
Hicks GE, Fritz JM, Delitto A, McGill SM.
Preliminary development of a clinical prediction
rule for determining which
patients with low back pain will respond
to a stabilization exercise program. Arch
Phys Med Rehabil. 2005;86:1753–1762.
Flynn T, Fritz J, Whitman J, et al. A clinical
prediction rule for classifying patients
with low back pain who demonstrate
short term improvement with spinal
manipulation. Spine. 2002;27:2835–2843.
Childs JD, Fritz JM, Flynn TW, et al. A
clinical prediction rule to identify patients
likely to benefit from spinal manipulation:
a validation study. Ann Intern Med. 2004;
Childs JD, Fritz JM, Piva SR, Whitman JM.
Proposal of a classification system for
patients with neck pain. J Orthop Sports
Phys Ther. 2004;34:686–696.
Werneke M, Hart DL, Cook D. A descriptive
study of the centralization phenomenon:
a prospective analysis. Spine. 1999;
Jensen MP, Turner JA, Romano JM. What
is the maximum number of levels needed
in pain intensity measurement? Pain.
Vernon H, Mior S. The Neck Disability
Index: a study of reliability and validity.
J Manipulative Physiol Ther. 1991;14:
Waddell G, Newton M, Henderson I, et al.
Fear-Avoidance Beliefs Questionnaire and
the role of fear-avoidance beliefs in
chronic low back pain and disability.
Jacob T, Baras M, Zeev A, Epstein L. Low
back pain: reliability of a set of pain measurement
tools. Arch Phys Med Rehabil.
George S, Fritz J, Erhard E. A comparison
of fear-avoidance beliefs in patients with
lumbar spine pain and cervical spine pain.
Flynn TW, Whitman J, Magel J. Orthopaedic
Manual Physical Therapy Management
of the Cervical-Thoracic Spine and
Ribcage. San Antonio, Tex: Manipulations
Kendall FP, McCreary EK, Provance PG.
Muscles: Testing and Function. 4th ed.
Baltimore, Md: Williams & Wilkins; 1993.
Griegel-Morris P, Larson K, Mueller-Klaus
K, Oatis CA. 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:
McKenzie RA. Cervical and Thoracic
Spine: Mechanical Diagnosis and Therapy.
Minneapolis, Minn: Orthopaedic
Physical Therapy Products; 1990.
Harris KD, Heer DM, Roy TC, et al. Reliability
of a measurement of neck flexor
muscle endurance. Phys Ther. 2005;85:
Maitland G, Hengeveld E, Banks K, English
K. Maitland’s Vertebral Manipulation.
6th ed. Oxford, United Kingdom: Butterworth-
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.
Magee D. Orthopedic Physical Assessment.
4th ed. Philadelphia, Pa: Saunders;
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.
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.
New York, NY: Churchill Livingstone
Jaeschke R, Guyatt GH, Sackett DL. Users’
guides to the medical literature, III: 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 Evidence-
Based Medicine Working Group. JAMA.
Erhard RE. The Spinal Exercise Handbook:
A Home Exercise Manual for a
Managed Care Environment. Pittsburgh,
Pa: Laurel Concepts; 1998.
Jaeschke R, Singer J, Guyatt G. Measurement
of health status: ascertaining the
minimal clinically important difference.
Controlled Clin Trials. 1989;10:407–415.
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
trial. Spine. 1992;17:28 –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:
Farrar J, Young JJ, La Moreaux L, et al.
Clinical importance of changes in chronic
pain intensity measured on an 11-pont
numerical pain rating scale. Pain. 2001;
Hurst H, Bolton J. Assessing the clinical
significance of change scores recorded on
subjective outcome measures.
J Manipulative Physiol Ther. 2004;27:
Bolton JE. Sensitivity and specificity of
outcome measures in patients with neck
pain: detecting clinically significant
improvement. Spine. 2004;29:
Hoving JL, Koes BW, de Vet HC, et al.
Manual therapy, physical therapy, or continued
care by a general practitioner for
patients with neck pain: a randomized,
controlled trial. Ann Intern Med. 2002;
Freedman DA. A note on screening regression
equations. The American Statistician.
Deyo RA, Centor RM. Assessing the
responsiveness of functional scales to clinical
change: an analogy to diagnostic test
performance. J Chronic Dis. 1986;39:
Concato J, Feinstein AR, Holford TR. The
risk of determining risk with multivariable
models. Ann Intern Med. 1993;118:
Cohen J. A coefficient of agreement for
nominal scales. Educ Psychol Meas. 1960;
Cohen J. Weighted kappa: nominal scale
agreement with provision for scaled disagreement
or partial credit. Psychol Bull.
Shrout PE, Fleiss JL. Intraclass correlations:
uses in assessing rater reliability.
Psychol Bull. 1979;86:420–426.
Cleland JA, Childs JD, Fritz JM, Whitman
JM. Inter-rater reliability of the historical
and physical examination in patients with
mechanical neck pain. Arch Phys Med
Kuijpers T, van der Windt DA, Boeke AJ,
et al. Clinical prediction rules for the prognosis
of shoulder pain in general practice.
Lee SJ, Lindquist K, Segal MR, Covinsky
KE. Development and validation of a prognostic
index for 4-year mortality in older
adults. JAMA. 2006;295:801–808.
Redelmeier DA, Lustig AJ. Prognostic indices
in clinical practice. JAMA. 2001;285:
Nederhand MJ, Ijzerman MJ, Hermens HJ,
et al. Predictive value of fear avoidance in
developing chronic neck pain disability:
consequences for clinical decision making.
Arch Phys Med Rehabil. 2004;85:
Laupacis A, Sekar N, Stiell IG. Clinical
prediction rules: a review and suggested
modifications and methodological standards.
Landis JR, Koch CG. The measurement of
observer agreement for categorical data.
Norlander S, Aste-Norlander U, Nordgren
B, Sahlstedt B. Mobility in the cervicothoracic
motion segment: an indicative
factor of musculo-skeletal neck-shoulder
pain. Scand J Rehabil Med. 1996;28:
Norlander S, Gustavsson BA, Lindell J,
Nordgren B. Reduced mobility in the cervico-
thoracic motion segment—a risk factor
for musculoskeletal neck-shoulder
pain: a two-year prospective follow-up
study. Scand J Rehabil Med. 1997;29:
Greenman P. Principles of Manual Medicine.
2nd ed. Philadelphia, Pa: Lippincott
Williams & Wilkins; 1996.
Johansson H, Sojka P. Pathophysiological
mechanisms involved in genesis and
spread of muscular tension in occupational
muscle pain and in chronic musculoskeletal
pain syndromes: a hypothesis.
Med Hypotheses. 1991;35:196–203.
Knutson GA. Significant changes in systolic
blood pressure post vectored upper
cervical adjustment vs resting control
groups: a possible effect of the cervicosympathetic
and/or pressor reflex.
J Manipulative Physiol Ther. 2001;24:
Wainner R, Gill H. Diagnosis and nonoperative
management of cervical radiculopathy.
J Orthop Sports Phys Ther. 2000;12:
Daffner S, Hilibrand A, Hanscom B, et al.
Impact of neck and arm pain on overall
health status. Spine. 2003;2817:2035.
Borghouts JA, Koes BW, Bouter LM. The
clinical course and prognostic factors of
non-specific neck pain: a systematic
review. Pain. 1998;77:1–13.
McGinn T, Guyatt G, Wyer P, et al. Users’
guides to the medical literature, XXII: how
to use articles about clinical decision
rules. JAMA. 2000;284:79–84.
Return to the CLINICAL PREDICTION RULE Page
Return to the CHIROPRACTIC AND CHRONIC NECK PAIN Page