69

ORIGINAL ARTICLE

Psychometric Properties of the Neck Disability Index and

Numeric Pain Rating Scale in Patients With Mechanical
Neck Pain
Joshua A. Cleland, PT, PhD, OCS, John D. Childs, PT, PhD, MBA, OCS, Julie M. Whitman, PT, DSc, OCS
ABSTRACT. Cleland JA, Childs JD, Whitman JM. Psychometric properties of the Neck Disability Index and numeric
pain rating scale in patients with mechanical neck pain. Arch
Phys Med Rehabil 2008;89:69-74.
Objective: To examine the psychometric properties including test-retest reliability, construct validity, and minimum levels of detectable and clinically important change for the Neck
Disability Index (NDI) and the numeric rating scale (NRS) for
pain in a cohort of patients with neck pain.
Design: Single-group repeated-measures design.
Setting: Outpatient physical therapy (PT) clinics.
Participants: Patients (N137) presenting to PT with a
primary report of neck pain.
Interventions: Not applicable.
Main Outcome Measures: All patients completed the NDI
and the NRS at the baseline examination and at a follow-up. At
the time of the follow-up, all patients also completed the global
rating of change, which was used to dichotomize patients as
improved or stable. Baseline and follow-up scores were used to
determine the test-retest reliability, construct validity, and minimal levels of detectable and clinically important change for
both the NDI and NRS.
Results: Test-retest reliability was calculated using an intraclass correlation coefficient (ICC) (NDI ICC.50; 95% confidence interval [CI], .25.67; NRS ICC.76; 95% CI, .51.87).
The area under the curve was .83 (95% CI, .75.90) for the
NDI score and .85 (95% CI, .78 .93) for the NRS score for
determining between stable and improved patients. Thresholds
for the minimum clinically important difference (MCID) for
the NDI were 19-percentage points and 1.3 for the NRS.
Conclusions: Both the NDI and NRS exhibit fair to moderate
test-retest reliability in patients with mechanical neck pain. Both
instruments also showed adequate responsiveness in this patient
population. However, the MCID required to be certain that the
change in scores has surpassed a level that could be contributed to
measurement error for the NDI was twice that which has previously been reported. Therefore the ongoing analyses of the properties of the NDI in a patient population with neck pain are warranted.

From the Department of Physical Therapy, Franklin Pierce College, Concord, NH

(Cleland); Rehabilitation Services, Concord Hospital, Concord, NH (Cleland); Manual
Therapy Fellowship Program, Regis University, Denver, CO (Cleland, Whitman); U.S.
Army-Baylor University Doctoral Program in Physical Therapy, San Antonio, TX
(Childs); and Department of Physical Therapy, Regis University, Denver, CO (Whitman).
Supported by the Orthopaedic Section of the American Physical Therapy Association, the American Academy of Orthopaedic Manual Physical Therapists, and Steens
Physical USA.
No commercial party having a direct financial interest in the results of the research
supporting this article has or will confer a benefit upon the author(s) or upon any
organization with which the author(s) is/are associated.
Reprint requests to Joshua A. Cleland, PT, PhD, OCS, Physical Therapy Program,
Franklin Pierce College, 5 Chenell Dr, Concord, NH 03301, e-mail:
joshcleland@comcast.net.
0003-9993/08/8901-00470$34.00/0
doi:10.1016/j.apmr.2007.08.126

Key Words: Neck; Neck pain; Rehabilitation.

2008 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and
Rehabilitation
EALTH OUTCOME MEASURES are commonly used in
H
both the clinical and research environment to determine if
treatment has impacted the patients health status. Prior to
1

using self-report measures to guide clinical decision-making

regarding individual patients, the psychometric properties
(minimum detectable change [MDC], minimal clinically important difference [MCID]) of the particular instrument must
be identified to allow the clinician to categorize if a particular
patient has experienced a clinically important change with a
degree of confidence. MDC is the amount of change that must
be observed before the change can be considered to exceed the
measurement error,2 whereas MCID is the smallest difference
that patients perceive as beneficial.3
The Neck Disability Index (NDI) is a commonly used health
outcome measure to capture perceived disability in patients
with neck pain.1 Riddle and Stratford4 identified a significant
correlation between the NDI and both the physical and mental
health components of the Medical Outcomes Study 36-Item
Short-Form Health Survey (SF-36). The authors also confirmed
that the NDI possesses adequate sensitivity to detect the magnitude of change that occurred for patients reaching their functional goals, work status, and if the patient was currently in
litigation.4 Jette and Jette5 further substantiated the sensitivity
to change by calculating the effect sizes for change scores of
both the NDI and SF-36. However, these data do not provide
useful information to assist clinicians in determining the minimal amount of change necessary to represent a clinically
important difference from the patients perspective.6 It has
been suggested that indices of responsiveness that indicate a
cutoff point for identifying the MCID are of greatest value to
clinicians when they are determining if a meaningful change
has occurred.7
Two studies8,9 with small sample sizes have identified the
MDC, which is the amount of change that must be observed
before the change can be considered to exceed the measurement error for the NDI. Westaway et al8 identified the MDC as
5 points in a group of 31 patients with neck pain. Additionally,
Stratford et al9 identified the MDC to be 5 points in a group of
48 patients with neck pain. Although these studies reported that
a change of 5 points (10%) must be observed to be certain that
the change in scores is greater than measurement error, no
values for the MCID have been reported in the literature for
patients with neck pain.10
No consensus exists on the ideal reference standard for
measuring functional change.4 Westaway8 and Stratford9 used
a clinicians prognostic rating as the reference standard to
determine the responsiveness of the NDI. In these studies,
experienced clinicians were asked a priori to identify those
patients likely to experience little change in status or full
Arch Phys Med Rehabil Vol 89, January 2008

70

PSYCHOMETRIC PROPERTIES OF THE NDI AND NRS, Cleland

recovery. This assessment was based on the patients initial

presentation, degree of identified impairments, and duration of
symptoms.9 However, recent evidence suggests considerable
variation exists among expert clinicians reports of identifying
the probability of disease,11 which raises some concerns regarding the validity of a prognostic rating scale as a reference
standard.
In addition to the NDI, the numeric rating scale (NRS) for
pain is also a commonly used outcome measure for patients
with neck pain.12-14 The responsiveness of the NRS in a broad
population of patients with various musculoskeletal conditions
has been investigated and the MCID has been identified to be
2 points.15 Additionally, in a patient population with low back
pain, the scales also showed an MCID of 2 points.16 However,
the responsiveness of the NRS in a patient population with
neck pain has yet to be determined. Based on the limited scope
of previous work, the purpose of this study was to identify the
psychometric properties of the NDI and the NRS in a large
cohort of patients with neck pain.
METHODS
We collected data on consecutive patients presenting to 1 of
5 outpatient orthopedic physical therapy (PT) clinics (Rehabilitation Services of Concord Hospital, Concord, NH; NewtonWellesley Hospital, Boston, MA; Centennial Physical Therapy,
Colorado Springs, CO; Groves Physical Therapy, St Paul, MN;
Sharp HealthCare, San Diego, CA) between July 2004 and July
2006 with a primary report of neck pain from 2 clinical trials
that were included in the analysis.17,18 Both studies included
identical eligibility criteria. Inclusion criteria included patient
age between 18 and 60 years, an NDI score greater than 10%,
and a primary complaint of neck pain with or without referral
of symptoms to the upper extremity or extremities. Exclusion
criteria included any signs or symptoms consistent with a
nonmusculoskeletal etiology for the patients symptoms, a history of a whiplash injury within the past 6 weeks, evidence of
central nervous system involvement, 2 or more signs consistent
with nerve root compression (myotomal weakness, sensory
deficits in a dermatomal pattern, decreased or absent muscle
stretch reflexes), prior surgery to the cervical or thoracic spine,
or pending legal action regarding their neck pain. All subjects
who agreed to participate in either study signed an informed
consent approved by the institutional review board at the respective clinical site.
Twelve physical therapists participated in the examination
and treatment of all patients in this study. All therapists underwent a standardized training regimen, which included studying
a manual of standard procedures with the operational definitions of each examination and treatment procedure used in this
study. All participating therapists underwent training provided
by a current fellow in the Regis University Manual Therapy
Fellowship Program, Denver, CO. During this training session,
all participating therapists were required to demonstrate the
examination and treatment techniques to ensure that all study
procedures were performed in a standardized fashion. Participating therapists had a mean standard deviation (SD) of
9.76.8 years (range, 119y) of clinical experience.
All patients provided demographic information and completed a number of self-report measures, followed by a standardized history and physical examination at baseline. Selfreport measures included a body diagram,19 NRS,20 the NDI,21
and the Fear-Avoidance Beliefs Questionnaire (FABQ).22
Neck Disability Index
The NDI contains 10 items7 related to activities of daily
living, 2 related to pain, and 1 related to concentration.23 Each
Arch Phys Med Rehabil Vol 89, January 2008

item is scored from 0 to 5, and the total score is expressed as

a percentage (total possible score, 100%), with higher scores
corresponding to greater disability.
NRS for Pain
We used the NRS to capture the patients level of pain.
Patients were asked to indicate the intensity of current, best,
and worst levels of pain over the past 24 hours using an
11-point scale, ranging from 0 (no pain) to 10 (worst pain
imaginable).24 The average of the 3 ratings was used to represent the patients level of pain over the previous 24 hours. This
procedure has been shown to have adequate reliability, validity,
and responsiveness in patients with low back pain,16,25 but has
not been specifically examined in patients with neck pain.
Standardized History, Physical Examination,
and Interventions
The standardized history 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 included a neurologic screen,26 postural
assessment,27 cervical range of motion measurements and
symptom response,28 assessment of the length26 and strength27
of the muscles of the upper quarter, and endurance of the deep
neck flexor muscles.29 The amount of mobility and symptom
response was recorded for spring testing30 of the cervical 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 A,31 cervical distraction test,23 and the upper-limb neurodynamic test.32 All patients who completed the baseline evaluation underwent 1 PT treatment session consisting of manual
therapy techniques directed at the thoracic spine. They returned
within 2 to 4 days for a re-evaluation and again completed the
self-report measures as well as the global rating of change scale
(GRCS). After 1 session of PT, we expected to see a dramatic
change in patient status for a subgroup of patients with mechanical neck pain, but no change in another subgroup that did
not respond positively to the intervention provided.17
Global Rating of Change
At the follow-up evaluation, each patient completed a GRCS
as described by Jaeschke et al.3 Patients were asked to rate their
overall perception of improvement since beginning treatment
on a scale ranging from 7 (a very great deal worse) to zero
(about the same) to 7 (a very great deal better). It has been
recommended3 that scores on the GRCS between 3 and 1
represent small changes, scores between 4 and 5 represent
moderate changes, and scores of 6 or 7 large changes. A
GRCS has also been used to identify responsiveness and the
MCID for health outcome measures.33-35 Use of the GRCS to
calculate responsiveness has been criticized because the patients must recall their initial status weeks or months after the
initial examination.36 Schmitt and Di Fabio37 recently reported
that a retrospective GRCS does not accurately reflect change
over time. However, the follow-up time frame in this study by
Schmitt and Di Fabio37 was 3 months and shorter follow-up
periods (1wk) could potentially reduce the biases associated
with patient recall. Hence we elected to again collect the
measures at the time of the patients first follow-up which was
scheduled within 2 to 4 days of the initial examination. Intuitively, it also makes sense that a patients perception of improvement gives a more accurate assessment if a true change
has occurred rather than a prognostic rating.25

71

PSYCHOMETRIC PROPERTIES OF THE NDI AND NRS, Cleland

Table 1: Baseline Statistics for Stable and Improved Patients

Characteristics

Age (y)
Symptom duration (d)
Sex (% female)
Symptoms distal to shoulder
(% yes)
Initial NDI score
Initial NRS score
FABQ physical activity subscale
FABQ work subscale
Currently taking medications
(% yes)
Currently on workers
compensation (% yes)
Currently seeking litigation
(% yes)
Days between evaluations

Stable
Patients
(n89)

Improved
Patients
(n48)

42.112.0
82.474.8
60.069.0

43.212.0
46.440.3
25.052.0

.620
.002
.160*

27.031.0
32.211.6
4.61.9
12.34.4
13.511.0

15.031.0
35.79.8
5.11.6
10.94.4
11.58.8

.330*
.080
.110
.030
.280

51.059.0

25.052.0

.590*

8.09.0

4.08.0

.760*

6.07.0
2.4.87

4.08.0
2.51.1

.730*
.540

NOTE. Values are means SD unless otherwise indicated.

*Chi-square tests; independent-samples t tests.

Data Analysis
We dichotomized patients into 2 groups based on GRCS
scores; those scoring between 3 and 3 were considered
stable (minimal to no change), those scoring greater than 3
were considered to have exhibited clinically important improvement, and patients scoring less than 3 were considered
to have experienced a worsening in status. Baseline variables
were compared between groups using independent t tests for
continuous data, and chi-square tests of independence for categoric data. Test-retest reliability of the NDI and NRS were
investigated in the group of stable patients using an intraclass correlation coefficient (ICC2,1) with the 95% confidence interval (CI), calculated according to procedures described by Shrout and Fleiss.38 Assessment of reliability was
performed using criteria described by Shrout39 with values
less than .10 indicating no agreement, values between .11
and .40 indicating slight agreement, values between .41 and
.60 indicating fair agreement, values between .61 and .80
indicating moderate agreement, and values greater than .81
indicating substantial agreement. Additionally, a Pearson product-moment correlation coefficient (r) was calculated for preand post-test measurements on the NDI to allow direct comparison with the results of other studies.21
Construct validity of the NDI and NRS were examined by
comparing the change in scores for the stable and improved
groups using separate 2-way analysis of variance for the repeated measures at baseline and re-evaluation. We hypothesized that stable patients would have NDI and NRS scores that
did not change, whereas improved patients would show a
significant change in disability. This would be represented by a
significant group by time interaction.
We analyzed responsiveness, or the ability of a test to
recognize change,40 of the NDI and NRS with 2 methods. The
first method used receiver operator characteristic (ROC) curves
constructed by plotting sensitivity values (true positive rate) on
the y axis and 1 specificity values (false positive rate) on the
x axis for each level of change score for distinguishing improved from stable patients. Separate ROC curves were constructed for the NDI and NRS. The area under the curve (AUC)
and the 95% CI were obtained as a method for describing the

ability of each measure to distinguish improved patients from

stable patients. An AUC of .50 indicates the measure has no
diagnostic accuracy beyond chance, whereas a value of 1
would indicate perfect accuracy.41 Responsiveness was also
analyzed by correlating the change scores of the NDI and NRS
to the GRCS scores in all patients. Change scores were calculated by subtracting each patients baseline score from the
score obtained at the re-evaluation.
Minimum detectable change, or the amount of change that
must be observed before the change can be considered to
exceed the measurement error, was calculated by determining
the standard error (SE) of measurement for the NDI and NRS
for the stable group.2 The SE of measurement was calculated
using the formula (SD [1 r]1/2) where r is the test-retest
reliability coefficient and the SD is the square root of the total
variance. The SE of measurement was multiplied by 1.65 to
determine the 90% CI.34 This value was multiplied by the
square root of 2 to account for the errors taken with repeated
measurements.42 Minimal clinically important change, the
smallest difference that patients perceive as beneficial,3 was
calculated by identifying the point on the ROC curve nearest
the upper left-hand corner, which is considered to be the best
cutoff score for distinguishing improved and stable patients.34
Sensitivity and specificity values for the selected cutoff score
were calculated.
RESULTS
Of the 209 consecutive patients with neck pain screened for
eligibility, 138 patients satisfied inclusion and exclusion criteria and agreed to participate (mean, 42.511.9y). Of the 71
patients that were not included, 21 had a recent history of
whiplash, 13 had signs of nerve root compression, 12 presented
with contraindications to the interventions, 8 had prior surgery
to the cervical or thoracic spine, 2 had signs of central nervous
system involvement, 1 had insufficient English skills to complete the questionnaires, and 14 patients declined to participate.
Only 1 patient exhibited a worsening of status (GRCS score,
4) and was not included in further analyses. The mean GRCS
score for the remaining 137 patients was 2.12.3. Eighty-nine
patients were considered to have remained stable (GRCS score
range, 3 to 3) and 48 were considered to have improved
(GRCS score, 4). Baseline characteristics for both groups can
be found in table 1. The mean follow-up time between the first
and second measurements was 2.500.95 days. A significant
difference existed in the duration of symptoms and initial
FABQ and FABQ physical activity scores between the group
that remained stable and the group that improved. The ICC
values calculated from the stable patients were .50 (95% CI,
.25.67) for the NDI and .76 (95% CI, .51.87) for the NRS
(table 2). The Pearson r value calculated between pre- and
post-test measurements in the group of patients considered
stable for the NDI was .56.
Figure 1 shows the mean initial and follow-up scores for the
NDI for the stable and improved groups. There was a significant interaction (P.001) between groups for the pre- and
post-test scores indicating that the change in NDI with time
differed between stable and improved patients (mean, 12.9;

Table 2: Mean Change Scores and 95% CIs for Stable Patients
(n89) for the NDI and NRS
Scale Type

Baseline

Re-Evaluation

Change Score

ICC2,1 (95% CI)

NDI
NRS

32.211.6
4.61.9

26.212.2
3.91.8

6.910.4
.711.1

.50 (.25.67)
.76 (.51.87)

Arch Phys Med Rehabil Vol 89, January 2008

72

PSYCHOMETRIC PROPERTIES OF THE NDI AND NRS, Cleland

1.2

35

30
25
Stable

20

Improved

15

Sensitivity

40

0.8
0.6
0.4

10

0.2

0
0

0
Initial Exam

0.2

0.4

0.6

0.8

1.2

1 - Specificity

Follow-up

Fig 1. NDI scores for the groups of subjects defined as stable and
improved based on global rating of change. The interaction between time and group was significant (P<.001).

95% CI, 9.316.5). Figure 2 shows the initial and follow-up

scores for the NRS for the stable and improved groups. A
significant interaction (P.001) was also found between the
initial and follow-up evaluations for the NRS between improved and stable patients (mean, 2.1; 95% CI, 1.6 2.6) indicating that the change in NRS scores differed between patients
determined to be stable or improved based on the GRCS. The
AUC for the NDI was .83 (95% CI, .75.90) and .85 (95% CI,
.78 .93) for the NRS. Figures 3 and 4 represent the ROC
curves for the NDI and NRS, respectively. Pearson r value
between change scores of the NDI and NRS with the GRCS for
the entire group was .58 (P.01) and .57 (P.01), respectively
(table 3).
The SE of measurement values calculated from the stable
patients were 8.4 for the NDI and .91 for the NRS, respectively.
These values corresponded with MDC values of 19.6-percentage points for the NDI and 2.1 for the NRS. Thresholds for
MCID for the NDI were 19-percentage points (sensitivity, .83;
specificity, .72) and 1.3 for the NRS (sensitivity, .88; specificity, .71).
DISCUSSION
It is essential for clinicians to have an understanding of the
psychometric properties of measures, including reliability and

Fig 3. ROC curve for the NDI at the follow-up. The point nearest the
uppermost left-hand corner of the graph represents the MCID. The
circled value is the point nearest the left-hand corner and represents
the MCID for the NDI.

responsiveness. Instruments should exhibit acceptable reliability and validity prior to being used to guide clinical decisionmaking. To determine the reliability and validity of self-report
measures, it is useful to compare them with a construct that
indicates when a true change has occurred.34 Frequently this
construct of true change is a patient global rating of
change.43,44 We examined the test-retest reliability of the NDI
and NRS for a subgroup of patients with mechanical neck pain.
The results of this study suggest that the NDI exhibits only fair
test-retest reliability (ICC.50), which is considerably lower
than the values reported by Vos et al45 for the Dutch version of
the NDI. Other studies have identified the NDI to exhibit high
test-retest reliability when using correlation coefficients as the
method of data analysis.21,46 However, our study exhibited a
much lower correlation coefficient than that previously reported for the test-retest reliability of the NDI. The NRS
exhibited moderate test-retest reliability, which is similar to the
test-retest reliability identified in a patient population with
cervical radiculopathy.35
Construct validity for both outcome measures was examined
by comparing the baseline and follow-up scores for both the
stable and improved groups. Both the NDI and NRS exhibited
significantly greater reductions in disability among patients

1.2

6
1

4
Stable

Improved

Sensitivity

5
0.8
0.6
0.4
0.2

0
0

0
Initial Exam

Follow-up

Fig 2. NRS scores for the groups of subjects defined as stable and
improved based on global rating of change. The interaction between time and group was significant (P<.001).

Arch Phys Med Rehabil Vol 89, January 2008

0.2

0.4

0.6

0.8

1.2

1 - Specificity

Fig 4. ROC curve for the NRS score at the follow-up. The point
nearest the uppermost left-hand corner of the graph represents the
MCID. The circled value is the point nearest the left-hand corner and
represents the MCID for the NRS score.

PSYCHOMETRIC PROPERTIES OF THE NDI AND NRS, Cleland

Table 3: Pearson r Values for Change Scores (N137)
Scale

NDI

NRS

GRCS score

.58

.57

NOTE. Correlation at P.001.

rating themselves as improved versus stable.4,5,9,43,44 These

data further substantiate the findings of other studies that have
investigated the validity of the NDI as compared with other
measures.4,5,8,9,21
We used 2 methods to investigate the responsiveness. In the
first, we used ROC curves to calculate the AUC for both the
NDI and NRS. The AUC for the NDI and NRS were .83 and
.85, respectively. The AUC of .83 (95% CI, .75.90) is slightly
lower than that reported by Stratford et al9 who identified an
AUC of .9. However, the AUC of .85 (95% CI, .78 .93) for the
NRS is greater than that identified by Childs et al16 in patients
with low back pain, suggesting that the responsiveness of the
NRS may differ depending on the patient population it is
applied. Second, we investigated responsiveness by calculating
correlation coefficients between the NDI and NRS for the
group of patients who were identified as improved on the
GRCS. Results showed a moderate but significant correlation
for both the NDI and the NRS, further substantiating other
reports supporting the validity of both measures.1,6,15
The NRS exhibited an MDC of 2 points and an MCID of 1.3
points, which is consistent with the findings in heterogeneous
groups of patients with musculoskeletal conditions15 and patients with low back pain.16 However, Bolton6 investigated the
psychometric properties of the 7 questions on the Bournemouth
Questionnaire using the NRS. The results showed that for the
questions on the Bournemouth the MCID was 3 points.6 Although pain is 1 dimension measured by the Bournemouth, it
appears that the NRS possesses different psychometric properties when used to measure other dimensions (disability, affective, and cognitive-behavioral).
Although the NDI showed responsiveness in the ability to
detect change, the MDC required to be certain that the patient
has exhibited a change that exceeds that of measurement error
(19 percentage points) was double that which was previously
reported.9 Perhaps the difference is a direct result of using a
prognostic rating to identify which patients exhibit a true
change and that of the transitional scale, the GRCS.8,9 Vos
et al45 used a different GRCS and identified a much smaller MDC
(1.66 points) for the Dutch version of the NDI. It is also
possible that the dichotomous scoring system we used for the
GRCS (status quo or improved) may fail to detect small but
meaningful improvements in patient status.25 For example, if a
GRCS cutoff between 2 and 2 were considered stable, the SE
of measurement for the NDI would have been 8.26 with an
MDC of 13.6%. Additionally, the MCID would have been 14%
(or 7 points). A GRCS cutoff between 3 and 3 is most often
used in the literature to examine the responsiveness of patient
reported outcome measures34,35; however, the use of a GRCS
between 2 and 2 has also been reported.16 This suggests that
the values reported using a GRCS score between 2 and 2
might be a more accurate value for the MCID. Although
arguments can be made for and against both methods,37,47 the
results indicate that further investigation into the MDC and
MCID in patients with neck pain is warranted.
Study Limitations
We have examined the responsiveness of the NDI and NRS
in a population with mechanical neck pain. Hence, the findings

73

of this study can only be generalized to a similar population

and not to patients who present with neck pain related to other
underlying pathologies. Although we have proposed that the
short term follow-up could potentially reduce the likelihood of
recall bias associated with the GRCS, the short-term follow-up
could have also been the reason why the responsiveness of the
NDI was lower than that reported in other studies.8,9 It has been
reported that instruments are usually less responsive in shortterm follow-ups.16 Future studies should first identify the most
appropriate reference standard for examining the responsiveness of self-report outcome measures.37,48 Additionally, examining the psychometric properties of instruments should be
considered an ongoing process and we suggest that future
studies should continue to examine the responsiveness of the
NDI in patient populations with neck pain.
CONCLUSIONS
The results of our study indicate that both the NDI and NRS
exhibit fair to moderate test-retest reliability. Both instruments
also showed adequate responsiveness in this patient population.
However, the MDC required to be certain that the change in
scores has surpassed a level that could be attibuted to measurement error for the NDI was twice that which has previously
been reported in the literature.9,16 This warrants further investigation. The responsiveness, as well as the MDC and MCID,
for the NRS were consistent with studies performed on different patient populations.
Acknowledgment: None of the funding organizations played any
role in the design, conduct, or reporting of the study or in the decision
to submit the study for publication.
References
1. Pietrobon R, Coeytaux RR, Carey TS, Richardson WJ, DeVellis
RF. Standard scales for measurement of functional outcome for
cervical pain or dysfunction. Spine 2002;27:515-22.
2. Beaton DE, Bombardier C, Katz JN, et al. Looking for important
change/differences in studies of responsiveness. J Rheumatol
2001;28:400-5.
3. Jaeschke R, Singer J, Guyatt GH. Measurement of health status:
ascertaining the minimal clinically important difference. Control
Clinl Trials 1989;10:407-15.
4. Riddle D, Stratford P. Use of generic versus region-specific functional status measures on patients with cervical spine disorders.
Phys Ther 1998;78:951-63.
5. Jette DU, Jette AM. Physical therapy and health outcomes in
patients with spinal impairments. Phys Ther 1996;76:930-41.
6. Bolton JE. Sensitivity and specificity of outcome measures in
patients with neck pain: detecting clinically significant improvement. Spine 2004;29:2410-7.
7. Husted JA, Cook RJ, Farewell VT, Gladman DD. Methods for
assessing responsiveness: a critical review and recommendations.
J Clin Epidemiol 2000;53:459-68.
8. Westaway M, Stratford P, Binkley J. The Patient-Specific Functional Scale: validation of its use in persons with neck dysfunction. J Orthop Sports Phys Ther 1998;27:331-8.
9. Stratford PW, Riddle DL, Binkley JM, Spadoni G, Westaway
MD, Padfield B. Using the Neck Disability Index to make decisions concerning individual patients. Physiother Can 1999;51:
107-12.
10. Hoving JL, de Vet HC, Koes BW, et al. Manual therapy, physical
therapy, or continued care by the general practitioner for patients
with neck pain: long-term results from a pragmatic randomized
clinical trial. Clin J Pain 2006;22:370-7.
11. Cahan A, Gilon D, Manor O, Paltiel O. Clinical experience did not
reduce the variance in physicians estimates of pretest probability
in a cross-sectional survey. J Clin Epidemiol 2005;58:1211-6.
Arch Phys Med Rehabil Vol 89, January 2008

74

PSYCHOMETRIC PROPERTIES OF THE NDI AND NRS, Cleland

12. Hurwitz EL, Morgenstern H, Harber P, et al. A randomized trial

of chiropractic manipulation and mobilization for patients with
neck pain: clinical outcomes from the UCLA neck-pain study.
Am J Public Health 2002;92:1634-41.
13. 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;136:713-22.
14. Keating J, Kent PM, Davidson M, Duke RG, Mckinnon L, De
Nardis R. Predicting short-term response and non-response to
neck strengthening exercise for chronic neck pain. J Whiplash
Relat Disord 2005;4:43-55.
15. Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM.
Clinical importance of changes in chronic pain intensity measured
on an 11-point numerical pain rating scale. Pain 2001;94:149-58.
16. Childs JD, Piva SR, Fritz JM. Responsiveness of the numeric pain
rating scale in patients with low back pain. Spine 2005;30:1331-4.
17. Cleland JA, Childs JD, Fritz JM, Whitman JM, Eberhart SL. Development of a clinical prediction rule for guiding treatment of a subgroup of patients with neck pain: use of thoracic spine manipulation,
exercise and patient education. Phys Ther 2007;87:9-23.
18. Cleland JA, Glynn P, Whitman JM, Eberhart SL, MacDonald C,
Childs JD. Short-term response of thoracic spine thrust versus
non-thrust manipulation in patients with mechanical neck pain: a
randomized clinical trial. Phys Ther 2007;87:431-40.
19. Werneke M, Hart DL, Cook D. A descriptive study of the centralization phenomenon. A prospective analysis. Spine 1999;24:
676-83.
20. Jensen MP, Turner JA, Romano JM. What is the maximum
number of levels needed in pain intensity measurement? Pain
1994;58:387-92.
21. Vernon H, Mior S. The Neck Disability Index: a study of reliability and validity. J Manipulative Physiol Ther 1991;14:409-15.
22. Waddell G, Newton M, Henderson I, Somerville D, Main CJ. A
Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of
fear-avoidance beliefs in chronic low back pain and disability.
Pain 1993;52:157-68.
23. Wainner RS, Fritz JM, Irrgang JJ, Boninger ML, Delitto A,
Allison S. Reliability and diagnostic accuracy of the clinical
examination and patient self-report measures for cervical radiculopathy. Spine 2003;28:52-62.
24. Jensen MP, Karoly P, Braver S. The measurement of clinical pain
intensity: a comparison of six methods. Pain 1986;27:117-26.
25. Pengel HM, Refshauge K, Maher C. Responsiveness of pain,
disability, and physical impairment outcomes in patients with low
back pain. Spine 2004;29:879-83.
26. Flynn TW, Whitman J, Magel J. Orthopaedic manual physical
therapy management of the cervical-thoracic spine and rib cage.
San Antonio: Manipulations Inc; 2000.
27. Kendall FP, McCreary EK, Provance PG. Muscles: testing and
function. 4th ed. Baltimore: Williams & Wilkins; 1993.
28. McKenzie RA. Cervical and thoracic spine: mechanical diagnosis
and therapy. Minneapolis: Orthopaedic Physical Therapy Products; 1990.
29. Harris KD, Heer DM, Roy TC, Santos DM, Whitman JM, Wainner RS. Reliability of a measurement of neck flexor muscle
endurance. Phys Ther 2005;85:1349-55.

Arch Phys Med Rehabil Vol 89, January 2008

30. Maitland G, Hengeveld E, Banks K, English K. Maitlands vertebral manipulation. 6th ed. Oxford: Butterworth-Heinemann;
2000.
31. Spurling RG, Scoville WB. Lateral rupture of the cervical intervertebral discs: a common cause of shoulder and arm pain. Surg
Gynecol Obstet 1944;78:350-8.
32. 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, editors. Grieves
modern manual therapy. 2nd ed. New York: Churchill Livingstone; 1994. p 577-85.
33. Stratford PW, Binkley J, Solomon P, Gill C, Finch E. Assessing
change over time in patients with low-back pain. Phys Ther
1994;74:528-33.
34. Fritz JM, Irrgang JJ. A comparison of a modified Oswestry Low
Back Disability Questionnaire and the Quebec Back Pain Disability Scale. Phys Ther 2001;81:776-88.
35. Cleland JA, Fritz JM, Whitman JM, Palmer JA. The reliability and
construct validity of the Neck Disability Index and patient specific
functional scale in patients with cervical radiculopathy. Spine
2005;31:598-602.
36. Norman G, Stratford P, Regehr G. Methodological problems in
the retrospective computation of responsiveness to change: the
lesson of Cronbach. J Clin Epidemiol 1997;50:869-79.
37. Schmitt J, Di Fabio RP. The validity of prospective and retrospective global change criterion measures. Arch Phys Med Rehabil
2005;86:2270-6.
38. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing
rater reliability. Psychol Bull 1979;86:420-8.
39. Shrout PE. Measurement reliability and agreement in psychiatry.
Stat Methods Med Res 1998;7:301-17.
40. Portney LG, Watkins MP. Foundations of clinical research: applications to practice. 2nd ed. Upper Saddle River: Prentice Hall
Health; 2000.
41. Hanley J, McNeil B. The meaning and use of the area under
receiver operating characteristic (ROC) curve. Radiology 1982;143:
29-36.
42. Wyrwich K, Tierney W, Wolinsky F. Further evidence supporting
a SEM-based criterion for identifying meaningful intra-individual
changes in health related quality of life. J Clin Epidemiol 1999;
52:861-73.
43. Stratford PW, Binkley JM, Riddle DL, Guyatt GH. Sensitivity to
change of the Roland Morris Back Pain Questionnaire: part 1.
Phys Ther 1998;78:1186-96.
44. 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;47:258-63.
45. Vos CJ, Verhagen AP, Koes BW. Reliability and responsiveness
of the Dutch version of the Neck Disability Index in patients with
acute neck pain in general practice. Eur Spine J 2006;15:1729-36.
46. Ackelman BH, Lindgren U. Validity and reliability of a modified
version of the neck disability index. J Rehabil Med 2002;34:
284-7.
47. Guyatt G, Walter S, Norman G. Measuring change over time:
assessing the usefulness of evaluative instruments. J Chronic Dis
1987;40:171-8.
48. Guyatt GH, Norman GR, Juniper EF, Griffith LE. A critical look
at transition ratings. J Clin Epidemiol 2002;55:900-8.