Validity and reliability of a modified version of the neck disability index.
ABSTRACT The Neck Disability Index was tested for validity and reliability. Fifty-nine Swedish patients (28 men, 31 women) were included. Twenty patients were in the acute phase after a neck sprain, 19 had chronic neck pain and 20 had no neck pain but had other musculoskeletal symptoms. On 5 occasions, the patients completed the Neck Disability Index, the Disability Rating Index, the MOS 36-item short-form health survey, 2 visual analogue scales, for pain and overall activity and some complementary questions. Levels of sensitivity, test-retest reliability and validity were acceptable. In order to increase specificity, we modified the Neck Disability Index by clarifying that the items only referred to the pain in the neck in 9 of 10 items. Thirty-eight patients (16 men, 22 women) were included in a study of the modified version. Twenty patients with acute neck sprain and 18 with other musculoskeletal symptoms filled out the modified version of the Neck Disability Index, which provided a more specific measure of disability due to neck pain.
- SourceAvailable from: Wei-Li Hsu[Show abstract] [Hide abstract]
ABSTRACT: Study Design. Translation and psychometric testing of a questionnaireObjective. To adapt the Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire (JOACMEQ) cross-culturally to Chinese language and to investigate the psychometric properties of the translated Chinese JOACMEQ.Summary of Background Data. The recently developed JOACMEQ is a self-reported questionnaire used to assess the severity of the Cervical Spondylotic Myelopathy (CSM). Despite its acceptance in the literature, the JOACMEQ has not been translated and validated for Chinese speaking patients.Methods. Translation and cross-cultural adaptation of the original JOACMEQ was performed according to published guidelines. A total of 105 patients with a predominant complain of neck pain which included but not limited to patients with a clinically confirmed CSM were included. Psychometric testing of the responses included construct validity (factor analysis), internal consistency (Cronbach's alpha), test-retest reproducibility (Intra-class correlation coefficient) and concurrent validity (Spearman's rho). The discriminatory power of differentiating those with myelopathy from those without was assessed by the area under the receiver operating characteristic (ROC) curve and with the optimal cut-off points determined.Results. Construct validity testing revealed a comparable factorial structure of the Chinese JOACMEQ with the original JOACMEQ. The Cronbach's alpha of the Chinese JOACMEQ was 0.88 indicating good internal consistency. Good to excellent test-retest reliability was demonstrated for all five domain scores (ICC range = 0.79-0.90). A significant strong association between the Chinese NDI and the quality of life domain of the Chinese JOACMEQ (Spearman's rho = -0.76, p<0.01) was identified. The upper and lower extremity function domain scores were found to be significant with the area under the ROC curve found to be 0.75 (p = 0.04) and 0.82 (p = 0.02) respectively.Conclusion. The results indicated that the Chinese JOACMEQ successfully retained the psychometric properties of the original JOACMEQ and support the usefulness of the Chinese JOACMEQ as an appropriate supplementary diagnostic and outcome measure for Chinese patients suspected of CSM.Spine 04/2014; · 2.16 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Different types of exercises can help manage chronic neck pain. Supervised exercise interventions are widely used, but these protocols require substantial resources. The aim of this trial, which focused on adherence, was to evaluate two home exercise interventions. This parallel group randomized controlled trial included 57 women randomly allocated into two groups - a strength training group (STRENGTH, 34 subjects) and a stretching group (STRETCH, 23 subjects). The interventions focused on the neck and shoulder muscles and lasted for 12 months. The STRENGTH group performed weight training and ended each session with stretching exercises. These stretching exercises constituted the entirety of the STRETCH group's training session. Both groups were instructed to exercise three times per week. All the participants kept an exercise diary. In addition, all participants were offered support via phone and e-mail. The primary outcomes were pain intensity and function. The trial included a four- to six-month and a twelve-month follow-up. A completer in this study exercised at least 1,5 times per week during eight unbroken weeks. A responder in this study reported clinically significant improvements on pain and function. The statistical analyses used the Mann Whitney U-test, Wilcoxon signed-rank test, and X2 test. At four- to six-months, the numbers of completers were 19 in the STRENGTH group and 17 in the STRETCH group. At twelve months, the corresponding numbers were 11 (STRENGTH) and 10 (STRETCH). At four- to six-months, the proportions of subjects reporting clinically important changes (STRENGTH and STRETCH) were for neck pain: 47% and 41%, shoulder pain: 47% and 47%, function: 37% and 29%. At twelve months, the corresponding numbers were for neck pain: 45% and 40%, shoulder pain: 55% and 50%, function: 55% and 20%. No differences in the two primary outcomes between the two interventions were found, a finding that may be due to the insufficient statistical power of the study. Both interventions based on home exercises improved the two primary outcomes, but the adherences were relatively low. Future studies should investigate ways to improve adherence to home exercise treatments.Trial registration: ClinicalTrials.gov Id: NCT01876680.BMC Musculoskeletal Disorders 01/2014; 15(1):6. · 1.88 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The Neck Disability Index (NDI) is the most commonly used outcome measure for neck pain. This study aimed to determine the psychometric properties of a German version of the NDI. Cross-cultural translation and psychometric testing of the NDI were performed. The 10-item NDI was translated into German and administered to 558 patients with chronic unspecific neck pain (Mean age 49.9 +/- 11.4 years, 76% female). The factor structure and reliability of the NDI were assessed using factor analysis, Cronbach's alpha, split-half reliability (Spearman-Brown coefficient), and intra-class correlation (ICC2,1). To determine convergent validity, pain intensity (visual analog scale; VAS), pain on movement (VAS), and quality of life (Short Form 36 Health Survey Questionnaire; SF-36) were correlated with the NDI. Correlation with range of motion and sensitivity to change were also assessed in a subsample of 49 patients. The mean NDI score was 32.75 +/- 13.09. Factor analysis revealed a single factor that explained 39.8% of the variance. Cronbach's alpha was 0.81; Spearman-Brown coefficient was 0.80; and intra-class correlation was 0.81 (95% confidence interval = 0.78, 0.83). Significant correlations were found for pain intensity (r = 0.22, p < 0.01), pain on movement (r = 0.39, p < 0.01), quality of life (r = -0.30 to -0.45, p < 0.01), and range of motion (r = -0.34, p = 0.02). Patients who reported global improvement of health after an exercise or yoga intervention showed a higher decrease on the NDI than patients who reported no global improvement (p < 0.01). The German version of the NDI has a comparable factor structure as the original version, acceptable psychometric properties, and is sensitive to change after physical activity. Neck disability is associated with other measures of neck pain.BMC Musculoskeletal Disorders 03/2014; 15(1):91. · 1.88 Impact Factor
VALIDITY AND RELIABILITY OF A MODIFIED VERSION OF THE NECK
Birgitta Helmerson Ackelman and Urban Lindgren
From the Departments of Physiotherapy and Orthopedic Surgery, Karolinska Institutet, Huddinge University Hospital,
The Neck Disability Index was tested for validity and reli-
ability. Fifty-nine Swedish patients (28 men, 31 women) were
included. Twenty patients were in the acute phase after a
neck sprain, 19 had chronic neck pain and 20 had no neck
pain but had other musculoskeletal symptoms. On 5 occa-
sions, the patients completed the Neck Disability Index, the
Disability Rating Index, the MOS 36-item short-form health
survey, 2 visual analogue scales, for pain and overall activity
and some complementary questions. Levels of sensitivity,
test-retest reliability and validity were acceptable. In order
to increase speci?city, we modi?ed the Neck Disability Index
by clarifying that the items only referred to the pain in the
neck in 9 of 10 items. Thirty-eight patients (16 men, 22
women) were included in a study of the modi?ed version.
Twenty patients with acute neck sprain and 18 with other
musculoskeletal symptoms ?lled out the modi?ed version of
the Neck Disability Index, which provided a more speci?c
measure of disability due to neck pain.
Key words: neck pain, whiplash injuries, disability
evaluation, outcome assessment.
J Rehabil Med 2002; 34: 284–287
Correspondence address: Birgitta Helmerson Ackelman,
Department of Physiotherapy R 41, Huddinge University
Hospital, SE-141 86, Stockholm, Sweden. E-mail:
Submitted June 13, 2001; accepted June 4, 2002
Valid and reliable tests are cornerstones in clinical research. To
evaluate a patient’s functional status, several instruments are
available, for example, the Sickness Impact Pro?le (SIP) (1) or
the Medical Outcomes Study 36-Item Short-Form Health Survey
(SF-36) (2–4). The Neck Disability Index (NDI) described by
Vernon & Mior (5) is based on the Oswestry Low Back Pain
Disability Index (6) and speci?cally measures activity limita-
tions due to neck pain. The NDI has been tested for face validity,
test-retest reliability, internal consistency, construct validity and
concurrent validity, but the authors suggested that larger group
studies should be conducted to strengthen the overall relevance
of the NDI (5). The Quebec Task Force also later suggested such
studies (7). A translation of the NDI has been used at Huddinge
University Hospital and at Linko ¨ping University Hospital in
Sweden, but was never tested for validity.
The purpose of this study was tovalidate a Swedish version of
the Neck Disability Index and to test a version that had been
modi?ed in order to improve its speci?city.
The subjects studied were 2 groups of patients with neck pain and 1
groupof patients without neck pain but with some other musculoskeletal
disorder (Table I). One group of patients with neck pain was in the acute
phase after a neck sprain and another group consisted of patients with
chronic neck pain who had not received any form of therapy during the
study period. A general inclusion criterion was the ability to read and
speak Swedish. For the acute group, only patients with no previous neck
disorder were included. For the chronic group, painfor 3 months or more
The?rst part of thestudyconsistedof 59 patients. Nineteen subjects in
the acute phase after a neck sprain were referred from the emergency
room on the day of the ?rst visit, 20 subjects with chronic neck pain
previously treated at the department of physiotherapy and 20 subjects
with no neck pain but having other musculoskeletal disorders were
recruited from the department of physiotherapy.
The second part of the studywas designed totest amodi?ed versionof
the NDI. In that part a different cohort of 38 subjects were included (20
with acute neck sprain and 18 with no neck pain).
The NDI consists of 10 items; pain intensity, personal care, lifting,
sleeping, driving, recreation, headaches, concentration, reading and
work. The 10 items, with 6 possible answers in each are scored 0 (no
activity limitations) to 5 (major activity limitations) and summed up to
yield a total score. In the Swedish version used in the study the item
concerning driving a car was provided with an additional alternative,
The SF-36 questionnaire (2–4), produces a pro?le of eight domain
scores, including physical functioning (PF), physical role limitations,
emotional role limitations, social functioning, bodily pain, general
mental health, vitality and general health perception. Each domain is
scored from 0 (poor health) to 100 (optimal health). The only domain
studied by us was PF.
The Disability Rating Index (DRI) (8), consists of 12 items related to
physical functioning: dressing, outdoor walks, climbing stairs, sitting for
a longer time, standing bent over a sink, carrying a bag, making a bed,
running, light work, heavy work, lifting heavy objects and participation
in exercise programs or sports. On 100-mm visual analogue scales
(VAS) with anchor points 0 (without dif?culty) and 100 (unable to
perform), the patient marks his/her presumed ability to perform the
activity. The mean value of these measurements provides the DRI
expressed as a percentage of the highest possible rating.
In addition to those instruments we distributed complementary
questions concerning the adequacy and completeness of the NDI, items
concerning neck disorders, cervical range of motion and the need for
pain relievers. Two100-mm visual analogue scales were also added: one
concerning pain (anchor points “no” and “worst possible”) and one
concerning overall activity (anchor points “fully active” and “prevented
In the second part of the study we modi?ed the NDI to clarify that the
items referredspeci?cally toneck conditions. Thus “neckpain” wasused
Ó 2002 Taylor & Francis. ISSN 1650–1977J Rehabil Med 34
J Rehabil Med 2002; 34: 284–287
instead of “pain” or “due to neck pain” was added (not for the item
concerning headache). This modi?cation involved 9 of the 10 items.
Data collection for the ?rst part was carried out between 1997 and 1998.
The studies were approved by the Ethical Committee at Huddinge
UniversityHospital. In the?rstpart of thestudythesubjects ?lledout the
forms for the NDI, the SF-36, the DRI, the 2 VAS (pain and overall
activity) and answered the complementary questions. Questionnaires
were ?lled out on 5 occasions (Fig. 1); the ?rst at the initial visit to the
department of physiotherapy, the other 4 at home. However, all 5
questionnaires were completed at home for the previously treated
chronic neck pain patients. All forms completed at home were returned
individually in prepaid envelopes. In the second part of the study the
subjects with acute necksprain?lledinthe modi?ed NDI and theDRI on
2 occasions; ?rst on the day of their visit to the Emergency Room and
second in the department of physiotherapy at least 2 hours later. The
group with no neck pain having other musculoskeletal symptoms ?lled
out the questionnaires once at the department of physiotherapy.
Since the NDI originally is scored from 0 (no activity limitations) to 50
(major activity limitations), and since the scoring scheme does not
provide a strategy for dealing with questions that are unanswered, we
transformed the scoring to a percentage score as described for the
Oswestry Low Back Pain Index (6). If the alternative not applicable in
the item concerning driving a car was chosen, that item was excluded.
If less than 8 of 10 items were answered the case was considered a
The Spearman rank order correlation coef?cient was used for the test-
retest reliability of the NDI and for the correlations between different
instruments. The Wilcoxon rank sum test was used for comparisons
between instruments, groups and measurements within a group.
In the ?rst part of the study, the patients in the chronic group
?lled out 94% of the items satisfactorily; the percentage in the
group without neck symptoms and in the acute group was 86%
and 88%, respectively. In the second part of the study the items
of the modi?ed NDI were satisfactorily ?lled out by 92% of the
group with no neck pain and by 88% of the group with acute
Face validity was estimated based on the subject’s response
during the initial evaluation as to whether the questionnaire was
relevant to his/her disorder. Sixteen patients in the chronic neck
pain group, 13 of the patients in the acute neck pain group and 2
in the no neck pain group felt that it was relevant or partially
relevant (Table II).
The content validity was estimated based on the response to
the question: “Is there something important you think should be
added?”. For results see Table III.
The concurrent validity was calculated as a rank correlation
using the initial evaluation of NDI/DRI, NDI/PF, NDI/VAS pain
and NDI/VAS activity. For results see Table IV.
The sensitivity was estimated based on the response to the
statement in the initial complementary questionnaire, “I have a
neck disorder”. Twenty subjects in the chronic neck pain group,
19 subjects in the acute neck pain group answered “correct” or
“partially correct”. The NDI for the neck pain subjects was
plotted in a diagram (Fig. 2).
The speci?city was estimated for the initial evaluation using a
diagram, in which the percentages of the NDI for the subjects
with no neck pain were plotted (Fig. 3). Six subjects inthe group
with no neck pain had an NDI score over 10%, 4 of these over
20%. The items most frequently misunderstood were; “dif?culty
in lifting” (3 subjects), “working” (6 subjects), “sleeping” (5
subjects) due to pain and the item concerning pain (3 subjects).
When analysing the individual answers, 2 subjects in the group
without neck pain answered “partially correct” to the statement
“I have a neck disorder”. These 2 subjects did not mark any pain
on the VAS.
The test-retest reliability coef?cient for the subjects with
chronic neck pain within 48 hours (initially/3 months), 3 weeks
and 3 months were rs= 0.94–0.99 (Table V). For the measure-
ments within 48 hours for the subjects with acute neck pain
rs= 0.81–0.89 (Table V).
The modi?ed NDI for the neck pain patients is presented in Fig.
4. The percentage scores of the modi?ed NDI were 18–64% for
the subjects with acute neck pain. The speci?city of the modi?ed
NDI is presented in Fig. 5. The scores of subjects with no neck
pain were all below 20% and there was a signi?cant difference
when comparing the modi?ed NDI and the DRI (p<0.0001).
The test-retest correlation coef?cient for the modi?ed NDI was
rs= 0.97 (n = 16).
Table I. Patient groups in study
M/WDuration of pain
Chronic neck pain 20 7/13
No neck pain
Acute neck pain
No neck pain
Acute neck pain
46.5 (25–63)8 months–20 years
20 10/10 31 (18–70)
19 11/838 (21–62)
Fig. 1. Test-retest reliability conducted in the study.
J Rehabil Med 34
Neck Disability Index validity and reliability
When measuring disability or activity limitation one can choose
between generic, condition speci?c and patient speci?c instru-
ments. The SF-36 and the DRI belong to the generic group (8).
Condition speci?c instruments like the NDI are thought to be
more sensitive to changes in symptoms (9–11) or easier to use in
clinical practice (12, 13). The study by Riddle & Stratford (13)
describes similarities in results when using the SF-36 and the
NDI (5). Westaway et al. (14) compared the Patient-Speci?c
Functional Scale (PSFS) to the NDI. In the PSFS the patient
chooses which 3 important personal activities are the most
dif?cult to perform (14). The authors concluded that the PSFS is
an excellent tool for working with individual patients and that it
should be supplemented with a generic or condition-speci?c
measure when group decision-making (i.e. quality assurance
assessments or research) is the goal. They propose that when
assessing the functional status of patients with cervical spine
problems, either the physical component summary scale or the
mental component summary scale of the SF-36 or the NDI can
be used since there is considerable overlap between the two
tests. The Swedish version of the NDI, which was used in the
?rst part of our study, demonstrated good validity, sensitivity
and test-retest reliability, but not optimal speci?city. Therefore,
in a modi?ed version, we made itclear that the items speci?cally
referred to neck pain. In the study by Vernon & Mior (5), test-
retest reliability of the NDI was measured on a subset of 17
subjects with neck pain of different origin. The subjects
completed the NDI within 48-hours. The correlation coef?cient
was rp= 0.89. In the ?rst part of our study we chose groups of
neck pain patients who represented an acute or a chronic stage.
The results showed a similar test-retest reliability within 48
Table II. Initial evaluation, face validity of the NDI. Answer to the
question “Do you feel that the questions are relevant to your
disorder?” Answers each group respectively
Chr = chronic neck pain, Nl1= no neck pain, Ac1= acute neck pain.
Table III. Part one. Content validity of the Neck disability index. All
answers, collected at time-points 1–3 to the question “Is there
something important that you think should be added?” are listed.
Each statement refers to 1 subject
Group Chronic neck pain
.“my life is almost destroyed because of pain”—“when one has
pain even the family suffers”
.“I have numbness in my ?ngers and hands which I believe is
connected to the neck disorder”
.“not the pain itself but it’s consequences-isolation , numbness,
.“the items are to distinct”
.“if numbness appears in different situations. Expand the item on
lifting also to include carrying”
.“the alternatives are not entirely relevant”
.“which kind of pain, the degree of self con?dence, dizziness when
moving my head”
.“hard to ?nd correct answer, a place to write my own comments”
Group Acute neck pain.
.“sore muscles when I bend forward and to the right”
.“locking sensation when opening my mouth”
.“depression …, no use …, a sense of not ?tting in …, pain when
lifting heavy objects from the ?oor”
Table IV. Rank correlations on initial evaluation between Neck
Disability Index (NDI) and Disability Rating Index (DRI) or
Physical Functioning of the SF-36 (PF) or VAS pain or VAS
Chr = chronic neck pain, Nl1= no neck pain, Ac1= acute neck pain.
n.s. = non signi?cant, all other correlations signi?cant p<0.01.
Fig. 2. Distribution of the Neck Disability Index (NDI) for the
chronic neck pain group (squares n = 20) and the acute neck pain
group (triangles n = 17) in part 1. Every dot represents 1 subject.
Fig. 3. Neck Disability Index (NDI) for the group with no neck pain
(n = 20) in the ?rst part. Every dot represents 1 subject. Four
subjects reach an NDI over 20%.
J Rehabil Med 34
B. Helmerson Ackelman and U. Lindgren
hours for the group with acute neck pain (correlation coef?cient
of rs= 0.81–0.89). For the group with chronic neck pain we
found rs= 0.97 within 48 hours and within 3 months rs= 0.94.
The NDI for the neck pain subjects was well distributed and
neither ceiling nor ?oor-effects could be seen. Because of the
results from the ?rst part of the study, we made the modi?ed
version. This version was clearly better, since non-speci?c
ailments or co-morbidity did not produce false increases in the
scores. Co-morbidity/other ailments may still be a problem
when patients try to estimate the level of a particular disability.
For example, the item dealing with headache is not necessarily a
part of the neck pain syndrome. As for the concurrent validity, a
high degree of pain for the chronic neck pain patients did not in
most cases correspond with high levels of the NDI. This could
indicate that the pain remains but the patient has learned to
compensate for his/her disorder in different kinds of activities.
Finally, in this study the suggestions for additional items
indicated that the social consequences of the pain isan important
part of the subjects’ situation which is not covered when using
In conclusion, it should be clearly stated for each item that
only disability due to neck pain is of interest. This modi?cation
of the NDI is a valid and reliable instrument to measure dis-
ability due to neck pain.
This work was supported by grants from Karolinska Institutet, Board of
Research for Health and Caring Sciences.
1. Bergner M, Bobbitt R, Carter W, Gilson B. The sickness impact
pro?le: development and ?nal revision of a health status measure.
Med Care 1981; 8: 787–805.
2. Sullivan M, Karlsson J, Ware J. The Swedish health survey—I.
Evaluation of data quality, scaling assumptions, reliability and
construct validity across general populations in Sweden. Soc Sci
Med 1995; 10: 1349–1358.
3. Ware JE, DonaldSherbourne C. The MOS 36-itemshort-formhealth
survey (SF-36). I. Conceptual framework and item selection. Med
Care 1992; 30: 473–483.
4. Ware JE, Kosinski M, Bayliss M, McHorney C, Rogers W, Raczek
A. Comparison of methods for the scoring and statistical analysis of
SF-36 health pro?le and summary measures: summary of results
from the Medical Outcomes Study. Med Care 1995; 33: AS264–279.
5. Vernon H, Mior S. The neck disability index: a study of reliability
and validity. J Manipulative Physiol Ther 1991; 14: 409–415.
6. Fairbank JT, Couper J, Davies JB, O’Brien JP. The Oswestry low
back pain disability questionnaire. Physiother 1980; 66: 271–273.
7. Quebec Task Force on WAD. Scienti?c monograph of the Quebec
Task Force on Whiplash-Associated Disorders: Rede?ning ‘whip-
lash’ and its management. Spine 1995; 20: 1–273.
8. Sale ´n BA, Spangfort EV, Nygren AL, Nordemar R. The disability
rating index: an instrument for the assessment of disabilityinclinical
settings. J Clin Epidemiol 1994; 47: 1423–1434.
9. Kopec JA, Esdaile JM, Abrahamowicz M, Abenhaim L, Wood-
Dauphinee S, Lamping DL, Williams JI. The Quebec Back Pain
Disability Scale. Measurement properties. Spine 1995; 20: 341–352.
10. Patrick DL, Deyo RA. Generic and disease-speci?c measures in
assessing health status and quality of life. Med Care 1989; 27: S217–
11. Stucki G, Liang M, Fossel A, Katz J. Relative responsiveness of
condition-speci?c and generic health status measures in degenera-
tive lumbar spinal stenosis. J Clin Epidemiol 1995; 48: 1369–1378.
12. Feinstein AR, Josephy BR, Wells CK. Scienti?c and clinical
problems in indexes of functional disability. Ann Intern Med 1986;
13. Riddle DL, Stratford PW. Use of generic versus region-speci?c
functional status measures on patients with cervical spine disorders.
Phys Ther 1998; 78: 951–963.
14. Westaway MD, Stratford PW, Binkley JM. The patient-speci?c
functional scale: validation of its use in persons with neck dys-
function. J Orthop Sports Phys Ther 1998; 27: 331–338.
Table V. Neck disability index reliability. The Spearman correla-
tion coef?cient (rs) between evaluation 1 and 2 (48 hours), 1 and 3
(3 weeks), 1 and 4 (3 months), 4 and 5 (48 hours after 3 months) for
the chronic group. Correlation between evaluation 1 and 2 (48
hours) and 4 and 5 (48 hours after 3 months) for the acute neck
n = 18
n = 16
n = 18
n = 19
n = 18
n = 15
Chr = chronic neck pain, Ac1= acute neck pain.
** p<0.01. *** p<0.001.
Fig. 4. Neck Disability Index (NDI) modi?ed for the acute neck
pain group in part 2, n = 16. Every dot represents 1 subject.
Fig. 5. Modi?ed Neck Disability Index (NDI) for the group with no
neck pain in part 2 (n = 17). Every dot represents 1 subject.
J Rehabil Med 34
Neck Disability Index validity and reliability