ARTICLE IN PRESS

1356-689X/$ -

doi:10.1016/j.m

�Correspondfax: +1412 38

E-mail add

Manual Therapy 11 (2006) 321–330

www.elsevier.com/locate/math

Technical and measurement report

Inter-tester reliability of passive intervertebral and active movementsof the cervical spine

Sara R. Pivaa,�, Richard E. Erhardb,c, John D. Childsd, David A. Browderd

aSchool of Health and Rehabilitation Sciences, Department of Physical Therapy, University of Pittsburgh, 6035 Forbes Tower,

Pittsburgh, PT 15260, USAbUniversity of Pittsburgh Medical Center Health System’s Spine Specialty Center, USA

cSchool of Health and Rehabilitation Sciences, Department of Physical Therapy, University of Pittsburgh, USAdDepartment of Physical Therapy, Wilford Hall Medical Center, San Antonio, TX, USA

Received 4 February 2004; received in revised form 1 August 2005; accepted 21 September 2005

Abstract

Measurements of active range of motion (AROM) and passive intervertebral movements (PIM) of the cervical spine

are frequently used for patients with neck pain. However, there is a paucity of studies that investigate the psychometric pro-

perties of these measurements. Objectives of this study were to: (1) determine the inter-tester reliability of PIM, AROM,

and the effects of AROM on symptom provocation; (2) establish the minimal detectable change (MDC) in cervical AROM;

and (3) determine the association between AROM and disability. Thirty subjects (age 41712) with neck pain participated

in this study. Two masked examiners performed the measurements during the same testing session. PIM was assessed manually

and recorded as hypomobile or normal. AROM was measured in degrees with a gravity goniometer. The effect of

AROM on patient’s symptoms was recorded as no change, decreased, increased, centralization, or peripheralization.

Measures of AROM had moderate to substantial reliability (.78–.91) and resulted in a MDC adequate for clinical use

(from 91 to 161). The effect of AROM on symptom provocation resulted in Kappa values that ranged from slight to substantial

(.25–.87). Measures of PIM resulted in substantial and moderate reliability of assessing occipital–atlas mobility, tenderness

of the transverse processes of atlas, and symptom provocation during PIM testing of the lower cervical segments. Fair Kappa

values were observed during judgment of mobility in the C2 segment and symptom reproduction during PIM of C2 and C5.

The additional PIM had Kappa values that ranged from none to slight. Low prevalence of positive findings likely resulted

in an artificial deflation of the Kappa statistic during some PIM measures. Measures of AROM in saggital and transverse

planes were associated with disability scores (r ¼ :43 and :40; respectively). Findings are relevant to the planning of future

studies to establish the criterion validity of these tests to guide the selection of interventions and establish prognosis in patients

with neck pain.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: Consistency; Measurement; Neck pain; Repeated measures study

1. Introduction

Measurements of active range of motion (AROM)and passive intervertebral movements (PIM) of thecervical spine are routinely used as part of the physical

see front matter r 2005 Elsevier Ltd. All rights reserved.

ath.2005.09.001

ing author. Tel.: +1412 383 6712;

3 6629.

ress: srpst24@pitt.edu (S.R. Piva).

examination of patients with neck pain. Tests of AROMand PIM are used to identify impairment in motion andpossible segmental level that may be the source of thepatient’s complaint.(Magee, 1997) Measurements ofAROM are performed to determine limitations inmotion, patient’s willingness to move, and to identifythe range of movement in which the patient reportssymptoms. During AROM examiners note the quantity,quality, and provocation of the patient’s complaints in

ARTICLE IN PRESSS.R. Piva et al. / Manual Therapy 11 (2006) 321–330322

each direction of movement. (Magee, 1997) The effect ofeach movement on the patient’s symptoms is also usefulto guide treatment decisions (i.e. knowing whether thesymptoms increase or decrease or whether the symptomscentralize or peripheralize, etc.) (McKenzie, 1990;Werneke and Hart, 2001). It helps clinicians understandwhich movements should be avoided and which move-ments should be used during treatment. For example,centralization of the patient’s symptoms during flexionprovides some indication that exercises in a flexedposition may be beneficial.

PIM are performed to assess the amount of motionavailable at each spinal segment and the provocation ofsymptoms during movement of each segment of thecervical spine. PIM are necessary for normal physiologica range of motion to occur. (Maitland, 1986) It isbelieved that the decreased PIM motion (hypomobility),which may be caused by muscle spasm or ligamentoustightness, may be associated with neck dysfunction.(Erhard, 1996; Childs et al., 2003)

In general, previous studies examining the reliabilityof AROM measurements of the cervical spine havedemonstrated acceptable levels of reliability. (Jordan,2000; Mannion et al., 2000; Petersen et al., 2000;Solinger et al., 2000) However, many of these studiesinvestigated reliability in asymptomatic individuals,(Youdas et al., 1991; Nilsson, 1995; Hole et al., 2000)others performed inappropriate statistics to estimatereliability, (Kadir et al., 1981; O’Driscoll and Tomen-son, 1982; Capuano-Pucci et al., 1991) some used onlyvisual estimation of movements, (Viikari-Juntura, 1987;Pool et al., 2004) and some used measurement tools notpractical for clinical use (Rheault et al., 1992). Fewstudies used simple measurement tools such as theuniversal goniometer or gravity goniometer (Tucci et al.,1986; Youdas et al., 1991). Furthermore, these studieshave not reported the precision or the error associatedwith these measurements. A systematic review thatevaluated 21 studies that assessed the reliability of toolsto measure cervical range of motion concluded thatmore rigorous studies were necessary (Jordan, 2000). Amore recent study reported acceptable reliability formeasures of cervical AROM when a gravity goniometerwas used and, to our knowledge, this is the only studythat determined the error associated with their measure-ments (Wainner et al., 2003).

Not many studies have investigated the reliability ofassessing PIM. To date, only one study investigated theinter-observer reproducibility of the patient’s painresponse to movement, and a small number of studieshave reported on the reliability of PIM. Pool et al.(2004) performed a study in which an 11-pointnumerical rating scale was used to report pain responseduring movement. In addition to assess the reliability ofpatient’s pain response, Pool et al. studied the consis-tency of assessing PIM from the occiput to T2. They

reported Kappa values from �.09 to .63. However,definitions of the assessment techniques were notclear. They only described that movements betweenthe occiput and atlas used a flexion technique, assess-ment of atlas/axis mobility used a rotation technique,and segments from C2 to T2 included fixation of thelower segmental level and lateral flexion to the right andto the left (Pool et al., 2004). Fjellner et al. (1999)studied PIM on normal healthy subjects. Therefore,their results may not apply to patients with neck pain.Smedmark et al. (2000) studied inter-tester reliability inassessing PIM of four tests of the cervical spine: C1/C2rotation, C3/C4 lateral flexion, C7 flexion/extension,and movement of the first rib. They reported Kappavalues from .28 to .43.

In addition to further investigate the reliability oftesting AROM, PIM, and symptom provocation duringmovement, it would be helpful to determine if measuresof AROM are associated with measures of disability,thus helping to establish validity for these measure-ments. Therefore, the objectives of this study were to: (1)determine the inter-tester reliability of PIM, AROM,and the effects of AROM on symptom provocation; (2)establish the minimal detectable change (MDC) incervical AROM; and (3) determine the associationbetween AROM and disability.

2. Methods

This study utilized a single group repeated measuresdesign.

2.1. Subjects

This study consisted of consecutive patients referredto the University of Pittsburgh Medical Center HealthSystem’s Spine Specialty Center with a primary com-plaint of neck pain. The following inclusion criteria wereused: (1) age between 18 and 75 years; (2) presence ofsymptoms in the neck, scapula, or head areas observedon the pain diagram; (3) less than 60% score on theNeck Disability Index (NDI) (Vernon and Mior, 1991).Based on our clinical experience, scores above 60%indicate that the patient is experiencing a high level ofdisability in which repeating the examination procedurefor reliability purposes could excessively exacerbate thepatient’s symptoms. Patients were excluded if the neckpain was associated with inflammatory or congenitalanomalies, presence of dizziness, or neurological signsand symptoms.

This study was approved by the University ofPittsburgh Institutional Review Board, and all subjectsprovided informed consent prior to participate inthe study. Thirty subjects with neck pain agreed toparticipate.

ARTICLE IN PRESS

Fig. 1. Measurement of active neck extension.

Fig. 2. Measurement of active neck flexion.

S.R. Piva et al. / Manual Therapy 11 (2006) 321–330 323

2.2. Measure of disability

The NDI was used to quantify the level of disability.The NDI is a reliable, valid and frequently usedcondition-specific disability scale for patients with neckpain (Vernon and Mior, 1991; Stratford et al., 1999).The disability score from the NDI was used in theassessment of the relationship between disability andAROM.

2.3. Measurements of AROM and PIM

2.3.1. Cervical AROM

Extension, flexion, rotation in full flexion, left andright lateral bending, and left and right rotation weretested in the order as described. Active movements weremeasured in degrees using a gravity goniometer (MIEMedical Research Ltd, Leeds, UK). Measures ofextension, flexion, rotation in full flexion, and lateralbending were performed with the patient seated on anexamination table. Before initiating measurements,subjects were asked to ‘‘sit up and look straight ahead’’.Measurement of rotation was performed with thepatient positioned in supine. Prior to the measurements,the gravity goniometer was zeroed by placement on ahorizontal surface.

Extension was measured by placing the gravitygoniometer on the top of the patient’s head in thesaggital plane. Patients were asked to bend the headbackward as far as possible (Fig. 1).

Flexion was measured with the gravity goniometerin the same position as for the extension measure-ment. The gravity goniometer was not removed fromthe patient’s head during extension and flexion.Patients were asked to bend the head forward asfar as possible and try to touch the chest with the chin(Fig. 2).

Rotation in full flexion was used to grade theatlantoaxial rotation. Patients were in full neck flexionand the gravity goniometer was positioned on the backof patient’s head in the frontal plane. Patients wereasked to rotate the head to the left and right as far aspossible (Fig. 3).

Lateral bending was measured with the gravitygoniometer in the frontal plane on the top of thepatient’s head. To measure lateral bending, patientswere asked to touch the left and right ear to the left andright shoulder respectively.

Rotation was measured with the patient in the supineposition with the head resting on a pillow. To allow pureaxial rotation, the patients were permitted to lift theirhead of the pillow before rotation. The gravitygoniometer was in the transverse plane on the top andmidline of the forehead. To measure rotation, patientswere asked to rotate the head to the left and after to theright as far as possible (Fig. 4).

ARTICLE IN PRESSS.R. Piva et al. / Manual Therapy 11 (2006) 321–330324

2.3.2. Effect of movement on symptoms

The effect of AROM on participant’s symptoms wasrecorded such as: no effect, increases symptoms,decreases symptoms, centralizes symptoms, or periph-

Fig. 3. Measurement of active neck rotation in full flexion.

Fig. 4. Measurement of active

eralizes symptoms. Centralization is the patient reportthat the neck movement has caused the symptoms tomove from an area more distal or lateral in the arms orshoulder girdle to a location more central or near themidline position in the cervical spine. Alternatively,peripheralization is the patient’s report that the neckmovement moved the symptoms from an area moreproximal in the cervical spine to an area more distal orlateral (McKenzie, 1990; Werneke and Hart, 2001).

2.3.3. Cervical PIM

PIM testing was used to qualitatively determine theamount of motion that occurred at each spinal segment.PIM for the cervical spine involves the palpation of eachmotion segment during passive movement of the headand neck. Measurements were performed with thepatient in the supine position and in the order described.Mobility was recorded as: (1) normal; or (2) hypomo-bile. Pain reproduction during each movement wasrecorded as: (1) pain; (2) no pain.

Atlanto-occipital joint—lateral glide: The examinerheld the patient’s head with a neutral relation-ship between the occiput and atlas and performed aleft lateral-glide of the occiput on atlas, followed by aright lateral-glide (Erhard, 1996; Magee, 1997). Ifdecreased mobility was noted to one side compared tothe opposite side, the test was considered positive forhypomobility.

Atlanto-occipital joint—lateral displacement of axis:The examiner stabilized the axis by placing the leftthumb on the left side of the spinous process of the axis.Then the examiner used the right hand to laterally bendthe head to the right. The test was then repeated to the

neck rotation in supine.

ARTICLE IN PRESSS.R. Piva et al. / Manual Therapy 11 (2006) 321–330 325

opposite side (Erhard, 1996). If decreased mobility wasnoted to one side compared to the opposite side, the testwas considered positive for hypomobility.

Tenderness over the transverse processes of atlas: Theexaminer gently palpated the transverse processes of theatlas and recorded the presence of symptoms on eitherside as positive.

Atlanto-axial joint—full flexion: The examiner sup-ported patient’s head with both hands and passivelymoved the patient’s neck into maximal flexion. Whilemaintaining this position, the patient’s head was rotatedfirst to the left and then to the right (Erhard, 1996). Ifthe rotation to one side was decreased compared to theopposite side, the test was considered positive forhypomobility.

Atlanto-axial joint—full lateral bending: The examinerpassively moved the patient’s neck into end range ofright lateral-bending. While maintaining this position,the patient’s head was rotated to the left. The test wasthen repeated to the opposite side (Erhard, 1996). If therotation to one side was decreased compared to theopposite side, the test was considered positive forhypomobility.

Mid and lower cervical: The examiner stood atthe head of the examination table and used the abdomento exert a constant pressure against the apex ofthe patient’s skull to stabilize the head but allowfree neck movements. Each spinal level beginning atC2 was glided laterally to the left and right.The examination progressed inferiorly to C6. Ifdecreased mobility was noted to one side compared tothe opposite side, the test was considered positive forhypomobility (Erhard, 1996; Hertling and Kessler, 1996;Magee, 1997).

2.4. Procedures

Patients attended one testing session lasting approxi-mately 20min. During the testing session, each patientremained inside an examination room. To warrantexaminer’s masking, the two examiners entered theexamination room independently, performed and re-corded the measurements, and then left the room. Theassessment results were not shared with the otherexaminer. To minimize the possibility that the PIM bythe first examiner would cause a true change in thepatient’s symptoms and restriction in motion, bothexaminers performed the AROM tests before the PIM.Therefore, each examiner entered the room twice. TheAROM and PIM testing were always performed in thesame order. The order of the examiners was varied foreach new patient (i.e. examiner 1 performed the examfirst for subject 1; examiner 2 performed the exam firstfor subject 2, and so on). Examiners were trained inmanual therapy and had different levels of experience(10 and 2 years, respectively). Examiners and investiga-

tors met once during a 2-h session before data collectionto review operational definitions and practice theprocedures to ensure standardization.

2.5. Sample size estimation

The sample size was calculated a priori usingSamplePowerTM statistical software (SPSS Inc., Chica-go, Illinois) (SPSS, 1998) based on the calculation ofCohen Kappa coefficients on a dichotomous variable(i.e. hypomobile or normal). To ensure sufficientstatistical power to achieve a lower bound of the 95%confidence interval for Kappa of 0.30, assuming Kappawould be equal to 0.60, we would need a sample size of30 subjects (Cohen, 1988).

2.6. Data analysis

Descriptive statistics, including frequency countsfor categorical variables and means and standarddeviations for continuous variables were calculatedto summarize the data. Cohen’s Kappa statisticand the associated 95% confidence intervals wereused to calculate inter-tester reliability of the classifica-tion of mobility and presence of pain during thePIM, and the effect of AROM on the patient’ssymptoms (Cohen, 1960; Simel et al., 1991). Theagreement of the effect of AROM on the patient’ssymptoms was based on both examiners placing thepatient in the same category of the five possible ones(e.g. both say increases symptoms; or both saycentralizes symptoms). Intra-class correlation coefficient(ICC), formula 2,1, and its 95% confidence interval werecalculated to determine the inter-tester reliability formeasurements of cervical AROM (Shrout and Fleiss,1979; Simel et al., 1991). Values less than 0.10 indicatevirtually no agreement; 0.11–0.40 indicate slight agree-ment; 0.41–0.60 indicate fair agreement; values between0.61 and 0.80 indicate moderate agreement; and valuesgreater than 0.81 indicate substantial agreement(Shrout, 1998).

The results of the reliability analyses were used tocalculate the standard error of measurement (SEM) andthe MDC. The SEM was calculated as sdO1�r, where r

is the test–retest reliability coefficient and sd is thestandard deviation of the combined scores of bothexaminers (Stratford and Goldsmith, 1997). The MDCwas calculated as 1.96 O2 SEM (Portney and Watkins,1993). In this formula 1.96 is the standard normal scoreassociated with a two-tailed 95% confidence intervaland the O2 is included to reflect the fact that there ismeasurement error associated with both the first andsecond repeated measures when calculating test–retestreliability. The Pearson correlation coefficient wascalculated to determine the association between theNDI scores and AROM.

ARTICLE IN PRESSS.R. Piva et al. / Manual Therapy 11 (2006) 321–330326

3. Results

Demographic characteristics of the 30 subjects whoparticipated in the study are reported in Table 1.

Means and standard deviations, ICC values, thecorresponding SEM and MDC values of AROMmeasurements are depicted in Table 2. The ICC valuesfor measures of AROM ranged from moderate tosubstantial (from .78 to .91) and the respective MDCvalues varied from 91 to 161. Kappa values for symptomreproduction during AROM are also depicted in Table 2.Kappa values of right rotation in flexion and left lateralbending were only slight. Symptom reproduction during

Table 1

Summary statistics of participants

Number ¼ 30 Neck pain patients

Gender 60% female

Age Median 41.5

Mean (SD) 41 (12)

Race 87% White

7% Afro-american

3% Hispanic

3% Asian

Pain (Numeric pain scale) Median 4.5

Mean (SD) 4.7 (2.4)

NDI Median 20

Mean (SD) 24.3 (14.8)

Gender and race data is reported as a percentage of participants,

whereas age, pain, and Neck Disability Index (NDI) data are reported

as median, mean, and standard deviation (SD).

Table 2

Means, standard deviations, ICC, SEM and MDC values for measures of A

for symptom reproduction during AROM

N ¼ 30 Mean (SD)a

(degrees)

ICCb (95% CI)

AROM

SEMc (degre

Extension 48 (15) .86 (.73: .93) 5.6

Flexion 60 (13) .78 (.59: .89) 5.8

Le rotation in

flexion

39 (10) .89 (.78: .95) 3.2

Rf rotation in

flexion

39 (13) .78 (.60: .89) 5.3

L lateral bending 39 (11) .85 (.70: .92) 4.2

R lateral bending 41 (11) .87 (.75: .94) 3.7

L rotation 68 (13) .91 (.82: .96) 4.1

R rotation 68 (14) .86 (.74: .93) 4.8

aMeans and SDs are based on rater 1 whilst the SEM and MDC are basebIntraclass correlation coefficient.cStandard error of measurement.dMinimum detectable change.eLeft.fRight.

flexion had a substantial Kappa value (.87). Theadditional Kappa values of symptom reproduction weremoderate (ranged from .65 to .76).

Kappa values, percentage of agreement, and preva-lence of positive findings for measurements of PIM arereported in Table 3. Kappa values for PIM weresubstantial for occipital-atlanto joint hypomobilitytested with side glides (.81) and tenderness over thetransverse processes of atlas (.83). Kappa values weremoderate for symptom reproduction during test ofatlanto–axial joint in full lateral flexion (.61) andsymptom reproduction during PIM of C4 and C6 (.65and .76, respectively). Fair Kappa values were observedduring judgment of mobility in the C2 segment (.46) andsymptom reproduction during PIM of C2 and C5 (.42and .55, respectively). Other PIM tests demonstratedslight or no agreement.

Increased levels of disability on the NDI were fairlyassociated with decreased total saggital plane AROM(flexion+extension) (r ¼ :43) and total transverse planeAROM (left rotation+right rotation) (r ¼ :40).

4. Discussion

An important element of the validity of measure-ments, and the subsequent ability to accurately interpretthese measurements, relies on the evidence of satisfac-tory reliability and measurement error (Hains et al.,1998). Poor reliability and high levels of measurementerror reduce the usefulness of a test and limit the extentto which test results can be generalized (Hains et al.,1998). This study has shown that measures of cervical

ROM; and Kappa values, and percentage of agreement between raters

es) MDCd (degrees) Kappa (95% CI)

symptom

reproduction

Agreement

16 .65 (.54: .76) 83%

16 .87 (.81: .94) 93%

9 .69 (.59: .78) 83%

15 .25 (.12: .39) 63%

12 .28 (.15: .41) 63%

10 .75 (.66: .84) 87%

11 .74 (.64: .84) 87%

13 .76 (.67: .84) 87%

d on the combined raters’ scores.

ARTICLE IN PRESS

Table 3

Kappa values, percentage of agreement between raters, and prevalence of positive tests for measurements of PIM

Kappa (95% CI) Agreement Prevalence

Occipital-atlanto joint—lateral glide Mobility .81 (.72: .91) 93% 27%

Pain .32 (.15:. 49) 77% 27%

Occipital-atlanto joint—lateral displacement of axis Mobility .35 (.08: .62) 90% 10%

Pain .35 (.15: .55) 83% 17%

Tenderness over transverse processes of atlas .83 (.74: .92) 93% 30%

Atlanto-axial joint—full flexion Mobility .21 (.08: .34) 59% 63%

Pain .36 (.24: .49) 68% 54%

Atlanto-axial joint—full lateral flexion Mobility .30 (.17: .43) 64% 59%

Pain .61 (.5: .72) 89% 63%

C2 Mobility .46 (.33: .59) 76% 38%

Pain .42 (.28: .56) 76% 31%

C3 Mobility .25 (.12: .38) 62% 52%

Pain .29 (.16: .43) 66% 45%

C4 Mobility .27 (.13: .40) 63% 50%

Pain .65 (.54: .76) 83% 48%

C5 Mobility .18 (.03: .33) 63% 40%

Pain .55 (.43: .67) 79% 41%

C6 Mobility �.07 (�.34: .20) 77% 19%

Pain .76 (.64: .87) 92% 23%

Kappa is calculated for mobility (normal or hypomobile) and pain (pain or no pain).

S.R. Piva et al. / Manual Therapy 11 (2006) 321–330 327

AROM performed with a gravity goniometer arereliable and acceptable for clinical use. Reliability refersto the consistency of a measurement to yield the sameresults when the testing procedure is repeated on aspecific population and the construct measured by thetest has not changed (Guyatt et al., 1992; Shrout, 1998).Interpretation of the confidence intervals around theICC values for cervical AROM leads to the conclusionthat, even considering the worst-case scenario (lowerboundaries of the 95% CI ¼ .6), the reliability of anycervical AROM measured with a gravity goniometer isstill satisfactory for clinical use.

Measurement error, determined in this study bycalculating the MDC, provides a threshold for inter-preting the measurements over time. For example, whenthe AROM value for extension or flexion changes morethan 161, one can be reasonably confident that truechange has occurred beyond that which can beattributable to measurement error. Knowledge of theMDC is essential when investigating the effect ofinterventions on change in cervical AROM in patientswith neck pain. Our results of reliability of cervicalAROM were similar to prior studies that investigatedpatients with neck dysfunction. Tucci et al. (1986)reported ICC values from .80 to .91 and Wainner et al.(2003)reported ICC values from .63 to .84 . Our valuesof measurement error were similar to the resultsreported by Wainner et al. (2003) (SEM from 4.61 to7.31). No previous studies have reported the MDC.

In addition to the reliability and precision ofmeasurement, understanding the relationship betweencervical AROM and measures of disability helps to

establish the validity of AROM measurement and canhelp clinicians interpret the meaning of this measure-ment (Hains et al., 1998). Therefore, we tested if themeasures of AROM were associated with disabilityscores. Although the associations were only fair,(Portney and Watkins, 1993) explaining no more than18% (r2 ¼ .432) of the variability in disability, the resultsseem to indicate that physical therapists should payattention to changes in total saggital and transverseplanes of motion when assessing patients with neckpain. Improvement in AROM in these planes willprobably be relevant and may reflect in better function.These relationships make empirical sense, since ade-quate range of motion in the saggital and transverseplanes are required in most activities of daily living suchas desk or computer work, driving, housekeeping,grooming, and eating.

The results of this study indicate that Kappa valuesfor symptom reproduction during AROM can beconsistently reproduced and assessed for movements inthe saggital (flexion and extension) and transverseplanes (rotations). However, right rotation in full flexionand left lateral bending has low reliability. Because wecannot explain why left rotation in flexion and rightlateral bending had better reliability values than thesame movements to the contralateral side, we advisecaution regarding interpretation of the consistency ofmeasuring symptom reproduction during rotation inflexion and lateral bending in general. During clinicalpractice we have noticed more discrepancy in symptomreproduction during repeated lateral bending thanduring movements in the saggital or transverse planes.

ARTICLE IN PRESSS.R. Piva et al. / Manual Therapy 11 (2006) 321–330328

During repeated lateral bending sometimes the patientsreport local symptoms whereas at other times thesymptoms radiate to the arm or show no change at all.Therefore, it may be that the symptoms producedduring movements of lateral bending and rotation inflexion truly change when the test is repeated. Ourresults are comparable to the ones reported by Poolet al. (2004). They recorded pain reproduction using an11-point pain scale and calculated reliability using theICC. The study reported ICC of .71 for provoked painin extension, .63 for flexion, .70 for right rotation, .66for left rotation, .65 for right lateral bending and .45 forleft lateral bending. Rotation in full flexion was notinvestigated in that study (Pool et al., 2004).

We proposed to test AROM of neck rotation in fullflexion to grade atlanto-axial rotation. Although thereliability of this measurement was good, we are notsure if this test really measures atlanto-axial rotation oratlanto-axial dysfunction. We theorize that in full neckflexion, because the lower cervical spine segments are ina fully opened position, the additional movement inrotation would have to come from the upper neck, morelikely the atlanto-axial joint (Hertling and Kessler, 1996;Bogduk and Mercer, 2000). We have not observed anyassociation between this test and the traditional PIM totest the atlanto-axial joint in full flexion (Hertling andKessler, 1996). Because this study was not intended totest diagnostic accuracy, we cannot suggest these testsare specific to the hypothesized spinal segments ordysfunctions.

Our results for the PIM of the upper neck indicatethat the hypomobility in the occipital-atlanto joint canbe consistently reproduced and assessed during lateralglides of the occiput on the atlas (Kappa of .81,agreement of 93%) and during lateral bend of the headwith a fixed axis (Kappa of .35, agreement of 90%). Thislater measure of PIM showed a high percentage ofagreement between examiners but low Kappa coeffi-cients. This was probably due to low prevalence ofpositive findings (10%) (Table 3). The low prevalence ofpositive findings of the lateral displacement of axis likelyresulted in an artificial deflation of the Kappa statistic.The Kappa coefficient is influenced by the prevalence ofthe attribute (e.g. a disease or clinical sign). If theprevalence is high, chance agreement is also high andkappa is reduced accordingly (Sim and Wright, 2005).To overcome this problem and have a better distributionof positive findings, future studies should investigateupper cervical PIM using only patients with apparentupper neck complaints. In addition, in this study thepatient’s pain response may have also influenced theconsistency of measures of mobility during PIM of theoccipital-atlanto joints.

Tenderness over the transverse processes of atlas wasreliable. Because the transverse processes of atlas aregenerally tender, we clarify that we considered positive

for tenderness when the palpation reproduced the samesymptom that the patient was experiencing. Therefore,this test may be helpful to identify the source ofsymptoms. Regarding the atlanto-axial tests of mobility,low reliability was found in all techniques used: rotationduring full neck flexion or rotation during full necklateral flexion. The low reliability cannot be explainedby the prevalence of positive findings, since around 60%of the patients had positive tests. The presence of painduring atlanto-axial joint test using the technique ofrotation with the neck in full lateral flexion had amoderate reliability, with a fair lower bound of the 95%CI (.5). We believe this moderate reliability is due to theconsistency of pain reproduction during the full passivelateral flexion rather than the rotation component addedat the end of the movement.

Regarding the reliability of PIM for mobility and painin the mid and lower neck, we observed that judgmentsof hypomobility were not consistent. Since we usedprocedures to minimize the possibility that the PIM bythe first examiner would cause a true change in thepatient’s restriction in motion to the next examiner, webelieve that consistency of mobility findings during thesetests may not be attainable. Although true change insegmental restriction may play a role, it is a factor noteasy to control. In addition, the agreement between thetesters was relatively low, which supports the low Kappavalues. On the other hand, symptom reproduction hasshown fair to moderate reliability in the lower segmentsof C4–C6. Therefore, it may be that during the PIMtests, clinicians should somewhat rely on symptomreproduction of the lower neck when making treatmentdecisions. While we cannot do direct comparison of thereliability results for PIM between our results and otherstudies because the techniques used for testing weredifferent, in general, we have found better reliability forPIM than prior studies (Fjellner et al., 1999; Smedmarket al., 2000; Pool et al., 2004).

Some may argue that for the mobility classification weonly used the categories hypomobility or normal. Wedid not classify mobility as hypermobility because ineveryday practice when we perform PIM of the neck welook predominantly for hypomobility of one segmentrelative to the other segments. In this study, hypomo-bility was defined as when decreased mobility was notedto one side compared to the opposite side. Therefore, wedo not discard the possibility that in some segmentswhich were classified as having normal mobility mayhave shown some increased mobility.

We caution the reader that some of these tests with asomewhat lower reliability coefficient may demonstrateuseful validity in future investigations. To date, there isno accepted reference standard of cervical dysfunctionwith which to validate these tests and measures as usefuldiagnostic tools. However, instead of investigating thepotential diagnostic accuracy of these tests, future

ARTICLE IN PRESSS.R. Piva et al. / Manual Therapy 11 (2006) 321–330 329

studies could establish the criterion validity of these teststo guide the selection of interventions and establishprognosis in patients with neck pain.

5. Conclusions

Cervical spine AROM measurement demonstratedmoderate to substantial inter-tester reliability andresulted in a MDC adequate for clinical use. The effectof AROM on symptom provocation also resulted inmoderate to substantial reliability for tests of symptomreproduction for cervical flexion and rotation. Measuresof PIM resulted in substantial and moderate reliabilityof assessing occipital-atlas mobility, tenderness of thetransverse processes of atlas, and symptom provocationduring PIM testing of the lower cervical segments. Lowprevalence of positive findings likely resulted in anartificial deflation of the Kappa statistic during somePIM measures. Measures of AROM in the saggital andtransverse planes were significantly associated withdisability scores. Findings are relevant to plan futurestudies to establish the criterion validity of these tests toguide the selection of interventions and establishprognosis in patients with neck pain.

Disclaimers

The opinions or assertions contained herein are theprivate views of the authors and are not to be construedas official or as reflecting the views of the US Air Forceor Department of Defense.

References

Bogduk N, Mercer S. Biomechanics of the cervical spine. I: normal

kinematics. Clinical Biomechanics 2000;15(9):633–48.

Capuano-Pucci D, Rheault W, Aukai J, Bracke M, Day R, Pastrick

M. Intratester and intertester reliability of the cervical range of

motion device. Archives of Physical Medicine And Rehabilitation

1991;72(5):338–40.

Childs JD, Piva SR, Whitman JM. Lower cervical spine. In: Physical

Therapy for the cervical spine and temporomandibular joint. Home

study course 13.3, Orthopaedic Section of the American Physical

Therapy Association, La Crosse, WI. 2003.

Cohen J. A coefficient of agreement for nominal scales. Educational

and Psychological Measurement 1960;20:37–46.

Cohen J. Statistical power analysis for the behavioral sciences.

Hillsdale: Lawrence Erlbaum Associates, Inc.; 1988.

Erhard RE. Manual therapy in the cervical Spineorthopedic. Physical

therapy home study course. American Physical Therapy Associa-

tion, Orthopaedic Section. 1996.

Fjellner A, Bexander C, Faleij R, Strender LE. Interexaminer

reliability in physical examination of the cervical spine. Journal

of Manipulative and Physiological Therapeutics 1999;22(8):511–6.

Guyatt GH, Kirshner B, Jaeschke R. Measuring health status: what

are the necessary measurement properties? Journal of Clinical

Epidemiology 1992;45(12):1341–5.

Hains F, Waalen J, Mior S. Psychometric properties of the neck

disability index. Journal of Manipulative and Physiological

Therapeutics 1998;21(2):75–80.

Hertling D, Kessler RM. Management of common musculoskeletal

disorders, 3 ed. Philadelphia: Lippincott; 1996 Chapter 17, p. 540.

Hole DE, Cook JM, Bolton JE. Reliability and concurrent validity of

two instruments for measuring cervical range of motion: effects of

age and gender. Manual Therapy 2000;1(1):36–42.

Jordan K. Assessment of published reliability studies for cervical spine

range-of-motion measurement tools. Journal of Manipulative and

Physiological Therapeutics 2000;23(3):180–95.

Kadir N, Grayson MF, Goldberg AAJ, Swain MC. A new neck

goniometer. Rheumatology and Rehabilitation 1981;20:219–26.

Magee DJ. Orthopedic physical assessment. Philadelphia: Saunders

Co; 1997 Chapter 1, p. 9 & Chapter 3, p. 137.

Maitland GD. Vertebral manipulation, 5th ed. Sydney, New South

Wales, Australia: Butterworths; 1986.

Mannion AF, Klein GN, Dvorak J, Lanz C. Range of global motion

of the cervical spine: intraindividual reliability and the influence of

measurement device. European Spine Journal 2000;9(5):379–85.

McKenzie RA. Cervical and thoracic spine: mechanical diagnosis and

therapy. Minneapolis: Orthopaedic Physical Therapy Products;

1990.

Nilsson N. Measuring passive cervical motion: a study of reliability.

Journal of Manipulative and Physiological Therapeutics 1995;

18:293–7.

O’Driscoll SL, Tomenson J. The cervical spine. Clinics in Rheumatic

Diseases 1982;8:617–30.

Petersen CM, Johnson RD, Schuit D. Reliability of cervical range of

motion using the OSI CA 6000 spine motion analyser on

asymptomatic and symptomatic subjects. Manual Therapy 2000;

5(2):82–8.

Pool JJ, Hoving JL, de Vet HC, van Mameren H, Bouter LM. The

interexaminer reproducibility of physical examination of the

cervical spine. Journal of Manipulative and Physiological Ther-

apeutics 2004;27:84–90.

Portney LG, Watkins MP. Foundations of clinical research: applica-

tions to practice. Stamford: Appleton & Lange; 1993 [Chapter 5,

p. 58, and Chapter 26, p. 523–24].

Rheault W, Albright B, Byers C, Franta M, Johnson A, Skowronek

M. Intertester reliability of the cervical range of motion device.

Journal of Orthopaedic & Sports PhysicalTherapy 1992;15:147–50.

Shrout PE. Measurement reliability and agreement in psychiatry.

Statistical Methods in Medical Research 1998;7(3):301–17.

Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater

reliability. Psychological Bulletin 1979;86:420–8.

Sim J, Wright CC. The kappa statistic in reliability studies: use,

Interpretation, and sample size requirements. Physical Therapy

2005;85:257–68.

Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence:

sample size estimation for diagnostic test studies. Journal of

Clinical Epidemiology 1991;44(8):763–70.

Smedmark V, Wallin M, Arvidsson I. Inter-examiner reliability in

assessing passive intervertebral motion of the cervical spine.

Manual Therapy 2000;5(2):97–101.

Solinger AB, Chen J, Lantz CA. Standardized initial head position in

cervical range-of-motion assessment: reliability and error analysis.

Journal of Manipulative and Physiological Therapeutics 2000;

23(1):20–6.

SPSS for Windows 10.0, 1998. SPSS Inc, Chicago.

Stratford PW, Goldsmith CH. Use of standard error as a reliability

index of interest: an applied example using elbow flexor strength

data. Physical Therapy 1997;77(7):745–50.

Stratford PW, Riddle DL, Binkley FM, Spadoni G, Westaway MD,

Padfield B. Using the neck disability Index to make decisions

concerning individual patients. Physiotherapy Canada 1999;51:107–12.

Tucci SM, Hicks JE, Gross EG, Campbell W, Danoff J. Cervical

motion assessment: a new, simple and accurate method. Archives

of Physical Medicine and Rehabilitation 1986;67:225–30.

ARTICLE IN PRESSS.R. Piva et al. / Manual Therapy 11 (2006) 321–330330

Vernon H, Mior S. The neck disability index: a study of reliability and

validity. Journal of Manipulative and Physiological Therapeutics

1991;14:409–15.

Viikari-Juntura E. Interexaminer reliability of observations in physical

examinations of the neck. Physical Therapy 1987;67:1526–32.

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(1):52–62.

Werneke M, Hart DL. Centralization phenomenon as a prognostic factor

for chronic low back pain and disability. Spine 2001;26(7):758–64.

Youdas JW, Carey JR, Garrett TR. Reliability of measurements of

cervical spine range of motion—comparison of three methods.

Physical Therapy 1991;71(2):98–104.

Further reading

Christensen HW, Nilsson N. The reliability of measuring active and

passive cervical range of motion: an observer-blinded and

randomized repeated-measures design. Journal of Manipulative

and Physiological Therapeutics 1998;21(5):341–7.

Jette DU, Jette AM. Physical therapy and health outcomes in

patients with spinal impairments. Physical Therapy 1996;76(9):

930–41.

Jordan K, Dziedzic K, Jones PW, Ong BN, Dawes PT. The reli-

ability of the three-dimensional FASTRAK measurement

system in measuring cervical spine and shoulder range of

motion in healthy subjects. Rheumatology (Oxford) 2000;39(4):

382–8.