reliability of measures-pfps.pdf

BMC Musculoskeletal Disorders {)SiroMed Central

Research article

Reliabil ity of measures of impairments associated withpatellofemoral pain syndromeSara R Piva* 1, Kelley Fitzgeraldl, ]ames J lrrgangl, Scott Jones2,Benf amin R Hando3, David A Browdera and John D Childss

Adclress: lDepartment of Physical Therapy, School of I Ieal th and Rehabi l i tat ion Sciences, Universi ty of Pi t tsburgh, [JSA, 2Ramstein Outpat ientPhysical Medicine CI in ic, Ramstein Air Force Base, Cermany, r l ,g l in Regional tJospi ta l , Egl in Air Force Base, FL, t lSA, aWil ford Hal l Medical Center,Lackland Air Force Base, fi, llSA and 5LIS Army-Baylor University Doctoral Program in Physical -l'herapy, tlSA

Emai l : Sara R Piva. - [email protected] t t .edu; Kel ley Fi tzgerald - kf izger@pit t .edu; James I I r rgang - i r rgangj [email protected];Scott Jones - scott . jonesl @ramstein.af .mi l ; Benjamin R Hando - [email protected] .mi l ;David A Rrowder - [email protected] . rn i l ; lohn D Chi lds - chi ldsjd@sbcglobal .net* Corresponding author

Publ ished: 3 | March 2006 Received: 2l January 2006

BMCMuscutosketerol Disorders2005,7:33 doi :10. 1185/1471 -2474-7-33 Accepted: 3lMarch2006

This art ic le is avai lable f rom: ht to: / /ww.biomedcentral .coml l4Tl-747417133

@ 2005 Piva et a l ; l icensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.orgllicenses/b),/2.O),which permits unrestr icted use, distr ibut ion, and reproduct ion in any medium, provided the or ig inal work is proper ly c i ted-

Abstract

Background: The rel iabi l i ty and measurement error of several impairment measures used duringthe cl inical examination of patients with patel lofemoral pain syndrome (PFPS) has not beenestabl ished. The purpose was to determine the inter-tester rel iabi l i ty and measurement error ofmeasures of imoairments associated with PFPS in Datients with PFPS.

Methods: A single group repeated measures design was used. Two pairs of physical therapistspart icipated in data col lect ion. Examiners were bl inded to each others' measurements.

Resufts: Thirty patients (age 29 +l- 8; 17 female) with PFPS part icipated in this study. Inter-testerrel iabi l i ty coeff icients were substantial for measures of hamstrings, quadriceps, plantarf lexors, andITB/TFL complex length, hip abductors strength, and foot pronation ( lCCs from .85 to .97);moderate for measures of Q-angle, t ibial torsion, hip external rotat ion strength, lateral ret inaculart ightness, and quali ty of movement during a step down task ( lCCs lrom .67 to .79); and poor forfemoral anteversion ( lCC of .45). Standard error of measurement (SEM) for measures of musclelength ranged from 1.6 degrees to 4.3 degrees. SEM for Q-angle, t ibial torsion, and femoralanteversion were 2.4 degrees, 2.9 degrees, and 4.5 degrees respectively. SEM for foot pronationwas I mm. SEM for measures of muscle strength was | .8 Kg for abduction and 2.4 Kg lor externalrotat ion.

Conclusion: Several of the impairments associated with PFPS had suff icient rel iabi l i ty and lowmeasurement error. Further investigation is needed to test i f these impairment measurements arerelated to physical function and whether or not they are useful for decision-making.

BackgroundPatel lofemoral pain syndrome (PFPS) is a common kneeproblem among young active individuals [1-31. The

mechanism of PFPS is not well understood. I t has beenproposed that PFPS may arise from abnormal muscularand biomechanical factors that alter tracking of the patel la

of 13(page number not for citation purposes)

BMC Musculoskeletal Disorders 2006, 7:33

within the femoral trochlear notch contributing toincreased patellofemoral contact pressures that result inpain and dysfunction [4,5]. Authors have suggested a vari-ety of impairments involved in the et iology of PFPS [6-8].However, there is no evidence that these impairments areassociated with the patient 's functional l imitat ions. In theabsence of definitive impairments in which to focus theexamination or treatment in patients with PFPS, cl inicianstend to perform an extensive physical examination thatgeneral ly includes a mult i tude of impairment measuressuch as muscle weakness, soft t issue t ightness, structuralor postural alterat ions, and poor qual i ty of movement [31.

Reliabi l i ty and rneasurement error are essential propert iesof any measurement that need to be establ ished before themeasurement can be considered cl inical ly meaningfuland useful. Rel iabi l i ty is the abi l i ty of a test to consistentlyyield more or less the same results when administered onseveral occasions to stable subjects, whereas measurementerror provides the threshold for interpreting test resultsbeing reasonably confident that true change has occurred

[9,10]. Although several studies have investigated the rel i-abi l i ty of impairment rneasLlres associated with patel-lofernoral dysfunct ion in heal thy subjects [11'151, therel iabi l i ty and measurement error of impairment rneas-ures used during the cl inical examination of patients withPIrPS has not been establ ished.

Arnclng the measures of muscle strength performed inpatients with PFPS, rel iabi l i ty of hip abduction and hipexternal rotat ion strength tests have not been determinedin pat ients wi th PFPS. I l ip abductor and external rotat ionstrength are commonly rneasured in patients with PFPSbecause weakness of these muscles has been l inked withPI]PS [16, 17]. Authors have suggested these muscles helpto maintain pelv ic stabi l i ty by eccentr ical ly control l ingfemoral internal rotat ion during weight-bearing activi t ies.Weakness may result in increased medial femoral rotat ionand valgus knee moments, augment ing compressiveforces on the patel lofemoral jo int [16,171. I re land et a l

I lB] suggest that individuals with PFPS have weaker hipmuscles when compared to matched control groups.Another study has shown that hip abduction strength isone of the variables able to dist inguish between patientswith and without PFPS [191.

Soft t issue restr ict ions, such as shortening of the quadri-

ceps, hamstrings, and plantarf lexor muscles, shortening ofthe i l iot ibial band/tensor fascia lata (ITB/TFL) complex,and shortening ofthe lateral retinacular structures have allbeen associated with PFPS and are impairments com-monly measured in this populat ion l2O-221. I t is theo-rized that t ight quadriceps and hamstrings may increasecompression of the patel lofemoral ioint [20]. While twostudies agree support ing the associat ion of quadriceps

htto://www. biomedcentral. coml I 47 1 -247 417 133

flexibility and PFPS, the same studies conflict regardingthe association of hamstrings flexibility and PFPS 121,22]1.There is some evidence to support the associationbetween plantarf lexors t ightness and PFPS [21]. Concern-ing the ITB/TFL and lateral ret inacular t issues, although i thas been theorized that tightness of these tissues may dis-place the patel la lateral ly and increase the stress in thepatel lofemoral ioint or medial ret inacular t issue [1,231,evidence to support such theory does notyet exist. In gen-eral, studies investigating the measurement propert ies ofthe above mentioned soft t issue measures have not usedindividuals with PFPS, or have not determined the meas,urement error [1 1,12,24-301.

Studies examining the measurement propert ies clf testsused to determine structural or postural alterat ions inpatients with PI]PS are also lacking. Some structural orpostural alterat ions that have been l inked to PFPS areexcessive foot pronation, quadriceps angle (Q angle), t ib-ial torsion, and fernoral anteversion. Irvidence to supportthat increased foot pronation causes PIrPS is inconclusive

[6,31]. Regarding Q-angle, i twas reported that Q-angle ismore accentuated in runners with PFPS than rn runnerswithout PFPS l7l. ' l 'o our knowledge, just one study hasinvestigated the relat ionship between t ibial torsion andPIrl)S and reported that the lateral rotat ion of the t ibia rel,at ive to the femur was increased in patients with PI]PS

1321. Studies that investigated the associat ion of femoralanteversion and PIrPS have reported confl ict ing results

132,33]1. Although some measures of structural alterat ionshave shown good rel iabi l i ty lal , samples of patients withPIrPS have rarely been used I t 3- t 5 l . A recent study usingpatients with PFPS reported poor consistency of thesemeasurements [301.

Quali ty of movement, sometimes referred to as neuromo-tor control or movement coordination, refers to the bio-mechanics of the lower extremit ies, trunk and arms inrelat ionship with i ts surrounding during physical act ivi-t ies [4]. I t has been theorized that patients with PFPSexhibit altered movernent patterns in the lower extremi-t ies that may result in alterat ions of the load distr ibutionacross the patel lofemoral joint 11,21,341. Altered move-ment patterns may be recognized during physical act ivi-t ies as movements performed with poor qual i ty. We areunaware of studies that investigated the consistency ofmeasures of qual i ty of movement in patients with PFPS.

The purpose of this study was to determine the inter-testerrel iabi l i ty and measurement error of measures of irnpair-ments associated with PFPS in a populat ion of patientsdiagnosed with PFPS. We have selected to examine themeasurement properties of measures of hip abductionstrength, hip external rotation strength, quadricepslength, hamstrings length, plantar flexors length, ITB/TFL


BMC Musculoskeletal Disorders 2006. 7:33

complex length, lateral retinacular structures length, footpronation, Q-angle, t ibial torsion, femoral anteversionand quality of movement, because of their frequent use inthe examination of individuals with PFPS and the lack of

:13:-"t"" concerning their rel iabi l i ty and measurement

MethodsA single group repeated measures design was used in thisstudy. Data for this study was obtained as part of a largermult icenter study that investigated predictors of functionin persons with PF'PS.

SubjectsI l .rdividuals were el igible to part icipate in t l-r is study i f theywere diagnosed by a physician with PITPS, were between12 ancl 50 years ofage, had pain in one orboth knees, hadduration of signs and symptoms greater than 4 weeks, hadhistory of insidious onset not related to trauma, and hadpain in the patel lar region with at least three of the fol low-ing: manual compression of the patel la against the femurat rest or during an isometric knee extensor contraction,palpation of the postero-medial and postero-lateral bor-ders of the patel la, resisted isometric quadriceps femorismuscle contraction, squatt ing, stair cl imbing, kneeling, orprolonged sit t ing.

I lxclr-rsion cri teria included previous patel lar dislocation,knee surgery over the past 2 years, concomitant known orsuspected diagnosis of: peripatel lar bursit is or tendonit is,internal knee derangement, systemic arthri t is, l igamen-tous knee injury or laxity, pl ica syndrome, Sinding-l .arsen-

Johansson's disease, Osgood Schlatter 's disease, infect ion,malignancy, musculoskeletal or neurological lowerextrernity involvement that interferes with physical act iv-ity, and pregnancy. 'l'hirty patients were recruited from 2cl in ical s i tes: Wi l ford I Ia l l Medical Center, in San Anto-nio, ' lX, and University of Pit tsburgh's Centers for RehabServices, Pit tsburgh, PA). Al l subjects who agreed to par-t icipate signed a consent forrn approved by the Inst i tu-t ional Review l loard of the respective cl inical si te.Demographic characterist ics of the part icipants arereported in' l 'able l .

ProceduresSubjects had one lower extrernity tested unless they hadbilateral symptoms, in which case the most symptomaticside was tested. The most symptomatic knee was deter-mined by the patient 's self-report. Data were col lectedduring one assessment session that lasted approximately60 minutes. We col lected data during the same assessmentsession to ensure the subf ects remained as stable as possi-ble (did not change) in the parameters tested. Examinersmet once during a 2-hour session before the study was ini-tiated to review operational definitions and practice the

http ://www. biomed central. co m | 1 47 1 -247 4 17 I 33

Table l : Demographic character ist ics of the sample. Valuesrepresent the mean (Standard Deviation) unless otherwisestated.

Var iable (n= 30)Age in yearsNumber of subjects in each age range:From l4to l9yearsFrom 20 to 29 yearsFrom 30 to 39 yearsFrom 40 to 47 yearsNumber of femalesHeight in cmWeithr in ktBody Mass Index as kg/cm2Numeric Pain Rat ing Scalea scoreAct iv i ty of Dai ly Liv ing Scale*+ score

2e. | (8.4)

2 (7%)r6 (s3%)7 (23%)s (17%)t7 (se %)r7r ( i l . r )7e ( r8.6).26 (.0s)3.e (r .e)67.3 ( | 7.3)

* Numeric pain rating scale ranges from 0 (No Pain) to l0 (Worstlmaginable Pain) points.8x Activity of daily living scale ranges from 0 to 100 points. Onehundred indicates absence of symptoms and functional l imitat ions.

procedures to ensure standardization. l tach examiner wasprovided with the Manual of Standard Operating Proce-dures oIthe study, which contained detai led explanationsabout the performance of each test.

' Iwo pairs of physical therapists (1 pair from each site)with dif ferent levels of experience part icipated in data col-lect ion. One pair of testers had 3 and 5 years of cl inicalpractice (pair 1); whereas the other pair had 2 and 10years of cl inical experience (pair 2). During each data col-lect ion session, the subject remained inside an examination room. ' Io ensure that the examiners remainedblinded to each other's assessments, the two examinersentered the examination room independently, performedand recorded the measurements, and then left the room.The results were not shared with the other examiner. ' l 'he

ffreasurements were always performed in the same order.Order of test ing was based on patient posit ioning in thefol lowing order: supine, prone, side-lying, and standingposit ions. This was done to avoid excessive changing o[posit ions, ensure that the examiners were performing al ltests under the same condit ions, and ensure that any ordereffect would be the same for each examiner. Each exam-iner in the pair alternated serving as the init ial exarniner.

MeosuresEach part icipant completed a demographic questionnaireand self-reported measures of pain and function prior tothe physical examination. Subjects' age, gender, height,weight, prior history of knee problems, mechanism ofinjury, current episode duration, and symptom locationwere recorded.

Pain intensity was measured using an 11-point numericpain rat ing scale ranging from 0 (No Pain) to 10 (Worst



Imaginable Pain). Patients rated their current, best, andworst level of pain during the last 24 hours. The average ofthe three ratings was used to represent the patient's overallpain intensity. Numeric pain scales have been shown tobe rel iable and val id [35-38].

The Activi ty of Daily Living Scale (ADLS) of the Knee Out-come Survey was used as a knee-specific measure of phys-ical function [39]. The ADLS assesses the effects of kneeimpairment on activi t ies of dai ly l iv ing. The ADLS consistsof l4 i tems that measure the ful l spectrum of symptomsand functional l imitat ions during activi t ies of dai ly l iv ingthat one may experience as a result of a variety of kneepathologies. l 'he ADLS score is transformed to a 0 to 100point scale with 100 indicating the absence of symptomsand functional l i rnitat ions. Psychornetr ic test ing has dem-onstrated the ADLS to be rel iable, val id and responsive insubjects wi th PFPS [39,401.

Measurements performed during the physical examina-t ion were as fol lows:

Hamstrings length was determined by measuring thestraight leg raise using a gravity goniometer (MIl, MedicalResearch Ltd., Leeks, LIK). The subject was in the supineposit ion with the tested knee extended and the other legflat on the table to avoid excessive posterior pelvic t i l t .I lefore start ing the measurement, the goniorneter waszeroed on the lower half of the anterior border of the t ibia.' l 'hen, the lower extremity was passively l i f ted to the endrange of motion or f i rm end feel and the measurementrecorded in degrees (Figure I ) .

' l 'he average measurementof two tr ials with 5-second pause between tr ials wasrecorded.

Tightness of the lateral retinacular structures wasassessed with the patel lar t i l t test Iaal. The patel lar t i l t testwas performed with the subject in supine with the knee inful l extension and the femoral condyles placed in the hor-izontal plane. ' l 'he exarniner attempted to l i f t the lateraledge of the patel la from the lateral femoral condyle. Thepatel la was not al lowed to move lateral ly during the meas-urement ( l ; igure I ) . The inabi l i ty to l i f t the lateral boarderof the patel la above the horizontal plane indicated a pos-i t ive test for t ightness of the lateral ret inaculum. Adequatelength of the lateral retinaculum or negative test was indi-cated by the ability to lift the lateral boarder of the patellaabove the horizontal plane. Tightness of lateral ret inacu-lum was scored as t ight or normal.

Q-Angle was measured with the knee in full extensionwith the subject supine. The angle formed by the intersec-t ion of the l ine of appl icat ion of the quadriceps force ( l inefrorn the anterior superior iliac spine to the center ofpatel la) with the center l ine of the patel lar tendon ( l ine

http://www. biomedcentral.coml 1 47 I -247 417 133

from the center of the patella to the tibial tubercle) wasmeasured in degrees with a universal goniometer (Figure1) 142]1. The center of the patella and the tibial tuberclewere marked with a demographic penci l , which waswiped out after the measurement. Before the measure-ment the tester palpated the anterior superior i l iac spineand asked the subject to keep his second f inger point ingdown over this landmark during the measurement. Sub-

iect was also asked not to contract the quadriceps musclesduring the measurement.

Tibial torsion was measured with a universal goniometerwith the part icipant prone on a low table, and with thetested knee bent at 90'. I leight of the table was adjustedso the tester could comfortably visual ize the plantar sur-face of the subject 's foot. To faci l i tate visual izat ion, thetester marked the most prominent aspect of the medialand lateral malleolus with a srnal l dot. ' l 'he exanrrnermeasured the angle formed by the axis of the knee ( imag-inary l ine frorn the medial to lateral femoral epicondile)and an irnaginary l ine through the malleol i ( l i igure I ) . Weelected to measure t ibial torsion with the patient in aprone posit ion rather than the posit ion usually describedwith the patient si t t ing with knees in 90" because t ibialtorsion is a horizontal plane rotat ional malal ignment

143,441. We bel ieve using an inferior view o[ the leg ena-bles better observation of the talocrural ioint axis in thehorizontal plane.

Quadriceps length was determined by measuring thequadriceps femoris muscle angle during passive knee f lex-ion with the subiect in the prone posit ion. Care was takento avoid anterior t i l t ing of the pelvis and/or extension ofthe lumbar spine. The angle of knee f lexion in the proneposi t ion was measured using a gravi ty goniometer whichwas zeroed on a horizontal surface prior to the measure-ments. The gravity goniometer was placed over the distalt ibia (Figure 1). 1'he average measurement of two tr ialswith 5-second pause between tr ials was recorded.

Femoral anteversion was measured using the Craig's testwith the part icipant in the prone posit ion with the kneeflexed to 90' [451. Before start ing the measurement, thegravity goniometer was zeroed on a vertical surface andplaced on the medial surface of the lower leg, just proxi-mal to the medial malleolus ( ' I 'able 2). The examiner pal-pated the posterior aspect of the greater trochanter of thefemur. The hip was then passively rotated unti l the mostprominent port ion of the greater trochanter reached thehorizontal plane. The degree of anteversion was then esti-mated based on the angle of the lower leg with the vertical(Figure 1).

Plantar flexors length was determined by measuring theamount of ankle ioint dorsiflexion with the knee extended


BMC Musculoskeletal Disorders 2006, 7:33 http ://www. biom ed central. co ml 1 47 1 -247 4 17 I 33

l l luslrttions of ur*ururcs of irnprirmcnls

Figure Il l lustrat ion of the techniques used to measure impairments associated with PFPS(A) Hamstrings length - straight leg raise test(B) Tightness of the lateral ret inacular structures - patel lar t i l t test(C) Q-angle(D) Tibial torsion - angle formed between inter-epicondilar and intermalleolar l ines(E) Quadriceps muscle length - quadriceps femoris muscle angle(F) Femoral anteversion - Craig's test(G) Plantar flexors length(H) Hip external rotation strength(l) Hip abduction strength

0) ITB/TFL complex length- Ober's test(K) Foot pronation - navicular drop test(L) Quality of movement - example of lateral step down test trial using arm strategy

of 13(page number not for citaIon purposes)


and again with the knee flexed at 90'. Ankle dorsiflexionmeasured with the knee extended was used to account forthe influence of gastrocnemius tightness. Measurement ofankle dorsiflexion with the knee bent was used to detectt ightness of loint capsule or soleus muscle. The subf ectwas posit ioned in the prone posit ion with the foot hang-ing offthe table and the subtalar joint was maintained inthe neutral posit ion. Dorsif lexion was measured with astandard goniometer as the angle formed by the lateralmidl ine of the leg on a l ine from the head of the f ibula tothe t ip of the lateral malleolus and the lateral midl ine ofthe foot in l ine with the border of the rearfoot/calcaneus(Figure I ) . The average measurement of two tr ials with 5-second pause between tr ials was recorded.

Hip external rotation strength - Strength measures wereperformed using the Lafayette Manual Muscle Test (MMT)System (Lafayette Instrument, l .afayette, IN). Musclestrength was recorded in terms o[force, in ki lograms. Hipexternal rotat ion strength was examined with the subjectposit ioned in prone on a padded table with the test kneeflexed to 90' and the hip in neutral rotat ion. ' fhe contral-ateral lower extremity was posit ioned with the hip in neu-tral rotat ion and the knee in ful l extension. ' l 'o obtainoptirnal rnechanical advantage, the examiner stood on theside of the table opposite of the test l imb. Subjects exertedan isometric contraction of their hip external rotators for3-5 seconds in a posit ion of neutral hip rotat ion. Themanual resistance against the external rotat ion wasappliecl with the MMT just proximal to the medial malle-olus (Figure 1) To maintain uni formity in the nature ofverbal commands provided by the tester during test ing,the testers were instructed to always give a strong verbalencouragement during the perfonr-rance of every maxi-mum effort. 'l'he average force of two trials with onerninute of rest between tr ials was recorded.

Hip abduction strength was measured with the subject inside-lying with the test hip posit ioned superior withrespect to the contralateral hip. Subjects exerted an iso-metric contraction of their hip abductors for 3-5 secondsin a posi t ion of approximately 30'of h ip abduct ion and5" of hip extension. ' l 'he manual resistance was appliedwith the MMT just proximal to the lateral malleolus in thedirect ion of adduction (Figure 1). To maintain uniformityin the nature of verbal commands provided by the testerduring testing, the testers were instructed to always give astrong verbal encouragement during the performance ofevery maximum effort. The average force of two trials withone minute of rest between trials was recorded.

Length of the Iliotibial Band/Tensor Fascia Lata (lTB/TFL) Complex was examined using the Ober's test [461.The subject was posit ioned in side-lying with the testedleg positioned superior and the lower leg slightly flexed at

http://www. biomedcentral.coml 1 47 I -247 417 133

the hip and knee to maintain stability. The gravity goni-ometer was zeroed on a horizontal surface prior to themeasurement and was placed over the distal portion ofthe ITB/TFL complex (Figure 1 ). The test leg was flexed toa r ight angle at the knee and grasped just below the kneewith the examiner's distal hand. The examiner moved thesubiect's thigh first in flexion, then through abductioncombined with extension unti l the hip was posit ioned inmid-range abduction with neutral f lexion/extension.From this position the thigh was allowed to drop towardthe table unti l the point where the l imb stopped movingtoward the table. At that point the measurement wastaken. The result was recorded as a continuous variable.Negative values represent more t ightness whereas posit ivevalues (below horizontal) represent less t ightness. ' fhe

average measurement of two tr ials with 5-second pausebetween tr ials was recorded.

Foot pronation was measured by the navicular drop test

114,47];. Navicular drop test measures the dif ferencebetween height of the navicular at subtalar joint neutralposit ion and that of the relaxed stance posit ion 114,47|1.The sublect stood on a high hard surface with his feetshoulder width apart. ' fhe examiner stayed behind thesubject with the eyes leveled at subject 's feet. The examinermarked the subject 's navicular tuberosity with a demo-graphic penci l , which was wiped out after the measure-ment. ' l 'he examiner put the subject in the subtalar iointneutral posit ion. [ Ising an index card placed perpendicu-lar to the table, the examiner recorded the distance fromthe navicular to the f loor (Figure 1). The subject was theninstructed to relax from the subtalar neutral posit ion andthe measurement was repeated. ' l 'hen, with a metric ruler,the distance between the two dots, in the index card(which represents the dif ference in the posit ion of thenavicular tubercle with respect to the f loor between thesubtalar neutral and relaxed standing posit ions) wasrecorded in mil l imeters. Creater distances between thedots indicate greater pronation.

Quality of movement during the lateral step down testwas assessed using a scale designed for this purpose. ' I 'hesubject was asked to stand in single l i rnb support with thehands on the waist, the knee straight and the foot posi-tioned close to the edge of a 20 cm high step. 'I'he contral-ateral leg was posit ioned over the f loor adjacent to thestep and was maintained with the knee in extension. Thesubject then bent the tested knee unti l the contralateral leggently contacted the floor and then re-extended the kneeto the start position. This maneuver was repeated for 5repetitions. The examiner faced the subject and scored thetest based on 5 cri teria: 1) Arm strategy. I fsubject used anarm strategy in an attempt to recover balance, 1 point wasadded (Figure l) ;2) Trunk movement. I f the trunk leanedto any side, I point was added; 3) Pelvis plane. I f pelvis



rotated or elevated one side compared with the other, Ipoint was added; 4) Knee posit ion. I f the knee deviatedmedially and the tibial tuberosity crossed an imaginary

vertical line over the 2nd toe, add I point, or, if the knee

deviated medial ly and the t ibial tuberosity crossed an

imaginary vertical line over the medial border of the foot,

add 2 points, and; 5) Maintain steady uni lateral stance. I f

the subf ect stepped down on the non-tested side, or i f the

subject tested l imb became unsteady ( i .e. wavered from

side to side on the tested side), add 1 point. Total score of

0 or I was classif ied as good quali ty of movement, total

score of 2 or 3 was classif ied as medium quali ty, and total

score of 4 or above was classif ied as poor qual i ty of move-

ment.

Doto onolysisDescript ive stat ist ics, including frequency counts for cate-

gorical variables and measures of central tendenry and

dispersion for continuous variables were calculated to

summarize the data. Kolmogorov-Smirnov Z-tests were

perforrned to assess whether continuous data approxi-

mated a norrnal distr ibution. Inter-tester rel iabi l i ty for cat-

egorical or ordinal impair lnent measurelnents was

deterrnined by a Cohen's Kappa stat ist ics and i ts 95% Cl

[481. lntraclass correlat ion coeff icients ( lCC) and their

95o/o Cl were calculated for continuous measures [49,501.' l 'he ICC model (2, 1) was used when the unit of analysis

was a single lneasurement, and the model (2, 2) was used

when the r-rnit of analysis represented the mean of 2 rat-

ings [49,501. ' fhe mean square estimates to calculate the

ICC coeff icients were obtained from a random effects 2-

way analysis of variance with repeated measures [50].

htto://www. biomedcentral.coml 1 47 1 -247 417 133

Calculat ion of the standard error of measurement (SEM)

was used to determine measurement error. Results of thereliability analyses for the continuous measures were usedto calculate the SEM. The SEM was calculated as (SD * v 1

- r), where r is the test-retest reliability coefficient and SD

is the standard deviat ion of the combined scores 151,52]1.

The sample size was calculated a priori using Sample-

Power'" (Chicago, I l l inois) stat ist ical software based on

the calculat ion of Cohen Kappa coeff icients on a dichoto-mous variable ( i .e. t ight or not t ight during the patel lar t i l t

test). To ensure sufficient statistical power to achieve a

lower bound ofthe 95%o confidence interval for Kappa of

0.30, assuming Kappa would be equal to 0.60, a sample

size of 30 subjects was needed [aBl. This sample size

would also be adequate to calculate ICC coeff icients on

the continuous variables, given that we had 2 testers per

subject, hoping for an ICC of .85, and having determined

that rel iabi l i ty of .60 or higher would be acceptable [53].

ResultsAll the continuous variables were found to approximate a

normal distr ibution (Kolrnogorov Smirnov Z tests p >

.10). Results of the rel iabi l i ty analysis are in Table 2. I 'able

2 shows the means and standard deviat ions of the 4 testers

on the continuous variables, the percentage of f indings

and percentage of agreement for categorical or ordinal

variables, the rel iabi l i ty model used during the analysis,

the rel iabi l i ty coeff icient with the 95o/o Cl, and the stand-

ard error of measurement for continuous variables. ' I 'able

3 shows the rel iabi l i ty coeff icient values for the overal l

sample and for each of the two pairs of testers.

Table 2: Means and standard deviations ofthe 4 testers ofthe continuous variables, percentage offindings for categorical or ordinal

variables, the reliability model used during the analysis, the reliability coefficient, the percentage of agreement for Kappa calculation,

the 95% Cl, and the standard error of measurement for continuous variables.

Var iable (n = 30) Mean (SD) or Model used% of findings

Rel iabi l i ty coeff ic ient

and 7" of

agreement for Kappa

calculat ion

95%CI SEM

Hamstrings length (degrees)

Lateral retinacular length (tight, normal)

Q-angle (degrees)

Tibial torsion (degrees)

Quadriceps length (degrees)

Femoral anteversion (degrees)

Gastrocnemius length (degrees)

Soleus length (degrees)Hip external rotation strength (Kg)

Hip abduction strength (Kg)

ITB/TFL complex length (degrees)

Foot pronation (mm)

Quality of movement( f rom 0to I =good; f rom 2to 3 = medium;,

"n6 "6o"9 = poor)

8 r .s (rs.0)83% tightt2.2 (4.3)r7.6 (s.4)

r38.s (r2.3)r2.8 (6. r )e.3 (s.8)r6.0 (6.0)r7.r (s.2)t2.e (4.6)rs.s (r r . r )s.e (2.7)

33% good50% medium

l7% poor

rcc (2,2)Kappa

rcc (2, r )rcc (2, r )rcc (2, 2)rcc (2, r )rcc (2,2)rcc (2, 2)rcc (2,2)rcc (2,2)rcc (2, 2)rcc (2, r )

Kappa

.927t (e3%)

.70

.70

.91

.45

.92

.85

.79

.85

.97

.93.67 (8o%)

(.82; .e6) 4.3(.57; .86)(.46r .8s) 2.4(.4s; .8s) 2.9(.80; .96) 3.8(. l0; .70) 4.s(.83; .96) 1.6(.7 t ; .94) 2.2(.s6; .e l ) L.4(.68r .93) 1.8(.e3; .e8) 2.1(.84: .e7) 0.7(.58; .76)


BMC Musculoskeletal Disorders 2006. 7:33 http: //www. biomedcentra l. co m | 1 47 1 -247 4 /7 I 33

Table 3: Comparison of reliability coefficient for the overall sample and for each of the pairs of testers.

Variable

Hamstrings length (degrees)Lateral retinacular length (tight, normal)

Q-angle (degrees)Tibial torsion (degrees)

Quadriceps length (degrees)Femoral anteversion (degrees)Gastrocnemius length (degrees)Soleus length (degrees)Hip external rotation strength (Kg)Hip abduction strength (Kg)ITB/TFL complex length (degrees)Foot pronation (mm)

Qual i ty of movement( f rom 0to I =good:from 2to 3 = medium;4and above = poor)

Reliabil ity coefficient

Overal l sample (n = 30).92.71.70.70.91.45.92-db

.79

.oJ

.97

.93

.67

Pair I = ?O\ Pair2(n= l0).98.55.77.83.93.oa

.97

.92

.97

.59

.99

.7)

.ot

(n.88t.0.56.60.90.3 1.78.85.71.86.90.89.6L

DiscussionShrout has suggested a classif icat ion of rel iabi l i ty coeff i-cients in which values less than 0.10 are considered virtu-al ly none agreement; .1 1 to .40 indicate sl ight agreement;.41 to .60 indicate fair agreement; values between .61 and.80 indicate moderate; and values greater than .Bl indi-cate substantial agreement [501. I ]ased on this classif ica-t ion the inter-tester rel iabi l i ty coeff icients were substantialfor measures of hamstrings length, quadriceps length, gas-trocnemius length, soleus length, I ' l 'B/ ' lTl. complexlength, hip abductors strength, and foot pronation. Mod-erate values of rel iabi l i ty were observed for measures of Q-angle, t ibial torsion, hip external rotat ion strength, lateralret inacular t ightness, and test of qual i ty of movement.Measurement of femoral anteversion resulted in fair rel ia-bi l i ty .

' l 'o n-rake val id interpretat ion of measurerrrents, the meas-Llrements must f i rst dernonstrate reasonable rel iabi l i ty.lnterpretat ior.r of the confidence intervals around the val-ues with substantial agreement (above .80) leads to theestimation that the inter-tester rel iabi l i ty of these meas,ures fal ls anlnvhere between .68 and .98. Therefore, con-sidering the worst case ( lower bound of the 95% CI of hipabduction strength of .68), the rel iabi l i ty o[ these meas-ures are still satisfactory for clinical use. Measures with amoderate level of rel iabi l i ty had their confidence intervalsranging from .45 and .91, with the lower bound of theseintervals ranging from .45 to .58, which warrants somecaution when interpreting the f indings of Q-angle, t ibialtorsion, hip external rotat ion strength, t ightness of lateralretinacular structures, and quality of movement. Regard-ing the interpretation of femoral anteversion, both thereliability coefficient value and the confidence intervalssuggest that interpretation of this test's finding may not beconsistent.

We have chosen to focus our investigation on inter-testerrel iabi l i ty, rather than intra-tester rel iabi l i ty for two rea-sons. First, in today healthcare system it is becomingincreasingly common to have more than one cl iniciantreating a patient for the same episode of care. Second,data for this study was obtained as part of a larger mult i-center study that investigated predictors of function inpersons with PITPS. As a result, data were been col lected inmult iple sites by dif ferent cl inicians. Furthermore, whendesigning this study we assumed that the levels of intra-tester rel iabi l i ty would be at least equal or higher than thedetermined inter-tester rel iabi l i ty.

We are not aware of prior studies that determined the rel i-abi l i ty of measuring hamstrings length using the straightleg raise test in a populat ion of patients with PITPS. Ourresults support the f indings in three prior studies and arein confl ict with one study. lwo studies that were per-formed with healthy adults and used standard goniometerto measure the straight leg raises reported intersession cor-relat ion of r = .BB and an ICC for inter-tester rel iabi l i ty of.99 for this measure [54,551. Although we acknowledge i tmay not be appropriate to direct ly compare results of rel i-ability studies that calculated Pearson correlation coeffi-cient with studies that calculated ICC, such comparisongives us at least an approximation of the consistenry ofthe measurement. Another study with a populat ion ofpatients with low back pain that used a gravity goniometerto perform the measure reported an ICC of .87 for theinter-tester rel iabi l i ty and a SEM of 6.4 degrees [56]. Ourresults conflict with the findings of IIunt et al, performedwith healthy individuals [571. They reported fair inter-tester rel iabi l i ty of measuring straight leg raise with anelectronic incl inometer, with ICC of .54 and .48 for theleft and r ight leg respectively [57]. tsecause Hunt, et al, didnot provide a descript ion of subject inclusion cri teria or a



clear description of the test procedure used in their study

[57], i t is not possible to speculate why their measureswere less consistent than our f indings or those of otherstudies. Perhaps the day-long time interval for inter-restermeasures used in Flunt et al's study may have been toolong and al lowed for true variat ion in t issue complianceover t im€.

We elected to rneasure hamstrings length using thestraight leg raise test rather than the popli teal angle test toavoid the potential for ceiling effects with the later test

[46]. In our cl inical experience, the cei l ing effect wi l l hap-pen with several patients with PFPS who may completelyextend the knee before starting to feel the passive ham-str ings resistance during the popli teal angle test. ' l 'here-

fore, in individuals with less hamstrings t ightness, thepopl i teal angle wi l l be l imi ted on the abi l i ty to pick upsubtle t ightness.

Our study yielded better rel iabi l i ty for the patel lar t i l t testthan that reported by Watson et al [24] Watson et al 'sstudy included mainly asymptomat ic indiv iduals (19symptomatic and 76 asymptornatic) as subjects and stu-dents as testers. ' l 'hey reported inter,tester rel iabi l i ty withKappa values ctf .2O, .33, and .35 for the three pair of test-ers, with respective percent agreements of 57o/o, 47o/o, and620/o [241. We bel ieve our study may have had higher rel i-abi l i ty because we used experienced therapists who werefami l iar wi th the test in c l in ical pract ice. Another poten-t ial explanation for such dif ference is the exclusive use ofpatients diagnosed with PI;PS in our study. I Iaving onlypatients with PFPS may increase the incidence of posit ivef indings and result in a more real ist ic determination ofKappa values. Watson et al [24] did not report the inci-clence of posit ive f indings in their study.

I ' r ior studies that used the same methocl as we did toneasure Q-angle have reported lower levels of inter-testerrel iabi l i ty than in our study. ' fomsich et al used a sampleof healthy young individuals tested by therapists withexperience ranging frorn 2.5 to 5.5 years and reported anICC of .23 and a SEM of 3.7" [58] . Greene et a l had 25 test-ers measuring each other's knees, two of whom had patel-lofernoral pain symptonrs. They reported inter-testerrel iabi l i ty with ICC values of .20 and .26 for left and r ightknee respectively [1 11. The better rel iabi l i ty in our studycould be explained by better standardization of measure-ments and training of raters, or because al l our subjectswere diagnosed with PFPS. As increases and decreases rn

Q-angle are associated with increased patel lofemoral pres-sures, it is possible that patients with PFPS have more var-iabi l i ty in the measures of Q-angle than asymptomaticindividuals [42]. The decreased data variabi l i ty in theother studies may have artificially reduced the ICC values.Sutl ive et al measured Q-angle on individuals with PFPS

http://www. biomed central. co ml 1 47 1 -247 4 17 I 33

in a standing posit ion and reported an ICC of .40 (95o/oCl: .08; .70) and a SEM of 4.2 ' [30] . In Sut l ive et a l 's studythey do not give detai ls about the methodology of themeasure [30]. We have chosen to measure Q-angle in anon functional posit ion to avoid contraction of the quad-riceps. Control for quadriceps contraction in a standingposit ion is more dif f icult than in a supine posit ion. Quad-riceps contraction during this measurement could pul l thepatel la sideways and result in inconsistent readings. Webelieve measuring Q-angle with the part icipant in asupine posit ion may yield more consistent results.

Our f inding indicates a fair to poor rel iabi l i ty of theCraig's test to measure femoral anteversion, which is con-sistent with prior studies. One study reported Pearson cor-relat ion coeff icient of r = .47 for inter-tester rel iabi l i ty ofth is test I l5 l and another study reported ICC of .17 [301.The low rel iabi l i ty may be due to the dif f iculty in accu-rately palpating the greater trochanter and determining i tsmost lateral posit ion, especial ly in overweight individu-als. ' [ 'o test this hypothesis, we divided the sample according to body mass index (BMI), in which indiv iduals wi thBMI of .249 or below are classif ied as normal or under-weight, and those with BMI of .25 or above are classif iedas overweight or obese [59]. ' l 'he ICC for the 1l indiv idu,als wi th UMI of .249 or below was .81 (95o/o CI .39; .95),whereas for the l9 individuals with I IMI of .25 or abovewas .20 (95o/o Cl - .30; .60). ' l 'herefore, i t appears that inoverweight individuals rneasurerr lents of femoral antever-sion may be more dif f icult to perform and consequentlyless consistent. LInt i l further study investigates the associ-at ion of t lMI and the consistency of femoral anteversionmeasures we recommend that cl inicians make judgments

based on the results of this measurement with caution.

Measures of dorsif lexion with the knees extelrded orf lexed at 90" resulted in substantial rel iabi l i ty, which is indisagreement with prior studies. Elvery et al reported ICCof .50 for intertester rel iabi l i ty for ankle passive dorsif lex-ion [27]. In another study Youdas et al reported an ICC of.28 for measurements of act ive dorsif lexion [29]. A thirdstudy reported ICCs of .29 and .38 for ankle dorsif lexionwith knee extended and f lexed respectively [30]. Webelieve our study may have resulted in better rel iabi l i ty forseveral reasons: 1) We trained the testers to be consistentwith posit ioning the arms of the goniometer; 2) We srabi,lized the tibia during active dorsiflexion; 3) Measuringactive dorsiflexion performed by the subject removes theconfounding effect oftester strength that could be a prob-lem if dorsif lexion was measured passively; 4) We usedthe average of two tr ia ls.

Our results are in agreement with previous studies thathave indicated good rel iabi l i ty for measures of quadricepslength, hip abduction strength, ITB/TFL complex tight-



ness, and foot pronation. Eng & Pierrynowski have testedthe consistenry of measures of quadriceps length usingthe quadriceps femoris muscle angle in a populat ion offemale with PFPS and reported an ICC of .94 for intra-tester rel iabi l i ty [60]. A prior study that examined the rel i-abi l i ty of measuring hip abduction strength using a handheld dynamometer in runners with i l iot ibial band syn-drome reported substantial inter-tester reliability, with anICC of 0.96 [61]. Another study used Pearson correlat ioncoefficients to determine test-retest reliability using ahand held dynamometer in two boys with muscular dys-trophy and reported correlat ion coeff icients of .86 for hipabduction strength [62]. In a recent study Reese & Bandytested the rel iabi l i ty of measuring ITB/TFL complex inasymptolnatic individuals using the Ober test as a contin-uous measure as we did and reported an ICC of .90 [63].Sel l et al investigated the rel iabi l i ty of measuring foot pro-nation using the navicular drop test and reported an ICCvalue of.73 for the inter-tester rel iabi l i ty [ 141. In disagree,ment with our study and Sell et al study, Sutl ive et alreported an ICC of .51 for the navicular drop test [30].

We identi f ied only one study that investigated the rel iabi l-i ty of measuring t ibial torsion using the same method aswe did and they reporred an ICC of .32 (95o/o CI: .07; .53)with a Sl lM o1 6.4 ' . ' l 'he better rel iabi l i ty in our studycould be explained by better standardization of measure-ments and training of raters.

Another i rnportant point of d iscussion when cornpar ingour results with results from other studies is that the agecrf our subjects ranged from 14 to 47 years, which repre-sents a wider range than most of the other studies. Llavingincluded adolescents as well as adults formed a heteroge-neous sample and potential ly created considerable dif fer-ence between the measurements. I l igher variat ion in therleasurement inf luence the within and between subjectsvariance, both of which can increase the ICC [49].

'I'o our knowledge this is the first study that reports therel iabi l i ty of measuring hip external rotat ion strength andquali ty of movement in patients with PFPS. Quali ty ofmovement was tested during the lateral step down test.This test was developed by our group based on the mala-daptive alterations in lower extremity function that arenormally observed during physical examination inpatients with PFPS 11,4,64,651. In addit ion ro the stepdown test being shown to be rel iable, we bel ieve i t is ableto recognize altered movement patterns commonlyobserved in this populat ion [66]. Further studies shouldvalidate this test against referenced measures of function.

When cornparing the reliability coefficients calculatedwith the data from the overall sample with the valuesobtained from each pair of testers, we observe that the val-

http://www. biomedcentral.co ml 1 47 I -247 417 133

ues are consistent for most measurements. Measures thathave shown greater differences between both pairs werelateral retinacular length, femoral anteversion, and hipabduction strength. These findings reiterate the above dis-cussion that measures of femoral anteversion are not rel i-able and that measures of lateral retinacular lengthwarrant some caution in its interpretation. The differencebetween the two pairs of testers in the measure of hipabduction strength raises addit ional concerns about thismeasurement.

An important element of the val idity of measurem€nts,and the subsequent abi l i ty to accurately interpret thesemeasurements, rel ies on the evidence of satisfactory rel ia-bi l i ty and measurement error [671. Poor rel iabi l i ty andhigh levels of measurement error reduce the usefulness ofa test and l imit the extent to which test results can be gen-eral ized [671. Measurement error, determined in thisstudy by calculat ing the SIIM, refers to the hypotheticaldif ference between an examinee's observed score on anypart icular measurement and the examinee's true score forthe procedure 1671. Knowledge of the Sl,M al lows us roput confidence bands around scores and provides athreshold for interpreting the test results over t ime. LIsingthe SIIM of hamstrings length of 4.3 degrees as an exaln-ple, one can calculate a confidence interval around theobtained score. Let 's suppose that the hamstrings lengthduring the straight leg raise test was B0 degrees. l f one SIIMis added to the hamstrings length measure and one SIMis subtracted from it , an interval is created within whichwe can be 6B%o certain that the true measure fal ls. I f twostandard errors are added to the measure and two stand-ard errors subtracted from it , a wider interval is created,within which we can be 95%o certain that the true measurefal ls. In our example, i f a cl inician rreasures hamstringslength B0 degrees and the SIIM is 4.3 degrees, we can be687o certain that the true hamstrings length is between75.7 and 84.3 degrees and 95% certain that i t is between71.3 and 88.6 degrees. When interpreting changes overtirne, if the measure changes from 80 to 84.3 degrees fromone occasion to the next, one can be 68% confident thattrue change has occurred, if the measure changes from B0to 88.6 degrees, the level of confidence in such changeincreases to 95o/o. Further val idation might be gained infuture studies that determine how responsive to changethese measurements are fol lowin g interventions.

' fhere is currently no consensus regarding the number ofSEMs an individual 's score must change for that change toconfidently exceed measurement error. In other words,there is no agreement about what is the preferred level ofconfidence. Previous researchers have reported one SEMas the best measure of meaningful change on health-related quali ty of l i fe measures [52]. The number of SEMsthat would reflect meaningful change on measures of



physical impairments is not known. Moreover, the SEMhas several properties that make it an attractive statistic fordetermining cl inical ly meaningful change. First, the SEMaccounts for the possibility that some of the changeobserved with a particular measure may be attributable torandom error. Secondly, the SEM is independent of thesample under investigation; that is, the SEM is expected toremain relat ively constant for al l samples taken from agiven populat ion. Third, the SEM is expressed in the orig-inal metr ic of the measure, aiding i ts interpretat ion [521.

One l imitat ion of this study is that the rel iabi l i ty resultsfound may be an over-estimate compared to real cl inicalpractice. Many factors may have inf luenced the measure-ments col lected during this research. ' l -he experimentalenvironment may have been unrepresentative of meas-ures taken in a busy cl inic. Specif ic aspects ofcl inical prac-t ice may lower rel iabi l i t ies values of the measuresinvestigated in the present study. The testers in this studywere well trained to perform the measures and fol lowed astandardized protocol. In the real cl inic, cl inicians workunder t ime restraints and may fol low a less str ict set ofrules when test ing their patients. Furthermore, several var-iat ions in technique may exist across cl inicians. Eventhose who are very accurate in the use of the tests maynever have had the opportunity to standardize their owntechniques with those oIcol leagues. We bel ieve the infor-rnation about the rel iabi l i ty of the measures investigatedin th is study may be of c l in ical re levance i f the c l in ic ianswho intend to use such measures are r ig,orous in the useof the tests as here described and i f they rnake the effort tostandardize the technique with the col leagues.

' fo val idate the use of the measures of impairments asso-ciated with PFPS tested in this study, further research iswarranted in a nurnber of areas. I t should be determinedwhether these irnpairment measurements are related topain and function in individuals with PFPS. l t should alsobe determined whether changes in these impairmentrneasurements wil l be associated with improvement ofpain and function after completing a rehabil i tat ion pro-

8ram.

ConclusionSeveral of the impairments associated with PFPS had goodreliability. Inter-tester reliability coeffi cients were substan-tial for measures of hamstrings length, quadriceps length,plantar flexors length, ITB/TFL complex length, hipabductors strength, and foot pronation, which ensurevalid interpretation of these tests results in clinical prac-tice. Moderate values of reliability were observed formeasures of Q-angle, t ibial torsion, hip external rotat ionstrength, lateral retinacular tightness, and test ofquality ofmovement, which warrants some caution when interpret-ing the findings of these tests. Measurement of femoral

http://www. biomedcentral. coml I 47 1 -247 417 133

anteversion resulted in fair reliability, suggesting thatinterpretation of this test may not be consistent. Addi-tional evidence is needed to support their use by testing ifthese impairment measurements are related to physicalfunction and whether or not they can be used to guidetreatment planning which ult imately would result in suc-cessful treatment outcofires.

Competing interests' I 'he author(s) declare that they have no competing inter-ests.

Authors' contributionsSRP conceived and coordinated the study, performed sta-tistical analysis, and drafted the manuscript. KF, JII, andJDC participated in the study design and revision of man-uscript. SJ, BRFI, and DAB have acquired the data andwere involved in drafting the manuscript. All authors readand approved the final manuscript.

AcknowledgementsFunding from Clinical Research Grant Program of Orthopaedic Section of

American Physical Therapy Association and the Pennsylvania Physical Ther-

apy Associat ion Research Fund.

Referencesl. Brody LT, Thein JM: Nonoperative treatment for patellofemo-

raf pain. J Orthop Spons Phys Ther | 998, 28:335-344.2. Dehaven KE, Dolan WA, Mayer PJ: Chondromalacia patellae in

athletes. Clinical presentation and conservative manage-ment. Am J Spons fied | 979, 7:5- | | .

3. Rubin B, Col l ins R: Runner 's knee. Phys Sportsmed 1980,8:49-58.4. Grabiner MD, Koh TJ, Draganich LF: Neuromechanics of the

patef lofemoral ioint. Med Sci Sporu Exerc 1994,26: l0-2 | .5. Sikorski JM: lmportance of femoral rotation in chondromala-

cia pateffae as shown by serial radiography. I BonelointSurg(Br)t979,6l :435-442.

6. Duffey MJ, Martin DF, Cannon DW, Craven T, Messier SP: Etiologicfactors associated with anterior knee pain in distance run-ners. Med Sci Spons Exerc 2000, 32:1825-1832.

7. Messier SP, Davis SE, Curl WW, Lowery RB, Pack RJ: Etiologic fac-tors associat€d with patellofemoral pain in runners. rlled SciSpons Exerc 1991, 23: 1008-1015.

8. Witvrouw E, Bel lemansJ, Lysens R, Danneels L, Cambier D: Intr insicrisk factors for the development of patellar tendinitis in anathletic population, A two-year prospective study. AmJ SponsMed 20Q1,29: | 90- | 95.

9. Guyatt GH, Kirshner B, Jaeschke R: Measuring health status:what are the necessary measurement properties? I Clin Epide-miol 1997, 45: | 341 - | 345.

10. Kirshner B, Guyatt G: A methodological framework for assess-ing health indices. I Chronic Dis 1985, 38:27-36.

I l. Greene CC, Edwards TB, Wade MR, Carson EW: Reliability of thequadriceps angle measurement, Am J Knee Surg 200l,| 4:97- | 03.

17. Rome K: Ankle jo int dorsi f lexion measurement studies. Areview of the literature. I Am Podiotr MedAssoc 1995, 85:205-21 l.

13. Holmes CF, Wilcox D, Fletcher JP: Effect of a modified, low-dyemedial longi tudinal arch taping procedure on the subtalarjoint neutral position before and after light exercise. J OnhopSpons Phys Ther 2002,32:194-701.

14. Sell KE, Verity TM, Worrell TW, Pease BJ, Wiggleswonh J: Twomeasurement technigues for assessing subtalar jo int posi-tio n: a reliabil ity study. J O nhop Spo rts Phys Ther | 99 4, | 9 z | 62- | 67 .

15. Gunther KP, Kessler S, Tomczaz R, Pfei fer P, Puhl W: Femoralanteversion: s igni f icance of c l in ical methods and imaging


t5.

t7.

t8.

t9.

20.

2t .

22

L5

24


techniques in the diagnosis in chi ldren and adolescents.Zeitschrift fur Orthopodie uns lhre Grenzgebiete | 996, | 34:295-30 | .McConnell J, Fulkerson JP: The Knee: Patellofemoral and sofr!tissue injuries. In Athletic lnjuries ond Rehobilitotion Edited by: Zach-azewski JE, Magee DJ and Qui l len WS. Phi ladelphia, PA, W.B. Saun-ders Cor 1996:693-778.R H: Pelvic stability influences lower-extremity kinematics.Biomechanics | 998, 5:23-29.l re land ML, Wi l lson JD, Bal lantyne BT, Davis lM: Hip strength infemales with and without patellofemoral pain. / Orthop SportsPhys Ther 2003, 33:571-676.Piva SR, EA G, Chi lds JD: Strength around the hip and f lexibi l i tyof sof t t issues in indiv iduals wi th and without patel lofemoralpain syndrome. I Orthop Spons Phys fher 2005, In Review:.Hertling D, Kessler RM: Monogement of Common lVusculoskeletol Dis-orders 3rd edition. Philadelphia/PA, Lippincott; | 995:3 | 5-378.Witvrouw E, Lysens R, Bellemans J, Cambier D, Vanderstraeten G:Intrinsic risk factors for the development of anterior kneepain in an athletic population. A two-year prospective study.Am J Spons Med 2000, 28:480-489.Smith AD, Stroud L, McQueen C: Flexibi l i ty and anter ior kneepain in adolescent elite figure skaters. J Pediotr Orthop 1991,I l :77-87.Wilk KE, Davies GJ, Mangine RE, Malone TR: Patel lofemoral d isor-ders: a c lassi f icat ion system and cl in ical guidel ines for nonop-erative rehabilitation. J Onhop Sports Phys Ther | 998, 28:307-322.Watson CJ, Leddy HM, Dynjan TD, Parham JL: Rel iabi l i ty of the lat-eral pull test and tilt test to assess patellar alignment in sub-iects with symptomatic knees: student raters. J Onhop SportsPhys fher 200 l , 3 | :368-374.Baggett BD, Young G: Ankle joint dorsiflexion, Establishment ofa normaf range. J Am Podiotr Med Assoc 1993, 83:25 l -254.Boone DC, Azen SP: Normal range of motion of joints in malesubiects. J Bone Joint Surg Am 1979,612756-759.Elveru RA, Rothstein JM, Lamb RL: Goniometr ic rel iabi l i ty in acl in ical set t ing. Subtalar and ankle jo int measurements. Physfher 1988,58:677-677.Stefanyshyn DJ, EngsbergJR: Right to left differences in the ankleioint complex range of motion, /v1ed Sci Spons Exerc 1994,25:55 I -555.Youdas JW, Bogard CL, Suman VJ: Rel iabi l i ty of goniometr icmeasurements and visual estimates of ankle joint activerange of motion obtained in a clinical setting. Arch Phys MedRehobi/ 1993,74: l I l3- l | 18.Sut l ive TG, Mitchel l SD, Maxf ie ld SN, McLean CL, Neumann JC,Swiecki CR, Hal l RC, Bare AC, Flynn TW: ldent i f icat ion of Indiv id-uals With Patellofemoral Pain Whose Symptoms lmprovedAfter a combined Program of Foot Orthosis use and Modi-fied Activity: A Preliminary Investigation. Phys Ther 2004,84:49-5 | .Powers CM, Maffucci R, Hampton S: Rearfoot posture in subjectswith patellofemoral pain. J Orthop Sports Phys Ther 1995,22:1 55- 1 50.Eckhoff DG. Brown AW. Kilcoyne RF. Stamm ER: Knee versionassociated with anter ior knee pain. Ct in Orthop 1997:152-155.Eckhoff DG, Montgomery WK, Kilcoyne RF, Stamm ER: Femoralmorphometry and anterior knee pain. Ain Orthop 1994:64-68.Cesarelli M, Bifulco P, Bracale M: Study ofthe control strategy ofthe quadriceps muscles in anterior knee pain. |EEETrons Reho-bil Eng 2000, 8:330-34 | .Jensen MP, Karoly P, Braver S: The measurement of c l in ical painintensi ty: a comparison of s ix methods, Poin 1986, 27: l 17-176.Jensen MP, Turner JA, Romano JM: What is the maximumnumber of levels needed in pain intensi ty measurement? Poin1994, 58:387-392.KatzJ, Melzack R: Measurement of pain. Surg Clin Nonh Am 1999,79:23l -252.Stratford PW, Binkley FM, Riddle DL: Health status measures:strategies and analytic methods for assessing change scores.Phys Ther 1996, 7 6: l | 09- | | 23.l r rgangJJ, Snyder-Mackler L, Wainner RS, Fu FH, Harner CD: Devel-opment of a patient-reported measure of function of theknee. J Eone Joint Surg Am | 998, 80: | | 32- | | 45.Marx RG, Jones EC, Al len AA, Al tchek DW, O'Br ien SJ, Rodeo SA,Williams RJ, Warren RF, Wickiewicz TL: Reliability, validity, and

http://www. biomedcentral. co ml 1 47 1 -247 417 133

responsiveness of four knee outcome scales for athleticpatients. J Bone Joint Surg 200 | , 83-A: | 459- | 469.Kolowich PA, Paulos LE, Rosenberg TD, Farnsworth S: Lateralrelease of the patel la: indicat ions and contraindicat ions, AmI Sports Med | 990, I 8:359-365Huberti HH, Hayes WC: Patellofemoral contact pressures. Theinfluence of q-angle and tendofemoral contact, I Bone loint SurgAm 1984, 561715-72.4.Giallonardo LM: Clinical evaluation of foot and ankle dysfunc-tion. Phys lher | 988, 68: | 850- | 855.Gross MT: Lower quarter screening for skeletal malalign-ment--suggestions for orthotics and shoewear. J Onhop SportsPhys Ther | 995, 2 l:389-405.Magee D: Onhopoedic Physicol Assessment 2nd edition. Philadelphia,PA, W.B. Saunders Company: 1992.Kendall FP, McCreary EK, Provance PG: Muscles Testing ond Funaion4th edi t ion. Bal t imore, MD, Wi l l iams & Wilk ins; 1993.TG MP, HG K: A survey of foot types in normal femalesbetween the ages of | 8 and 30 years. J Onhop Spons Phys Ther| 985, 9:405-409.Cohen J: Coefficient of agreement for nominal scales. Educ Psy-chol Meos | 950, 20:37.Shrout PE, Fleiss JL: lntraclass correlat ions: Uses in assessingrater ref iabil ity. Psychol Bull | 97 9, 86:470 - 428.Shrout PE: Measurement reliability and agreement in psychi-atry. Stdt Methods Med Res | 998, 7:30 | -3 | 7.Lydick E, Epstein RS: lnterpretation of quality of life changes.Quol Life Res 1993, 2:221-226.Wyrwich KW, Nienaber NA, Tierney WM, Wol insky FD: Linkingclinical relevance and statistical significance in evaluatingintra-individual changes in health-related quality oI lile, MedCore 1999,37:469-478.SD W, M E, A D: Sample size and optimal designs for reliabilitystudies. stotistics in Medicine | 998, | 7: l0 | - I | 0.Clark S, Chr ist iansen A, Hel lman DF, HuguninJW, Hurst KM: Efrectsof ipsilateral anterior thigh soft tissue stretching on passiveunifateral straight-leg raise. J Onhop Spons Phys Ther 1999,29:4-9.Hsieh CY, WalkerJM, Gi l l is K: Straight- leg-rais ing test . Compar-ison of three instruments. Phys fher 1983,63:1429-1433.Fr i tz JM, Piva SR: Physical impairment index: rel iabi l i ty , val id-ity, and responsiveness in patients with acute low back pain.Sbine 2003.28: l 189- l 194.Hunt DG, Zuberbier OA, Kozlowski AJ, Robinson J, Berkowitz J,Schul tz lZ, Mi lner RA, CrookJM, Turk DC: Rel iabi l i ty of the lum-bar flexion, lumbar extension, and passive straight leg raisetest in normal populat ions embedded within a completephysicaf examination. Spine 200 l, 26t27 | 4 -17 | 8.Tomsich DA, Nifz AJ, Threlkeld AJ, Shapiro R: Patellofemoralafignment: reliability. J Orthop Spons Phys Ther 1995, 23:200-208.Flegal KM, Carrol l MD, Ogden CL, johnson CL: Prevalence andtrends in obesi ty among US adul ts, 1999-2000. JAMA2007,288:1723-1727.Eng JJ, Pierrynowski MR: Evaluation of soft foot orthotics in thetreatment of patellofemoral pain syndrome. Phys Ther 1993,7l:62-68.Freder icson M, Cookingham CL, Chaudhar i AM, Dowdel l BC, Oest-reicher N, Sahrmann SA: Hip abductor weakness in distancerunners with iliotibial band syndrome. CIin J Sport ftled 2000,l 0:1 69- 1 75.Hyde SA, Goddard CM, Scott OM: The myometer: the develop-ment of a cf inical tool. Physiotheropy 1983, 59:424-427.Reese NB, Bandy WD: Use of an inclinometer to measure flex-ib i l i ty of the i l iot ib ia l band using the Ober test and the mod-ified Ober test: differences in magnitude and reliability ofmeasu rements. I O rthop Sporu Phys lher 2003, 3 3:325-3 30.Fulkerson JP: Disorders ofthe Potellofemorol Joint3rd edition. Edired by:Fulkerson JP. Bal t imore, f4D, Wi l l iams & Wilk ins; 1997.Thomee R, Renstrom P, Kar lsson J, Gr imby G: Patel lofemoral painsyndrome in young women. l . A c l in ical analysis of a l ign-ment, pain parameters, common symptoms and functionalactivity f evef . Scond J Med Sci Sporu | 995, 5 :237 -744.Vaatainen U, Airaksinen O, Jaroma H, Kiviranta l: Decreasedtorque and electromyographic act iv i ty in the extensor th ighmuscfes in chondromalacia patellae, lnt ./ Sports Med 1995,| 6:45-50.


4l

42.

43.

44.

45.

46.

47.

48.

49.

50.

5t .

(')

25.

LO.

27.

53.

54.

55.

56.

i t .

16-

29

30.

Jt.

32.

33.

34.

58.

59.

o/

bJ

64

65

35.

JO.

37.

38.

39.

60.

5l

40.66


57. Joint committee on standards for educational and psychot-ogocal testing of the American Educational Research Asso-ciat ion, American Psychological Associat ion, Nat ionalCounci l on Measurement in Educat ion, Standards for Educa-tional and Psychological Testing. Washington DC, AmericanEducational Research Association: 2002.

Pre-publication history'Ihe pre-publication history for this paper can be accessedhere:

http://www.biomedcentral.com/ 1 47 1 -24 74/7/33/prepub

http://www. biomedcentral. coml 1 47 1 -247 417 133

Publish with BiolUled{.e,*r*ral and everyscient ist can read your work f ree of charge

"BioMed Central will be the most significant development for

disseminating the resu/ts of biomedical research in our l i fetime."

Sir Paul Nurse, Cancer Research UK

Your research papers wil l be:. avai lable f ree of charge to the ent i re biomedicalcommunity

. peer reviewed and publ ished immediately upon acceptance

. c i ted in PubMed and archived on PubMed Central

. yours - you keep the copyright

-...Submit your manuscr ipt herer ( | BioMedcentralhttp: / /wwwbromedcentrdl . (omlnlo/publ tshinq adu.asp V


reliability of measures-pfps.pdf

Documents