Reliability and discriminatory testing of a client-based metrology instrument, feline musculoskeletal pain index (FMPI) for the evaluation of degenerative joint disease-associated

The Veterinary Journal 196 (2013) 368–373

Contents lists available at SciVerse ScienceDirect

The Veterinary Journal

journal homepage: www.elsevier .com/ locate / tv j l

Reliability and discriminatory testing of a client-based metrology instrument,feline musculoskeletal pain index (FMPI) for the evaluation of degenerativejoint disease-associated pain in cats

J. Benito a, V. DePuy b, E. Hardie a, H. Zamprogno a, A. Thomson a, W. Simpson c, S. Roe a, B. Hansen a,B.D.X. Lascelles a,d,⇑a Comparative Pain Research Laboratory, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, Raleigh, NC, USAb Bowden Statistical Consulting, Oxford, NC, USAc Morrisville Cat Hospital, Morrisville, NC, USAd Center for Comparative Medicine and Translational Research, College of Veterinary Medicine, North Carolina State University, Raleigh, NC, USA

a r t i c l e i n f o

Article history:Accepted 19 December 2012

Keywords:AssessmentDegenerative joint diseaseFelinePain

1090-0233/$ - see front matter Published by Elsevierhttp://dx.doi.org/10.1016/j.tvjl.2012.12.015

⇑ Corresponding author at: Comparative Pain ResearClinical Sciences, College of Veterinary Medicine, NoRaleigh, NC, USA. Tel.: +1 919 513 6762.

E-mail address: [email protected] (B.D.X

a b s t r a c t

The objective of this study was to test the readability, reliability, repeatability and discriminatory abilityof an owner-completed instrument to assess feline degenerative joint disease (DJD)-associated pain(feline musculoskeletal pain index, FMPI). Readability was explored using four different formulas (Flesch,Fry, SMOG and FOG) and the final FMPI instrument was produced. To assess the instrument, client-ownedcats that were defined as normal (normal group) or as having DJD-associated pain and mobility impair-ment (pain-DJD group) were recruited. A total of 32 client-owned cats were enrolled in the study (nor-mal, n = 13; pain-DJD, n = 19). Owners completed the FMPI on two occasions, 14 days apart. Internalconsistency (reliability) and repeatability (test–retest) were explored using Cronbach’s a and weightedj statistic, respectively. Data from the two groups were compared using analysis of covariance (control-ling for age) to evaluate discriminatory ability.

The FMPI was constructed with 21 questions covering activity, pain intensity and overall quality of life.It had a 6th grade readability score. Reliability of the FMPI was excellent (Cronbach’s a > 0.8 for all group-ings of questions in normal and pain-DJD cats) and repeatability was good (weighted j statistic >0.74) fornormal and pain-DJD cats. All components of the FMPI were able to distinguish between normal cats andcats with DJD (P < 0.001 for all components). This initial evaluation of the FMPI suggests that this instru-ment is worthy of continued investigation.

Published by Elsevier Ltd.

Introduction

Published information indicates that degenerative joint disease(DJD) is common in domesticated cats (Beadman et al., 1964;Clarke and Bennett, 2006; Clarke et al., 2005; Godfrey, 2005; Har-die et al., 2002; Koeppel and Ebner, 1990; Langenbach et al., 1998;Lascelles et al., 2010b; Slingerland et al., 2011). Several studieshave identified cats with radiographic DJD and mobility impair-ment (Bennett and Morton, 2009; Clarke and Bennett, 2006; Las-celles et al., 2001, 2007, 2010a) and these studies have foundthat non-steroidal anti-inflammatory drug (NSAID) administrationsignificantly improved mobility (Clarke and Bennett, 2006; Las-celles et al., 2001, 2007).

Ltd.

ch Laboratory, Department ofrth Carolina State University,

. Lascelles).

In the USA there are no approved drug therapies for the treat-ment of feline chronic musculoskeletal pain. This is partly due tothe fact that there is no validated assessment tool for use in clinicaltrials, such as an owner-completed questionnaire. Previous work inour laboratory has identified items that might be valid for inclusionin such a questionnaire and we have also identified the preferredstructure of such a questionnaire (Zamprogno et al., 2010). Theaim of the study described here was to build on previous workand further develop and evaluate a questionnaire for the assess-ment of feline musculoskeletal pain.

We hypothesized that an appropriately developed subjectiveowner-completed instrument to assess chronic feline DJD-associ-ated pain (feline musculoskeletal pain index, FMPI) would provereliable and have discriminatory validity. The objectives of thisstudy were (1) to use information generated in previous work(Zamprogno et al., 2010) to construct a candidate instrument; (2)to perform readability testing of the candidate instrument andgenerate the test instrument (FMPI); (3) to perform reliability test-

J. Benito et al. / The Veterinary Journal 196 (2013) 368–373 369

ing of the FMPI (test–retest and internal consistency) in normalcats and cats with DJD-associated pain, and (4) to test the instru-ment for its ability to discriminate between normal cats and catswith DJD-associated pain.

Materials and methods

Animals

This study was approved by the Animal Care and Use Committee at NorthCarolina State University College of Veterinary Medicine (NCSU-CVM; IACUC08-124-O), and informed owner consent was granted in each case. The instrumentwas constructed, tested for readability and adjusted accordingly. Then, usingnormal cats and cats with DJD-associated pain and mobility impairment, the instru-ment was evaluated for reliability and repeatability and its ability to discriminatebetween normal and DJD-affected animals.

The aim was to recruit a total of 40 cats to the study, namely, 20 normal client-owned cats without reported mobility impairment, and radiographic DJD and painscores that were considered clinically insignificant (normal group), and 20 cats withmobility impairment, significant radiographic DJD and at least one joint with bothradiographic DJD and pain on manipulation (DJD group). Potential owner partici-pants were required to be 18 years or older, to appear to understand the studyand the study demands, and to be able to keep their household routine as constantas possible for the duration of the study.

Evaluation of potential study candidates (screening)

Prior to the screening visit, owners of cats that could potentially be enrolled inthe study were contacted and sent a consent form to review and a detailed ques-tionnaire to complete. The questionnaire contained approximately 70 questionspertaining to the cat’s lifestyle and mobility, and was reviewed to determine if therewere indicators of mobility impairment. Potential candidates were then screened atthe Veterinary Health Complex. Owners who reported any mobility impairmentcompleted a Client Specific Outcome Measures evaluation as previously described(Lascelles et al., 2007) to determine whether any specific activities were impaired.Cats were screened with a physical examination, orthopedic and neurologicalexaminations, CBC, blood chemistry, urine analysis and orthogonal radiographs ofall appendicular joints and all parts of the axial skeleton. Each cat was weighed,and body condition score (BCS) and temperament (Lascelles et al., 2012) recorded.

Inclusion criteria

To be included in the normal group, cats were required to have a total painscore (see below) of 65 with no one joint pain score being >2, and a total DJD score(see below) of 610, with no one joint score >4 and no evidence of owner-reportedmobility impairment. To be included in the pain-DJD group, cats were required tohave a total pain score >5 and a total DJD score >10, and have overlap of positivepain and DJD scores in at least one joint. Additionally, cats in the pain-DJD groupwere required to have a CSOM score of at least 7 (with at least two of the five activ-ities with a score of at least 2, the other three having a score of at least 1).

Orthopedic evaluation

Every joint (the manus and pes were considered single joints) and each part ofthe axial skeleton (cervical, thoracic, lumbar, and lumbo-sacral) was palpated andmanipulated. A pain response score resulting from the manipulation was assignedto each joint or axial skeleton segment on a 0–4/4 scale, as described previously(Lascelles et al., 2010a, 2012; Zamprogno et al., 2010). The total pain score wasthe addition of all the individual appendicular joint and axial skeleton segmentscores.

Under sedation, orthogonal radiographs of all the appendicular joints and axialskeleton segments were made. The orthopedic evaluation was repeated under seda-tion to rule out concomitant orthopedic diseases. Radiographs were reviewed andscored by two investigators (BDXL and JB) as previously reported (Freire et al.,2011; Lascelles et al., 2010b) resulting in a DJD score (out of 10) for each jointand part of the axial skeleton. The addition of these resulted in a total DJD score(maximum 200).

Exclusion criteria

Exclusion criteria for all cats included the presence of suspected or diagnosedinfectious diseases, symptomatic cardiac disease, immune-mediated disease, neo-plasia, moderate or severe renal disease (see later), inflammatory bowel disease,urinary tract infection, hyperthyroidism, and diabetes mellitus. These diagnoseswere ruled out by careful review of medical records, owner history, physical exam-ination, blood work, and urine analysis. For the normal group, cats were excluded if

any orthopedic disease (e.g. cruciate ligament rupture, joint luxation) or neurolog-ical disease (e.g. lumbo-sacral nerve impingement) that might impair mobility wasdetected.

Additionally, eligible cats were required to not have received any anti-inflam-matory medications for at least 4 weeks prior to the study, to have been on thesame diet with or without nutritional supplements for at least 4 weeks prior tothe study, to be free from clinically abnormal hematological or blood chemistry val-ues (creatinine increases up to 2.8 mg/dL were acceptable if they had been stablefor at least 4 weeks before the study and this was documented with plasma bio-chemistry results), and be housed indoor-only. Owners were required to have a sta-ble routine of daily living that was unlikely to change over the duration of the study.

Instrument design and readability testing

Using the question topics and the preferred instrument design previously re-ported (Zamprogno et al., 2010) an owner-directed questionnaire (the FMPI) wasconstructed. The questionnaire followed a descriptive rating scale design withdescriptors from left to right (normal to abnormal). The instrument was createdand adapted following input from all investigators, then tested for readability (ClearLanguage Group, CLG; Mettger Communications and Plain Language Works, LLC)based on four readability formulas, namely, Flesch, Fry, SMOG and FOG (Fry,1968; Kincaid et al., 1975; McLaughlin, 1969). Following changes to optimize read-ability, the final 21-question FMPI instrument was produced.

Reliability testing (internal consistency and test–retest)

Questionnaires from owners of normal and pain-DJD cats were used for reliabil-ity testing. At the screening visit, and prior to any evaluation of the cat, ownerscompleted the FMPI (D1), and they returned to the clinic 14 days later (D14) tocomplete the FMPI a second time. Completion of the instrument was done in thesame environment on each occasion, and administered by the same person eachtime. A standard paragraph of instructions was read to the owner prior to comple-tion of the instrument. Data from D1 and D14 were used to calculate reliability andinternal consistency (see below).

Discriminatory ability testing

FMPI data in normal and pain-DJD cats were collected as described above. FMPIinstrument scores at D1 and D14 were compared between groups. In an attempt tosee if the instrument could discriminate between degrees of impairment, the pain-DJD cats were divided into low DJD (L-DJD) and high DJD (H-DJD) by adding the to-tal radiographic DJD and total pain scores for each cat, and dividing the pain-DJDcats into two groups based on the median value of the combined score. D1 andD14 data were compared between groups.

Statistical analysis

Power analysis calculations for this study were based on preliminary data sug-gesting a standard deviation of 10 in a group of cats with DJD, and a decision thatwe wanted to discriminate between normal cats (mean score 0/83) and cats withsome mobility impairment (a positive score on half of the activity questions andtwo out of three of the pain and quality of life questions – score of 11/83). Theseparameters suggested that in the discrimination portion of the study, a total of38 cats were required to achieve statistical power of 90%.

Data were compared between groups using Fisher’s Exact and Mann–Whitneytests as appropriate. If owners selected ‘I don’t know’, or ‘Does not apply’, thesewere considered missing data points. An estimate of internal consistency reliabilityof the FMPI questions for both normal and pain-DJD cats was assessed using Cron-bach’s a. Repeatability was assessed by calculation of the weighted j statistic forFMPI scores from D1 and D14 visits in the normal and pain-DJD cats. The FMPIscores evaluated were the FMPI activity score (sum of questions 1–18), the FMPIpain score (sum of questions 19–20), FMPI quality of life (QoL) score (question21) and the FMPI total score (sum of questions 1–21).

In addition to calculation of the weighted j statistic, repeatability was evalu-ated by comparing D1 and D14 data within each group using a Wilcoxon signedrank test. To test discriminatory ability, FMPI scores were grouped together withinday of assessment, and ranked from smallest to largest, average values were as-signed to ties and ages of cats were similarly grouped together and ranked. An anal-ysis of covariance was performed on the ranked FMPI scores, using ranked age andnormal/pain-DJD group as covariates. The P value for the effect of group, after con-trolling for ranked age, was then reported. As a sensitivity analysis, the same regres-sion was performed using age instead of ranked age. Data from D1 and D14 werecompared between both DJD subgroups (L-DJD and H-DJD) using a Mann Whitneytest. In all analyses, FMPI activity, FMPI pain, FMPI QoL and FMPI total scores werecompared.

370 J. Benito et al. / The Veterinary Journal 196 (2013) 368–373

Results

Instrument design and readability

The constructed FMPI included 21 questions, with a maximumscore of 83 (maximum impairment/pain). Eighteen questionsasked about the cat’s ability to perform different activities (activityfactors), two related to pain intensity (severity factors), and one re-lated to overall QoL. The 18 questions (Q) regarding activities askedabout walking (Q1), running (Q2), jumping up (Q3), jumping up tokitchen counter height (Q4), jumping down (Q5), climbing upstairs (Q6), going down stairs (Q7), playing with/chasing toys(Q8), playing and interacting with other pets (Q9), getting up froma resting position (Q10), lying or sitting down (Q11), stretching(Q12), grooming (Q13), interacting with the owner (Q14), beingtouched or handled (Q15), eating (Q16), using the litter box(Q17) and overall activity (Q18). Question 19 asked about pain overthe last week, and question 20 asked about pain on the day ofassessment; question 21 asked about overall QoL.

Activity questions followed a design of a descriptive rating scalewith seven descriptors (from left to right: ‘above normal’; ‘normal’;‘not quite normal’; ‘somewhat worse than normal’; ‘barely, or withgreat effort’; ‘not at all’; ‘does not apply or I don’t know’). Scoresallotted to these options were: �1, 0, 1, 2, 3 and 4 respectively,with ‘does not apply or I don’t know’ not being scored. Pain sever-ity factor questions followed a design of a descriptive rating scalewith five descriptors (from left to right: ‘no pain’; ‘little pain’; ‘mildpain’; ‘moderate pain’; ‘severe pain’). Scores allotted to these op-tions were 0, 1, 2, 3 and 4 respectively. Overall QoL questions fol-lowed a descriptive rating scale with four descriptors (from left toright: ‘excellent’; ‘good’; ‘fair’; ‘poor’). The questionnaire startedout with a readability score of ‘college graduate’, and the final ver-sion had a 6th grade readability score.

Demographic data

Forty-nine cats were screened and 32 matched the inclusioncriteria for the normal (n = 13) and pain-DJD (n = 19) groups(Table 1). Age (P = 0.0015), total (P < 0.0001) and appendicular(P < 0.0001) radiographic DJD scores, and total (P < 0.0001) andappendicular (P < 0.0001) pain scores were significantly differentbetween the normal and pain-DJD groups. Within the pain-DJDgroup, the median value resulting from the addition of total DJDand total pain scores was 32. Of the 19 pain-DJD cats, seven weredesignated L-DJD and 12 H-DJD. Age (P = 0.020, Table 2), totalradiographic DJD scores (P = 0.042), and total (P = 0.0247) andappendicular (P = 0.0304) pain scores were significantly differentbetween the L-DJD and H-DJD groups.

Table 1Data for normal cat group and DJD group and results of statistical comparison between g

Normal cats (n = 13)

Sex (n) M (0); MC (9); F (1); FS (3)

Age, yearsMean (SD) 7.4 (3.6)Median (range) 6.7 (2.1–14.5)

Weight, kgMean (SD) 5.5 (1.6)Median (range) 5.2 (2.6–8.8)

BCSMedian (range) 5 (5–9)

TemperamentMedian (range) 0 (0–3)

SD, standard deviation; M, male intact; MC, male castrated; F, female intact; FS, female

Reliability testing

Internal consistency reliability of the FMPI for both normal andpain-DJD groups was found to be excellent. Cronbach’s a was 0.95in the normal cats and 0.90 in the pain-DJD cats for FMPI activityquestions (Q1–18), 0.84 in the pain-DJD cats for pain questions(19 and 20), 0.94 in the normal cats for all questions combined,and 0.91 in the pain-DJD cats for all questions combined. A valuefor the pain questions in the normal could not be calculated dueto the high number of zero values. The repeatability (test–retest)of FMPI scores and subsets of scores was good in both normaland pain-DJD groups, with all weighted j statistic values over0.74 (Table 3).

There were no statistical differences between D1 and D14 datain normal cats for any of the FMPI domain scores (P > 0.05), how-ever FMPI activity and FMPI total scores increased significantly(P = 0.0142 and P = 0.0141, respectively) between D1 and D14 forthe pain-DJD group (Table 4).

Discriminatory ability

All components of the FMPI (activity, pain, quality of life and to-tal scores) were able to discriminate between normal and pain-DJDcats on both D1 and D14, even after controlling for age (Table 5).When the individual FMPI questions were reviewed, the questionpertaining to eating (Q 16) did not discriminate between normaland pain-DJD cats at both D1 and D14 (Table 6). Additionally, thequestion pertaining to interacting with the owner (Q 14) did notdistinguish between normal and pain-DJD cats at D1. Due to theexploratory nature of these analyses, no adjustments were madefor multiple comparisons. No components of the FMPI were ableto distinguish between L-DJD and H-DJD subgroups (Table 7).

Discussion

In this study, we attempted to design a disease -specific owner-administered clinical metrology instrument to assess DJD-associ-ated pain in the cat. To our knowledge, this is the first attempt toperform such an exercise. The instrument had good internal con-sistency (reliability) for both normal and pain-DJD cats. Repeatabil-ity (test–retest) was good in both normal and pain-DJD cats butthere was a suggestion that there was some movement (increase)in the FMPI activity and total scores with time. Overall, the abilityof the FMPI to discriminate between normal and pain-DJD cats wasvery good, but the FMPI did not discriminate between diseaseseverity subgroups, although the numbers of cats in each of thesegroups was small.

roups.

DJD cats (n = 19) P value

M (0); MC (6); F (0); FS (13) 0.070

12.2 (3.0)12.9 (5.0–18.3) 0.0015

5.6 (0.9)5.5 (4.1–7.5) 1.0000

7 (4–9) 0.4072

1 (0–4) 0.1173

spayed; DJD, degenerative joint disease.

Table 2Data for cats in the low pain DJD (L-DJD) and high pain DJD (H-DJD) groups and results of statisticalcomparison between groups.

L-DJD cats (n = 7) H-DJD cats (n = 12) P value

Sex (n) MC (3); FS (4) MC (3); FS (9) 0.6613

Age, yearsMean (SD) 10.3 (2.8) 13.3 (2.2)Median (range) 11.2 (5.0–13.2) 13.6 (9.0–18.3) 0.0201

Weight, kgMean (SD) 5.8 (0.9) 5.3 (1.0)Median (range) 6.2 (4.3–7.0) 5.3 (4.1–7.5) 0.2535

BCSMedian (range) 7 (4–9) 6 (4–7) 0.4309

TemperamentMedian (range) 1 (0–3) 1 (0–4) 0.7928

SD, standard deviation; M, male intact; MC, male castrated; F, female intact; FS, female spayed; DJD,degenerative joint disease.

Table 3Weighted j statistic between D1 and D14 FMPI scores (95% confidence intervals) forthe normal and pain-DJD groups.

Normal cats Pain-DJD cats

FMPI – activity 0.91 (0.82–1.00) 0.82 (0.73–0.91)FMPI – pain 0.89 (0.67–1.00) 0.74 (0.59–0.90)FMPI – QoL 0.89 (0.67–1.00) 0.75 (0.55–0.95)FMPI – total score 0.90 (0.80–1.00) 0.83 (0.74–0.92)

DJD, degenerative joint disease; FMPI, feline musculoskeletal pain index; QoL,quality of life.

J. Benito et al. / The Veterinary Journal 196 (2013) 368–373 371

We followed similar approaches to the design of our instrumentas have been recommended in human medicine (McDowell andNewell, 1996; Streiner and Norman, 1995; Sudman and Bradburn,1982). Such approaches have recently been adopted in veterinarymedicine for canine metrology instruments (Brown et al., 2007,2008; Hercock et al., 2009; Hielm-Bjorkman et al., 2003, 2009).Our instrument was initially developed along the lines of theCanine Brief Pain Inventory (CBPI; Brown et al., 2007) but endedup being constructed somewhat differently following evaluationof instrument designs by owners, veterinarians and statisticians(Zamprogno et al., 2010). The FMPI was developed around ques-tions asking directly about the cat’s ability to perform activitiesrather than how the disease (DJD) could interfere with the cat’sability to perform activities (the approach taken by the CBPI).

Table 4Comparison of D1 and D14 FMPI scores, median (range), within the normal and pain-DJD

Normal (Day 1) Normal (Day 14) P

FMPI – activity 0 (�18 to 2) 0 (�17 to 4) 0.4FMPI – pain 0 (0–0) 0 (0–1) NAFMPI – QoL 0 (0–0) 0 (0–1) NAFMPI – total score 0 (�18 to 2) 0 (�17 to 4) 0.2

DJD, degenerative joint disease; FMPI, feline musculoskeletal pain index; NA, not analyz

Table 5Comparison of FMPI scores after controlling for age, median (range), between the normal

Normal (Day 1) Pain-DJD (Day 1) P v

FMPI – activity 0 (�18 to 2) 20 (9–43) <0FMPI – pain 0 (0–0) 4 (0–7) <0FMPI – QoL 0 (0–0) 1 (0–2) <0FMPI – total score 0 (�18 to 2) 25 (11–52) <0

DJD, degenerative joint disease; FMPI, feline musculoskeletal pain index.

In order to assess repeatability in normal and DJD affected catsand to assess the FMPI’s discriminatory ability, we carefully se-lected normal cats and cats we believed had DJD-associated painand mobility impairment. There is no validated method to deter-mine if a cat is ‘normal’ or has ‘DJD-associated pain and mobilityimpairment’. Our inclusion criteria were based on current clinicaljudgment. We have recently found that there is not a good correla-tion between radiographic DJD and pain on manipulation, (Las-celles et al., 2012) and so we required both to be evident in atleast one joint for a cat to be assigned to the ‘pain-DJD’ group. Inaddition, we required owner-assessed mobility impairment to bepresent.

Currently there is no way to objectively assess the level ofmobility impairment, and so we relied upon a subjective evalua-tion of the answers to questions regarding mobility and activity,as we have used in previous studies, (Lascelles et al., 2010a) andalso the CSOM system. It might be argued that selecting for catswhose owners consider that they have a mobility impairment isintroducing bias, as we might have been selecting owners whoare good at identifying pain or mobility impairment. This couldbe true to an extent, however all the cats that were not enrolledwere excluded due to other disease (e.g. neoplasia), or not meetingthe parameters of normality or DJD based on radiographic or phys-ical examination findings. We believe that the cohort of cats eval-uated here is reasonably representative of cat owners generally.However, the instrument should be tested on a randomly selected

groups.

value Pain-DJD (Day 1) Pain-DJD (Day 14) P value

375 20 (9–43) 24 (11–55) 0.01424 (0–7) 4 (0–8) 0.57711 (0–2) 2 (0–3) 0.0625

969 25 (11–52) 29 (12–66) 0.0141

ed due to high number of zero values. P values from Wilcoxon signed rank tests.

and pain-DJD groups at D1 and D14.

alue Normal (Day 14) Pain-DJD (Day 14) P value

.0001 0 (�17 to 4) 24 (11–55) <0.0001

.0001 0 (0–1) 4 (0–8) <0.0001

.0001 0 (0–1) 2 (0–3) <0.0001

.0001 0 (�17 to 4) 29 (12–66) <0.0001

Table 6Comparison of D1 and D14 individual FMPI question scores after controlling for age, median (range), in normal and pain-DJD groups.

Normal Day 1 Pain-DJD Day 1 P value Normal Day 14 Pain-DJD Day 14 P value

ActivityFMPI 1 0 (�1 to 0) 1 (0–3) <0.0001 0 (�1 to 0) 2 (0–3) <0.0001FMPI 2 0 (�1 to 0) 2 (0–3) <0.0001 0 (�1 to 0) 2 (0–4) <0.0001FMPI 3 0 (�1 to 0) 2 (0–4) <0.0001 0 (�1 to 1) 3 (1–4) <0.0001FMPI 4 0 (�1 to 1) 4 (1–4) <0.0001 0 (�1 to 1) 4 (2–4) <0.0001FMPI 5 0 (�1 to 0) 2 (0–4) <0.0001 0 (�1 to 1) 2 (0–4) <0.0001FMPI 6 0 (�1 to 0) 2 (0–3) 0.0004 0 (�1 to 1) 2 (0–3) <0.0001FMPI 7 0 (�1 to 0) 2 (0–3) 0.0009 0 (�1 to 1) 2 (0–3) 0.0002FMPI 8 0 (�1 to 0) 2 (0–4) 0.0004 0 (�1 to 1) 3 (0–4) <0.0001FMPI 9 0 (�1 to 0) 2 (0–4) 0.0016 0 (�1 to 1) 2 (0–4) 0.0003FMPI 10 0 (�1 to 0) 1 (0–3) 0.0005 0 (�1 to 1) 2 (0–3) 0.0002FMPI 11 0 (�1 to 1) 1 (0–2) 0.0061 0 (�1 to 1) 1 (0–3) 0.0056FMPI 12 0 (�1 to 0) 1 (0–4) 0.0003 0 (�1 to 1) 2 (0–4) 0.0002FMPI 13 0 (�1 to 0) 1 (0–3) 0.0013 0 (�1 to 1) 1 (0–3) 0.006FMPI 14 0 (�1 to 0) 0 (�1 to 2) 0.0908 0 (�1 to 1) 1 (0–3) 0.0174FMPI 15 0 (�1 to 0) 1 (�1 to 2) 0.0111 0 (�1 to 1) 1 (�1 to 2) 0.0409FMPI 16 0 (�1 to 0) 0 (�1 to 1) 0.3406 0 (�1 to 1) 2 (0–4) 0.9173FMPI 17 0 (�1 to 0) 0 (0–2) 0.0122 0 (�1 to 1) 0 (�1 to 3) 0.0327FMPI 18 0 (�1 to 0) 2 (1–3) <0.0001 0 (�1 to 1) 2 (1–3) <0.0001

PainFMPI 19 0 (0–0) 2 (0–3) <0.0001 0 (0–1) 2 (0–4) <0.0001FMPI 20 0 (0–0) 2 (0–4) <0.0001 0 (0–0) 2 (0–4) <0.0001

QoLFMPI 21 0 (0–0) 1 (0–2) <0.0001 0 (0–1) 2 (0–3) <0.0001

DJD, degenerative joint disease; FMPI, feline musculoskeletal pain index; QoL, quality of life.

Table 7FMPI scores, median (range) in each FMPI domain in pain-DJD cats, comparing scores obtained on D1 and D14 for different subgroups.

L-DJD (Day 1) H-DJD (Day 1) P value L-DJD (Day 14) H-DJD (Day 14) P value

FMPI – activity 19 (11–42) 23 (9–43) >0.05 21 (7–44) 29 (11–55) >0.05FMPI – pain 4 (0–5) 4 (2–7) >0.05 2 (0–6) 5 (3–8) >0.05FMPI – QoL 1 (0–2) 1 (0–2) >0.05 1 (0–2) 2 (1–3) >0.05FMPI – total score 24 (15–49) 29 (15–49) >0.05 24 (9–50) 36 (24–66) >0.05

DJD, degenerative joint disease; FMPI, feline musculoskeletal pain index; L-DJD, low degenerative joint disease; H-DJD, high degenerative joint disease.

372 J. Benito et al. / The Veterinary Journal 196 (2013) 368–373

population of cats to see if it can detect DJD-associated impair-ment. The problem is that there is currently no gold standard fordetermining DJD-associated impairment.

Recruitment of normal cats was difficult, largely because of thehigh prevalence of radiographic evidence of DJD in the feline pop-ulation (Lascelles et al., 2010b; Slingerland et al., 2011), and ourconcern that radiographic DJD might be associated with pain eventhough we could not elicit pain on manipulation. Several cats weredisqualified from the normal group because of this, despite theirowners not having recognized any mobility impairment. In es-sence, our strict criteria resulted in us comparing fairly extremegroups.

In comparing the individual questions between normal and DJDgroups, we did not adjust the P value for multiple comparisons. Ifwe had adjusted the P value (to 0.0024 using Bonferroni correc-tion), then questions 11, 14, 15, 16 and 17 would not have beensignificantly different between the groups. It is also these ques-tions (and in addition questions 10, 12, 13 and 16) that showedthe smallest difference in medians between the two groups. Fur-ther research should be directed at determining if these questionsare valid measures of DJD-associated pain. The questions about ac-tive activities and pain appeared to show the biggest differences inmedians between the groups. The questionnaire is heavily biasedtowards questions on activity, as we believe that musculoskeletalpain impacts activity. However, it is clear from other work we haveperformed on a QoL questionnaire (Benito et al., 2012) that catowners are focused as much on non-active behaviors as activebehaviors when considering their pet cat’s QoL, and so it might

well be that the contribution of the questions on overall pain andQoL are important contributors.

It is unknown if the FMPI can distinguish between different de-grees of impairment, or between normal cats and mild degrees ofimpairment. Indeed, in this study the FMPI was not able to distin-guish between low and high impairment (in the way we definedlow and high impairment). This could limit its clinical utility if de-grees of impairment cannot be distinguished. However, the divi-sion of cats into low and high impairment was arbitrary andthere were few cats in each group. Further studies need to be per-formed with larger numbers before conclusive statements aboutthe ability of the FMPI to distinguish between levels of impairmentcan be made. A better way to test this discriminatory ability wouldbe to define level of impairment using some objective measure andthen test the instrument. Due to our requirement for an absence ofradiographic DJD, the normal and DJD groups were not age-matched. Our analysis controlled for age, and because radiographicevidence of DJD is very common in older cats, this will likely be re-quired feature of any study comparing arthritis-affected and nor-mal cats.

The group size for this study was relatively small and thereforeit is uncertain how generalizable the results obtained here are tothe general cat population. There was also evidence of an increasein activity and total FMPI scores over the 2-week period in thepain-DJD group. This could have been because of worsening dis-ease (clinical signs associated with DJD might not be stable overtime), or, more likely, a result of owners evaluating their cats moreclosely and noticing a greater degree of impairment than initially

J. Benito et al. / The Veterinary Journal 196 (2013) 368–373 373

recognized. Such findings have implications for which data is usedas baseline data in studies evaluating feline behavior over time.

Conclusions

The FMPI showed suitable readability, had good internal consis-tency-reliability and good repeatability. The discriminatory abilityof this questionnaire was very good, but it did not distinguish be-tween disease subgroups containing small numbers of cats. Furtherstudies should be performed in order to evaluate responsivenessand criterion validity.

Conflict of interest statement

BDXL has received honoraria from Novartis Animal Health forsponsored continuing education programs, and has acted as a paidconsultant for Novartis Animal Health. BDXL serves on the SmallAnimal Scientific Advisory Board for Morris Animal Foundation(voluntary, unpaid). None of the other authors have any conflictof interest.

Acknowledgements

This study was funded by the Fellowship Research Program ofNovartis Animal Health (Global) and the Morris Animal Founda-tion. The lead author was funded by the Global Fellowship Pro-gram, Novartis Animal Health and the Morris Animal Foundation.The authors would like to thank Drs. B. Kirsch (The Cat Hospitalof Durham and Chapel Hill, Durham), A. Alley (Bowman AnimalHospital and Cat Clinic, Raleigh), B. Pondish (VCA Broad Street Ani-mal Hospital, Fuquay Varina), D.L. Holman (Gentle Care AnimalHospital, Raleigh) and K. Bush (Kate Bush’s Mobile Veterinary Prac-tice, Raleigh) in North Carolina, for their referral of cases for thesestudies. The authors are grateful for the assistance of Dr. MilaFreire in reviewing the manuscript.

References

Beadman, R., Smith, R.N., King, A.S., 1964. Vertebral osteophytes in the cat.Veterinary Record 76, 1005–1007.

Benito, J., Gruen, M.E., Thomson, A., Simpson, W., Lascelles, B.D., 2012. Owner-assessed indices of quality of life in cats and the relationship to the presence ofdegenerative joint disease. Journal of Feline Medicine and Surgery 14, 863–867.

Bennett, D., Morton, C., 2009. A study of owner observed behavioural and lifestylechanges in cats with musculoskeletal disease before and after analgesic therapy.Journal of Feline Medicine and Surgery 11, 997–1004.

Brown, D.C., Boston, R.C., Coyne, J.C., Farrar, J.T., 2007. Development andpsychometric testing of an instrument designed to measure chronic pain indogs with osteoarthritis. American Journal of Veterinary Research 68, 631–637.

Brown, D.C., Boston, R.C., Coyne, J.C., Farrar, J.T., 2008. Ability of the canine brief paininventory to detect response to treatment in dogs with osteoarthritis. Journal ofthe American Veterinary Medical Association 233, 1278–1283.

Clarke, S.P., Bennett, D., 2006. Feline osteoarthritis: A prospective study of 28 cases.Journal of Small Animal Practice 47, 439–445.

Clarke, S.P., Mellor, D., Clements, D.N., Gemmill, T., Farrell, M., Carmichael, S.,Bennett, D., 2005. Prevalence of radiographic signs of degenerative joint diseasein a hospital population of cats. Veterinary Record 157, 793–799.

Freire, M., Robertson, I., Bondell, H.D., Brown, J., Hash, J., Pease, A.P., Lascelles, B.D.,2011. Radiographic evaluation of feline appendicular degenerative joint diseasevs. macroscopic appearance of articular cartilage. Veterinary Radiology andUltrasound 52, 239–247.

Fry, E., 1968. A readability formula that saves time. Journal of Reading 11, 513–516.Godfrey, D.R., 2005. Osteoarthritis in cats: A retrospective radiological study.

Journal of Small Animal Practice 46, 425–429.Hardie, E.M., Roe, S.C., Martin, F.R., 2002. Radiographic evidence of degenerative

joint disease in geriatric cats: 100 cases (1994–1997). Journal of the AmericanVeterinary Medical Association 220, 628–632.

Hercock, C.A., Pinchbeck, G., Giejda, A., Clegg, P.D., Innes, J.F., 2009. Validation of aclient-based clinical metrology instrument for the evaluation of canine elbowosteoarthritis. Journal of Small Animal Practice 50, 266–271.

Hielm-Bjorkman, A.K., Kuusela, E., Liman, A., Markkola, A., Saarto, E., Huttunen, P.,Leppaluoto, J., Tulamo, R.M., Raekallio, M., 2003. Evaluation of methods forassessment of pain associated with chronic osteoarthritis in dogs. Journal of theAmerican Veterinary Medical Association 222, 1552–1558.

Hielm-Bjorkman, A.K., Rita, H., Tulamo, R.M., 2009. Psychometric testing of theHelsinki chronic pain index by completion of a questionnaire in Finnish byowners of dogs with chronic signs of pain caused by osteoarthritis. AmericanJournal of Veterinary Research 70, 727–734.

Kincaid, J.P., Fishburne, R.P., Rogers, R.L., Chissom, B.S., 1975. Derivation of newreadability formulas (automated readability index, FOG count, and Fleschreading ease formula) for navy enlisted personnel. In: Research Branch Report.Chief of Naval Technical Training, Memphis, TN, USA. http://library.ucf.edu/Web/purl.asp?dpid=DP0008946 (accessed 17 December 2012).

Koeppel, E., Ebner, J., 1990. Die Hueftgelenksdysplasie der Katze. Kleintierpraxis 35,281–298.

Langenbach, A., Green, P., Giger, U., Rhodes, H., Gregor, T.P., LaFond, E., Smith,G., 1998. Relationship between degenerative joint disease and hip jointlaxity by use of distraction index and Norberg angle measurement in agroup of cats. Journal of the American Veterinary Medical Association 213,1439–1443.

Lascelles, B.D., DePuy, V., Thomson, A., Hansen, B., Marcellin-Little, D.J., Biourge, V.,Bauer, J.E., 2010a. Evaluation of a therapeutic diet for feline degenerative jointdisease. Journal of Veterinary Internal Medicine 24, 487–495.

Lascelles, B.D., Dong, Y.H., Marcellin-Little, D.J., Thomson, A., Wheeler, S., Correa, M.,2012. Relationship of orthopedic examination, goniometric measurements, andradiographic signs of degenerative joint disease in cats. BMC VeterinaryResearch 8, 10.

Lascelles, B.D., Hansen, B.D., Roe, S., DePuy, V., Thomson, A., Pierce, C.C., Smith, E.S.,Rowinski, E., 2007. Evaluation of client-specific outcome measures and activitymonitoring to measure pain relief in cats with osteoarthritis. Journal ofVeterinary Internal Medicine 21, 410–416.

Lascelles, B.D., Henderson, A.J., Hackett, I.J., 2001. Evaluation of the clinical efficacyof meloxicam in cats with painful locomotor disorders. Journal of Small AnimalPractice 42, 587–593.

Lascelles, B.D., Henry 3rd, J.B., Brown, J., Robertson, I., Sumrell, A.T., Simpson, W.,Wheeler, S., Hansen, B.D., Zamprogno, H., et al.Freire, M., 2010b. Cross-sectionalstudy of the prevalence of radiographic degenerative joint disease indomesticated cats. Veterinary Surgery 39, 535–544.

McDowell, I., Newell, C., 1996. Pain measurements. In: Measuring Health: A Guideto Rating Scales and Questionnaires. Oxford University Press, New York, NY,USA, pp. 470–519.

McLaughlin, G.H., 1969. SMOG grading – A new readability formula. Journal ofReading 12, 639–646.

Slingerland, L.I., Hazewinkel, H.A., Meij, B.P., Picavet, P., Voorhout, G., 2011. Cross-sectional study of the prevalence and clinical features of osteoarthritis in 100cats. The Veterinary Journal 187, 304–309.

Streiner, D.L., Norman, G.R., 1995. Health Measurements Scales. A Practical Guide toTheir Development and Use. Oxford University Press, New York.

Sudman, S., Bradburn, N.M., 1982. Questionnaire from start to finish. In: Vinoth, J.(Ed.), Asking Questions. A Practical Guide to Questionnaire Design. Jossey-BassInc., San Francisco, CA, USA, pp. 281–286.

Zamprogno, H., Hansen, B.D., Bondell, H.D., Sumrell, A.T., Simpson, W., Robertson,I.D., Brown, J., Pease, A.P., Roe, S.C., et al., 2010. Item generation and designtesting of a questionnaire to assess degenerative joint disease-associated painin cats. American Journal of Veterinary Research 71, 1417–1424.