Accepted Manuscript Development and psychometric evaluation of the Musculoskeletal Pain Intensity and Interference Questionnaire for professional orchestra Musicians Patrice Berque, MSc, BSc (Hons), MMACP, MCSP Heather Gray, MSc, BSc, MCSP, Prof D Angus McFadyen, PhD, Statistical Consultant PII:

S1356-689X(14)00117-9

DOI:

10.1016/j.math.2014.05.015

Reference:

YMATH 1580

To appear in:

Manual Therapy

Received Date: 18 January 2014 Revised Date:

13 April 2014

Accepted Date: 30 May 2014

Please cite this article as: Berque P, Gray H, McFadyen A, Development and psychometric evaluation of the Musculoskeletal Pain Intensity and Interference Questionnaire for professional orchestra Musicians, Manual Therapy (2014), doi: 10.1016/j.math.2014.05.015. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Development and psychometric evaluation of the Musculoskeletal Pain Intensity and Interference Questionnaire for professional orchestra Musicians

RI PT

Patrice Berquea,b, MSc, BSc (Hons), MMACP, MCSP; Heather Graya,c, Prof D, MSc, BSc, MCSP; Angus McFadyend, PhD a

School of Health and Life Sciences, Glasgow Caledonian University, Glasgow, G4 0BA, Scotland, UK.

b

c

SC

Department of Physiotherapy, Glasgow Royal Infirmary, Glasgow, G31 2ER, Scotland, UK. Glasgow Dental School, University of Glasgow, Glasgow G3 3JZ.

d

M AN U

Statistical Consultant, AKM-STATS. Formerly Reader in Health Statistics, Glasgow Caledonian University, Glasgow, G4 0BA, Scotland, UK.

Lead and corresponding author: Patrice Berque

AC C

EP

TE D

Mr Patrice Berque Department of Physiotherapy Glasgow Royal Infirmary Alexandra Parade Glasgow G31 2ER Scotland UK

Tel: +44 1360 621896 Email: [email protected]

ACCEPTED MANUSCRIPT

Abstract Many epidemiological surveys on playing-related musculoskeletal disorders (PRMDs) have been conducted on professional musicians, but none have evaluated

RI PT

or confirmed the psychometric properties of the self-report instruments that were used. The aim of the present study was to develop and validate a self-report

instrument for professional orchestra musicians to measure musculoskeletal (MSK) pain and pain interference in terms of function and psychosocial constructs. 183

SC

professional orchestra musicians in Scotland were eligible to participate in the study,

M AN U

of which 101 (55% response rate) took part. Development of the Musculoskeletal Pain Intensity and Interference Questionnaire for Musicians (MPIIQM) involved the selection and modification of the most appropriate instruments measuring MSK pain, followed by psychometric evaluation of the new instrument. Face and content validity were ascertained by expert panels. 37 participants completed the

TE D

questionnaire. The percentage of missing scores was very low (2.7%). Exploratory factor analysis revealed that the MPIIQM had a strong and stable two-factor structure, with nine retained items explaining 71.3% of the variance in the data set.

EP

“Pain intensity” and “pain interference” were the two emerging factors. High internal consistency was achieved for each subscale (Cronbach’s alpha = 0.91). Substantial

AC C

test-retest reliability for the pain intensity items (range 0.78 – 0.82), and moderate to substantial test-retest reliability for the pain interference items (range 0.56 – 0.76) were obtained. The MPIIQM is a valid and reliable self-report instrument for the measurement and evaluation of MSK pain and pain interference in a population of professional orchestra musicians. Keywords: Musculoskeletal, Musicians, Psychometrics, Questionnaire.

ACCEPTED MANUSCRIPT

1 1.1

Introduction Background

The Performing Arts Medicine literature has grown substantially since 1980, and a

RI PT

definition of playing-related musculoskeletal disorders (PRMDs) affecting musicians was developed, with PRMDs defined as “pain, weakness, numbness, tingling, or other symptoms that interfere with (their) ability to play (their) instrument at the

SC

level (they) are accustomed to” (Zaza et al., 1998). Numerous prevalence studies on PRMDs affecting musicians have been carried out worldwide (Fishbein et al., 1988;

M AN U

Zaza & Farewell, 1997; Zetteberg et al., 1998; Yeung et al., 1999; Davies and Mangion, 2002; Engquist et al., 2004; Kaneko et al., 2005; Abreu-Ramos and Micheo, 2007; Ackermann and Driscoll, 2010; Leaver et al., 2011; Ackermann et al., 2012). The heterogeneity of these studies prohibits, however, a meta-analysis summary estimate of the overall prevalence. Two recent systematic reviews (Zaza et

TE D

al., 1998; Wu, 2007) highlighted the weaknesses of the studies, i.e. the lack of an operational definition of the observed outcome, low response rates, measurement bias, reporting errors and omissions, questionnaires that were not validated,

1.2

EP

inconsistent, poorly described, and deficient in collecting psychosocial factors. Lack of validated outcome measures

AC C

Measurement is central to evaluating many phenomena encountered in healthcare and epidemiology, and the quality of measurement instruments is therefore crucial. Instruments need to be valid, reliable, and responsive to change (de Vet et al., 2011). Very few of the recent prevalence studies mentioned in section 1.1 used existing validated instruments to measure musculoskeletal (MSK) pain and, if used, no attempt was made to evaluate or confirm the psychometric properties of these 1

ACCEPTED MANUSCRIPT

instruments in a population of musicians. Furthermore, none of the study authors made reference to using the biopsychosocial principles set out by the World Health Organisation (WHO) in the international classification of functioning, disability and

1.3

RI PT

health (ICF)(WHO, 2002). Literature review

1.3.1 Characteristics and psychometric properties of included instruments

SC

The aim of the literature review was to identify instruments measuring MSK pain

and pain interference which had been psychometrically tested and could be used and

M AN U

adapted for a population of professional orchestra musicians. Six instruments (Table 1) were retained following the literature search, and after applying inclusion/exclusion criteria (Supplementary file 1).

Each instrument was reviewed in detail and graded using levels of reliability, validity

TE D

and responsiveness for methodological quality assessment described by Terwee et al. (2007). Supplementary file 2 describes in detail the psychometric testing results that were reported for the selected instruments. Summary ratings of the psychometric

EP

testing results are compiled in Table 1, using the rating scales developed by McDowell (2006, p.7). These rating scales evaluate both the results and

AC C

thoroughness of reliability and validity testing. Table 2 summarises the characteristics of the instruments selected for this review. 1.3.2 Outcome of the literature review: the need for a new instrument

The literature review revealed that two existing instruments have been thoroughly psychometrically tested in numerous studies, i.e. the McGill Pain Questionnaire (LFMPQ) and its short-form (SF-MPQ), and the Brief Pain Inventory (BPI)(Table 1). Both instruments showed adequate construct validity, and were designed as 2

ACCEPTED MANUSCRIPT

instruments for evaluative purposes which could be used as outcome measures, and have been shown to have good responsiveness and test-retest reliability (Supplementary file 2). Neither of these instruments has, however, been validated in

RI PT

a population of musicians. Moreover, the LF-MPQ did not address biopsychosocial dimensions relating to pain interference with function, and distress relating from pain (Table 2), and the activity interference items of the BPI were not suitable for PRMDs

SC

in musicians due to potential floor effects (Table 2).

There were only two studies dealing specifically with instruments developed for

M AN U

musicians. The Musculoskeletal Load and Physical Health Questionnaire for Musicians was developed as part of a large-scale longitudinal five-year study focusing on the assessment of musculoskeletal pain and injury, and the identification of risk factors in professional orchestras in Australia (Ackermann & Driscoll, 2010; Ackermann et al., 2012). However, the questionnaire is extensive, takes at least 25

1).

TE D

minutes to complete, and its psychometric properties have not been evaluated (Table

EP

The Musculoskeletal Pain Questionnaire for Musicians (MPQM) was the first instrument developed and psychometrically tested to measure MSK pain in freelance

AC C

professional orchestra musicians (Lamontagne and Bélanger, 2012). The MPQM was tested using principal component analysis (PCA), and revealed a three-component structure, including 10 items (Supplementary file 2). The structure of this instrument (Table 2) should, however, be viewed with some caution. The authors accepted a final solution where one item loaded substantially on two components, and this item should have been considered for deletion (de Vet et al., 2011; De Vellis, 2012). It could be argued that deletion of the item could have greatly affected the component 3

ACCEPTED MANUSCRIPT

structure of the MPQM and, therefore, its construct validity. Secondly, the MPQM did not include items measuring the impact of pain on psychosocial and affective variables, and was not designed to gather prevalence data on PRMDs and pain

1.4

RI PT

location (Table 2). Study aims

The aim of the study was to develop and validate, for a population of professional

SC

orchestra musicians, a new biopsychosocial self-report instrument collecting

demographical data, prevalence of PRMDs and pain location, and measuring

-

M AN U

musculoskeletal pain and pain interference. This study involved two phases: Phase 1: Development of the new instrument, involving modification and adaptation of those selected following the literature review. -

Phase 2: Psychometric evaluation of the new instrument, including face and

TE D

content validity, construct validity, internal consistency, and test-retest

AC C

EP

reliability.

4

ACCEPTED MANUSCRIPT

2 2.1

Methods Phase 1

2.1.1 Ethics

RI PT

The study was approved by the Research Ethics Committee of the School of Health and Life Sciences at Glasgow Caledonian University. Written authorisation to use and modify the BPI and the DASH questionnaires was obtained.

SC

2.1.2 Subjects and inclusion criteria

Of the four professional classical orchestras in Scotland, three agreed to take part in

M AN U

the study. Only the permanent members of those three orchestras were eligible to participate (n=183). Freelance players, who may have very variable timetables and workloads, and may perform a wider type of repertoire, were excluded (Chan et al., 2000).

TE D

2.1.3 Criteria for the new instrument

Important criteria were defined for the new instrument. It should: be short and not take more than 15 minutes to complete; be specific to the population of professional

EP

orchestra musicians, especially with regard to their perceived impairment with workrelated functional activity, i.e. playing their musical instrument; follow the

AC C

international guidelines set out by WHO in the ICF; and have evaluative qualities, i.e. the ability to measure change over time, as this could be used to measure changes in health status following treatment interventions. The instrument was named the Musculoskeletal Pain Intensity and Interference Questionnaire for Musicians (MPIIQM). 2.1.4 Structure and content of the MPIIQM 5

ACCEPTED MANUSCRIPT

The questionnaire was designed to gather information on the following: demographical data on age, gender, and practice habits, prevalence of musculoskeletal pain and problems, pain location, pain frequency and duration, pain

RI PT

intensity, pain affective interference, and pain activity interference. Each topic/construct was designed by using and sometimes modifying items from previously evaluated existing instruments.

SC

The draft version of the MPIIQM is reproduced in Table 3 and details the 26 items used. When extraction or modification took place, the sources of the existing

M AN U

instrument are given. Questionnaire items are grouped according to the topics or constructs to be measured and indicate reference to the appropriate ICF component. Three “levels of functioning” are described within the ICF: “functioning” refers to all body functions, activities and participation; while “disability” refers to impairments, activity limitations, and participation restrictions. The ICF also includes external

TE D

environmental and personal factors (WHO, 2002). Since the BPI addressed dimensions not covered by the LF-MPQ, i.e. pain

EP

interference with function, and distress resulting from pain (Table 2), it was chosen as the reference instrument to measure the following core constructs of the new

AC C

instrument: pain intensity and affective interference (Table 3). The version of the optional performing arts and sports module of the Disabilities of the Arm, Shoulder, and Hand (DASH) questionnaire (Hudak et al., 1996), modified for the MPQM (Lamontagne and Bélanger, 2012), was used for the activity interference items. An 11-point numerical rating scale (NRS) was preferred to the 4point Likert scale used in the MPQM in order to improve responsiveness to change 6

ACCEPTED MANUSCRIPT

(Kirshner and Guyatt, 1985; Hjermstad et al., 2011), and to be consistent with the scales used in the BPI. 2.1.5 Content validity

RI PT

Guidelines from the COnsensus-based Standards for the selection of health

Measurement INstruments (COSMIN) checklist were followed (Terwee et al., 2007; Mokkink et al., 2010). A panel of four experts, with experience in the field of

SC

psychometrics, pain management, neurological and pain syndromes affecting

musicians, was chosen by the main researcher. Experts were asked to comment on

M AN U

item relevance by rating each as either “essential”, or “useful but not essential”, or “not necessary”, and content validity ratios (CVR) were then calculated (Lawshe, 1975). Experts also commented on item comprehensiveness, presence of ambiguous or confusing items, clarity, conciseness, and wording of the items (de Vet et al.,

TE D

2011; De Vellis, 2012). 2.1.6 Pilot testing

A sample of three professional musicians who were not part of the study sample was

EP

asked to comment on comprehensibility and relevance of each questionnaire item by rating each item as “easy to understand” or not, “relevant” or not. Respondent burden

AC C

(in minutes) was also evaluated. 2.1.7 Face validity Comments from the expert panel and the three musicians were used to address the concept of face validity (de Vet et al., 2011). 2.2

Phase 2

2.2.1 Data collection 7

ACCEPTED MANUSCRIPT

The MPIIQM was modified jointly by the authors following content validity and pilot testing. Questionnaire data were collected from participants during rehearsals or meetings organised with each of the three orchestras. Return of the questionnaire to

RI PT

the researcher (PB) was deemed evidence of consent, in order to respect anonymity. 2.2.2 Missing scores

The percentage of missing scores was evaluated (Mokkink et al., 2010), since

SC

missing scores may point to various problems regarding the formulation of some

2.2.3 Construct validity

M AN U

items. Less than 3% is considered acceptable (de Vet et al., 2011).

Construct validity aims to determine the dimensionality and internal structure of an instrument, i.e. how many groups of variables, or constructs, underlie a set of items, and to reduce its size by deleting items which do not clearly contribute to a construct

TE D

(de Vet et al., 2011).

Exploratory factor analysis (EFA) is considered by many as preferable to principal component analysis (PCA) (Floyd and Widaman, 1995; Fabrigar et al., 1999;

EP

Costello and Osborne, 2005). EFA attempts to represent only the shared variance (common variance) of each variable, not the total variance, contrary to PCA. EFA is,

AC C

therefore, a modelling process relying on assumptions, which derives a mathematical model from which common factors are estimated. PCA, in contrast, is mainly a data reduction method (De Vellis, 2012). Exploratory factor analysis (EFA), with principal axis factoring (PAF), was used in the study (Field, 2011). Guidelines regarding sample size indicate a subject-to-item ratio of 5:1 (Floyd and Widaman, 1995; De Vellis, 2012). In the present study, the 8

ACCEPTED MANUSCRIPT

EFA on the 14 items measured by NRS would require a minimum of 70 subjects. The Kaiser-Meyer-Olkin measure of sampling adequacy (KMO) and the Bartlett’s test of sphericity were run to check that the data were appropriate for EFA. KMO

RI PT

values greater than 0.5 indicate that patterns of correlation are relatively compact and that EFA should yield distinct and reliable factors for the sample size. A significant Bartlett’s test (p < 0.05) indicates that the inter-item correlation matrix is

SC

significantly different from an identity matrix with very low correlations between variables (Field, 2011). For item reduction, the cut-off for significance of factor

M AN U

loading was set to 0.4 (Floyd and Widaman, 1995). Factor loading represents the correlation between an item and a factor, and the square of the factor loading represents a measure of the substantive importance of a particular item to a factor, i.e. the percentage of the variance of an item that is explained by a factor (Field, 2011).

TE D

2.2.4 Internal consistency

Internal consistency, measured by Cronbach’s alpha, was calculated for each subscale separately (Mokkink et al., 2010). Cronbach’s alpha values between 0.70

EP

and 0.90 are considered adequate (Streiner and Norman, 2003). Item-to-total correlations and changes to alpha when an item is deleted were considered (de Vet et

AC C

al., 2011).

2.2.5 Test-retest reliability An intra-class correlation (ICC) Model (2,1) with 95% confidence intervals (CI) was chosen to assess test-retest reliability for the questionnaire items retained following PAF analysis (Fleiss and Cohen, 1973). The time period between measurements was five days, to prevent recall bias (Streiner and Norman, 2003). A minimum sample 9

ACCEPTED MANUSCRIPT

size of 50 subjects was recommended (de Vet et al., 2011), therefore, musicians from all three orchestras were handed a second questionnaire to be filled in five days after the first, and to be returned to the researcher.

RI PT

2.2.6 Statistical analysis

SPSS software version 19 (SPSS Inc., Chicago, Illinois, USA) was used for all

AC C

EP

TE D

M AN U

SC

analyses. All tests were performed using a 5% level of significance (α = 0.05).

10

ACCEPTED MANUSCRIPT

3 3.1

Results Face validity, content validity, and respondent burden

Members of the experts’ and professional musicians’ panels who reviewed the

RI PT

relevance, comprehensiveness, and clarity of the MPIIQM agreed unanimously that face validity was present. All three musicians completed the questionnaire in less than 10 minutes. Of the 14 core items measured by NRS, three did not reach the

SC

minimum agreement of at least half the experts considering the item as “essential”

(CVR = 0). These items were “relations with other people” (item 20), “sleep” (item

M AN U

21), and “playing your instrument as well as you would like” (item 25)(Table 3). These items were identified as being possibly problematic and were considered during EFA for potential deletion. 3.2

Changes made to the MPIIQM

The original instrument (Table 3) was modified following comments from experts

TE D

and musicians. The NRS for the items “duration of pain” and “frequency of pain” was replaced by a 10cm visual analogue scale (VAS), since an 11-point NRS (0 to 10) was considered confusing to express a 12-month time period. One demographic

EP

question was added to ascertain the proportion of full-time versus part-time musicians (item 4, Fig. 1), since this may impact on the reported weekly number of

AC C

hours of playing. Minor wording changes were made to improve clarity (item 6 and statements prior to items 13, 14, 18)(Fig. 1). 3.3

Participants’ characteristics

Of the 183 eligible professional orchestra musicians, 101 completed the MPIIQM (55% response rate), with almost an equal split between males (50.5%) and females (49.5%). The mean age of participants was 47.7 ± 10.4 (mean ± SD) years (range 2511

ACCEPTED MANUSCRIPT

65 years). On average, musicians had been playing professionally in an orchestra for 23.5 ± 11.1 (mean ± SD) years. 37 participants completed the MPIIQM after question 12 (Fig. 1). Missing scores

RI PT

3.4

For the 14 items of the MPIIQM measured by NRS and VAS, two items were

omitted by one participant, i.e. 2.7% missing scores. For demographic items (n =

SC

101), only two questions had missing entries, i.e. 1.9% and 2.9% missing scores

respectively. For prevalence items, there were four completion errors, representing a

3.5

M AN U

3.9% error rate. Exploratory factor analysis

PAF analyses were conducted, starting with the 14 items measured by VAS and NRS, with data from participants who reported current prevalence of PRMDs (n =

TE D

37). The process was iterative, i.e. a new PAF analysis was re-run after each item deletion. Factor rotations were run in the following order: no rotation, orthogonal rotation (varimax), oblique rotation (direct oblimin). Oblique rotations consistently

EP

yielded the best solutions. The KMO measure of sampling adequacy for each PAF analysis was in an acceptable range (values > 0.753). Bartlett’s tests of sphericity

AC C

was significant for each iteration (p < 0.001), indicating a justifiable factoring solution (Field, 2011). Several items were questionable and deleted in turn from the analyses. The items “duration of pain” and “spending your usual amount of time playing” consistently had factor loadings under 0.4 and low communalities, and were consequently deleted. The third deletion was for “frequency of pain” due to cross-loading onto two

12

ACCEPTED MANUSCRIPT

factors in all models. The item “sleep” was identified as correlating with the pain intensity items rather than the interference items. Since this item had a CVR value of 0, it was deleted and another iteration of the PAF analysis was conducted. Two

RI PT

factors had eigenvalues ≥ 1, and a two-factor solution with 10 items emerged with factors named “pain intensity” for Factor 1 and “pain interference” for Factor 2. All 10 items had acceptable communalities after extraction (range 0.446 – 0.915). The

SC

two-factor solution with 10 items explained 68.7% of the variation.

Following the poor test-retest reliability of the item “relations with other people”

M AN U

(section 3.7 below), another PAF analysis was run, retaining only the 9 items with significant test-retest reliability. All 9 items had substantial communalities after extraction (range 0.614 – 0.928). Table 4 shows the factor loadings after oblique rotation (direct oblimin). A two-factor solution emerged, with 9 items explaining 71.3% of the variance, which was higher than the 10-item solution. The final two-

3.6

TE D

factor and 9-item version of the MPIIQM is reproduced in Figure 1 (items 14-22). Internal consistency

EP

Cronbach’s alpha values for the four pain intensity items (Factor 1) and for the five pain interference items (Factor 2) were both 0.91. Any further item deletion within a

AC C

subscale did not improve these values. Cronbach’s alpha value for the overall 9-item scale was 0.88. 3.7

Test-retest reliability

Data from 19 participants were available for test-retest reliability. Table 5 summarises the results for the 10 items retained following preliminary PAF analyses. Results from the ICC Model (2,1) showed substantial test-retest reliability for the

13

ACCEPTED MANUSCRIPT

four pain intensity items (Factor 1), with ICC values ranging between 0.78 and 0.82, and a set of fairly narrow 95% CI. For the pain interference items (Factor 2), the ICC Model (2,1) revealed moderate to substantial test-retest reliability for five out of six

RI PT

items, with values ranging between 0.56 and 0.76, and a set of perhaps wider than desired, but reasonable 95% CI. The item “relations with other people” showed poor

AC C

EP

TE D

M AN U

SC

reliability with an ICC value of 0.13, and was not statistically significant.

14

ACCEPTED MANUSCRIPT

4 4.1

Discussion Conceptual justification of the MPIIQM

The results from the PAF analyses demonstrated that the MPIIQM had a strong two-

RI PT

factor and 9-item structure (Table 4) and also addresses several of the themes set out in the biopsychosocial ICF, i.e. structure/function or impairment, activity limitation, participation restriction, and personal factors, in keeping with the aim to measure

SC

health and disability within a biopsychosocial model (WHO, 2002).

The four pain intensity items were from the BPI (impairment). The pain interference

M AN U

factor comprised two items from the BPI, representing the impact of pain on psychosocial variables (personal factors and participation restriction), and three items from the optional performing arts and sports module of the DASH questionnaire, representing the impact of pain on function (activity limitation and participation

4.2

TE D

restriction)(Fig. 1).

Comparison of the MPIIQM to the MPQM and BPI instruments

The MPIIQM shares some similarities with the MPQM in terms of constructs, and

EP

uses the same activity interference items as the MPQM, although slightly modified (Fig. 1). The MPQM and the MPIIQM have low respondent burdens, being

AC C

completed in less than 10 minutes. However, there are several differences between the two instruments. The MPIIQM showed good construct validity, with a strong and stable two-factor structure similar to the factor structure of the BPI, in keeping with psychometric studies of the BPI carried out in large cohorts of MSK pain patients (Keller et al., 2004; Mendoza et al., 2006)(Supplementary file 2). A 9-item solution was preferred, with substantial 15

ACCEPTED MANUSCRIPT

communalities, and high factor loadings without cross-loadings (Table 4), thereby showing “strong” data in terms of EFA (Costello & Osborne, 2005). This is in contrast with what was obtained for the MPQM (section 1.3.2). Furthermore, there is

COSMIN checklist (Mokkink et al., 2010).

RI PT

no indication that the MPQM was developed following the guidelines set out in the

Internal consistency of the MPIIQM, with Cronbach’s alpha values of 0.91 for each

SC

subscale, is in keeping with values obtained by studies investigating the BPI

(Supplementary file 2), and shows better overall scale homogeneity than the MPQM,

M AN U

with a Cronbach’s alpha value of 0.88, compared to 0.77 for the MPQM. The MPIIQM development included test-retest reliability, the MPQM did not. 4.3

Strengths and Limitations

4.3.1 Strengths

TE D

The MPIIQM is the first instrument of its kind, validated specifically for a population of professional orchestra musicians and designed to gather epidemiological data on PRMDs. The process followed guidelines from the

EP

COSMIN checklist during development and psychometric testing, and guidelines from WHO – ICF, encompassing a biopsychosocial model. It also provided

AC C

confirmation regarding construct validity of the optional performing arts and sports module of the DASH questionnaire. However, there were several limitations. 4.3.2 Limitations due to the sample Firstly, the study was targeting professional classical orchestra musicians, limiting generalisability to other groups (de Vet et al., 2011). Secondly, the sample size available for EFA and internal consistency was smaller (n = 37) than the 16

ACCEPTED MANUSCRIPT

recommended size (n = 70), which may have affected the stability of the factor pattern (De Vellis, 2012), and the results of the present study should therefore be treated with some caution.

RI PT

There is, however, considerable controversy about what the “optimum” sample size should be for EFA, and strict rules regarding sample size have mostly disappeared (Costello & Osborne, 2005). Sample size is partly determined by the nature of the

SC

data, i.e. the magnitude of the factor loadings and communalities. Several authors

have indeed agreed that when the communalities are high (average value ≥ 0.70), and

M AN U

the factor loadings are high (values ≥ 0.80), solutions were highly stable across samples with as few as 50 participants (Floyd and Widaman, 1995; Fabrigar et al., 1999; Costello and Osborne, 2005; Field, 2011; De Vellis, 2012). In the present study, the average of the communalities after factor extraction in the 9-item solution was 0.71, and all factor loadings were high, with 7 out of 9 equal or higher than 0.80

TE D

(Table 4), thereby confirming the strong structure and stability of the MPIIQM. The high values obtained for the KMO measure of sampling adequacy (values ≥ 0.753) for each PAF analysis confirmed this trend. The sample size of the present study

EP

may, therefore, not be as limiting as previously thought.

AC C

4.3.3 Other aspects of reliability and validity Firstly, with regard to test-retest reliability, the sample available was smaller (n = 19) than desired (n = 50), which may partly explain the poor reliability score obtained for the item “relations with people” in the 10-item solution. Secondly, there are other important aspects of psychometrics which could be tested in future studies to strengthen the overall validity and evaluative characteristics of 17

ACCEPTED MANUSCRIPT

the MPIIQM: criterion validity, convergent validity, responsiveness, and interpretability. 4.4

Recommendations

RI PT

The MPIIQM could be used for longitudinal epidemiological studies exploring the

prevalence and incidence of PRMDs. Additionally, due to its evaluative properties, it could be used as an outcome measure in clinical practice or intervention studies

SC

dealing with MSK pain and pain interference in professional orchestra musicians.

The instrument could also be used as injury surveillance tool within the context of

M AN U

health and safety in professional orchestras. Future studies are required to confirm its validity and reliability on a larger sample of orchestra players, and with other groups

AC C

EP

TE D

of musicians.

18

ACCEPTED MANUSCRIPT

5

Conclusions

The MPIIQM is a promising instrument with robust psychometric properties that can be used to gather epidemiological data on PRMDs, and to measure MSK pain

RI PT

intensity and pain interference in a population of professional orchestra musicians in clinical and research settings. It has a short completion time of less than 10 minutes; face and content validity; good construct validity with a strong two-factor structure

SC

compliant with the WHO-ICF biopsychosocial themes; and is reliable, thereby

confirming its properties as a potential evaluative instrument capable of measuring

Competing interests

M AN U

change over time.

AC C

EP

TE D

The authors declare that they have no competing interests.

19

ACCEPTED MANUSCRIPT

References Abreu-Ramos AM, Micheo WF. Lifetime prevalence of upper-body musculoskeletal problems in a professional-level symphony orchestra: age, gender, and instrumentspecific results. Medical Problems of Performing Artists 2007; 22(3): 97-104.

RI PT

Ackermann B, Driscoll T, Kenny DT. Musculoskeletal pain and injury in professional orchestral musicians in Australia. Medical Problems of Performing Artists 2012; 27(4): 181-87.

SC

Ackermann B, Driscoll T. Development of a new instrument for measuring the musculoskeletal load and physical health of professional orchestral musicians. Medical Problems of Performing Artists 2010; 25(3): 95-101.

M AN U

Atkinson TM, Rosenfeld BD, Sit L, Mendoza TR, Fruscione M, Lavene D, Shaw M, Li Y, Hay J, Cleeland CS, Scher HI, Breitbart WS, Basch E. Using confirmatory factor analysis to evaluate construct validity of the Brief Pain Inventory (BPI). Journal of Pain & Symptom Management 2011; 41(3): 558-65. Baron S, Hales T, Hurrell J. Evaluation of symptom surveys for occupational musculoskeletal disorders. American Journal of Industrial Medicine 1996; 29: 60917. Berque P, Gray H. The influence of neck-shoulder pain on trapezius muscle activity among professional violin and viola players: an electromyographic study. Medical Problems of Performing Artists 2002; 17(2): 68-75.

TE D

Chan RFM, Chow C, Lee GPS, To L, Tsang XYS, Yeung SS, Yeung EW. Selfperceived exertion level and objective evaluation of neuromuscular fatigue in a training session of orchestral violin players. Applied Ergonomics 2000; 31: 335-41.

EP

Cleeland CS, Ryan KM. Pain assessment: global use of the Brief Pain Inventory. Annals Academy of Medicine Singapore 1994; 23(2): 129-38.

AC C

Cleeland CS, Ladinsky JL, Serlin RC, Thuy NC. Multidimensional measurement of cancer pain: comparisons of US and Vietnamese patients. Journal of Pain and Symptom Management 1988; 3(1): 23-7. Costello AB, Osborne JW. Best practices in exploratory factor analysis: recommendations for getting the most from your analysis. Practical Assessment, Research & Evaluation 2005; 10(7): 1-9. Daut RL, Cleeland CS, Flanery RC. Development of the Wisconsin Brief Pain Questionnaire to assess pain in cancer and other diseases. Pain 1983; 17: 197-210. Davies J, Mangion S. Predictors of pain and other musculoskeletal symptoms among professional instrumental musicians: elucidating specific effects. Medical Problems of Performing Artists 2002; 17(4): 155-68. 20

ACCEPTED MANUSCRIPT

Dawson AP, Steele EJ, Hodges PW, Stewart S. Development and test-retest reliability of an extended version of the Nordic Musculoskeletal Questionnaire (NMQ-E): a screening instrument for musculoskeletal pain. Journal of Pain 2009; 10(5): 517-26.

RI PT

De Vellis RF. Scale development: theory and applications. 3rd ed. London: Sage; 2012. p. 1-205. De Vet HCW, Terwee CB, Mokkink LB, Knol DL. Measurement in medicine: a practical guide. Cambridge: Cambridge University Press; 2011. p. 1-338.

SC

Dickinson CE, Campion K, Foster AF, Newman SJ, O’Rourke AM, Thomas P.G. Questionnaire development: an examination of the Nordic Musculoskeletal Questionnaire. Applied Ergonomics 1992; 23(3): 197-201.

Elliott AM, Smith BH, Smith WC, Chambers WA. Changes in chronic pain severity over time: the Chronic Pain Grade as a valid measure. Pain 2000; 88(3): 303-8.

M AN U

Engquist K, Ǿrbaek P, Jakobsson K. Musculoskeletal pain and impact on performance in orchestra musicians and actors. Medical Problems of Performing Artists 2004; 19(2): 55-61. Fabrigar LR, Wegener DT, MacCallum RC, Strahan EJ. Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods 1999; 4(3): 272-99. Field A. Discovering statistics using SPSS. 3rd ed. London: Sage; 2011. p. 1- 821.

TE D

Fishbein M, Middlestadt SE, Ottati V, Straus S, Ellis A. Medical problems among ICSOM musicians: overview of a national survey. Medical Problems of Performing Artists 1988; 3(1): 1-8.

EP

Fleiss JL, Cohen J. The equivalence of weighted kappa and the intraclass correlation coefficient as measures of reliability. Educational and Psychological Measurement 1973; 33: 613-9.

AC C

Floyd FJ, Widaman KF. Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment 1995; 7(3): 286-99. Gabel CP, Melloh M, Yelland M, Burkett B, Roiko A. Predictive ability of a modified Örebro Musculoskeletal Pain Questionnaire in an acute/subacute low back pain working population. European Spine Journal 2011; 20(3): 449-57. Hjermstad MJ, Fayers PM, Haugen DF, Caraceni A, Hanks GW, Loge JH, Fainsinger R, Aass N, Kaasa S. Studies comparing rating scales, verbal rating scales, and visual analogue scales for assessment of pain intensity in adults: a systematic review. Journal of Pain and Symptom Management 2011; 41(6): 1073-93.

21

ACCEPTED MANUSCRIPT

Holroyd KA, Holm JE, Keefe FJ, Turner JA, Bradley LA, Murphy WD, Johnson P, Anderson K, Hinkle AL, O’Malley WB. A multi-center evaluation of the McGill Pain Questionnaire: results from more than 1700 chronic pain patients. Pain 1992; 48(3): 301-11.

RI PT

Howe TE, Dawson LJ, Gray H, Dean S, Foster NE, Quin J, Syme G,Van der Windt DA. Recommendations for a Scottish Minimum Dataset of Outcome Measures for community based musculoskeletal rehabilitation, consensus event: musculoskeletal rehabilitation. Glasgow: HealthQWest; 12 May 2010.

Hudak PL, Amadio PC, Bombardier C. Development of an upper extremity outcome measure: the DASH (disabilities of the arm, shoulder, and hand). American Journal of Industrial Medicine 1996; 29(6): 602-8.

SC

Kaneko Y, Lianza S, Dawson WJ. Pain as an incapacitating factor in symphony orchestra musicians in São Paulo, Brazil. Medical Problems of Performing Artists 2005; 20(4): 168-74.

M AN U

Keller S, Bann CM, Dodd SL, Schein J, Mendoza TR, Cleeland CS. Validity of the brief pain inventory for use in documenting the outcomes of patients with non-cancer pain. Clinical Journal of Pain 2004; 20(5): 309-18. Kirshner B, Guyatt G. A methodological framework for assessing health indices. Journal of Chronic Disorders 1985; 38(1): 27-36.

TE D

Krebs EE, Bair MJ, Damush TM, Tu W, Wu J, Kroenke K. Comparative responsiveness of pain outcome measures among primary care patients with musculoskeletal pain. Medical care 2010; 48(11): 1007-14. Kuorinka I, Jonsson B, Kilborn A, Vinterberg H, Biering-Sørensen F, Andersson G, Jørgensen K. Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms. Applied Ergonomics 1987; 18(3): 233-7.

EP

Lamontagne V, Bélanger C. Development and validation of a questionnaire on musculoskeletal pain in musicians. Medical Problems of Performing Artists 2012; 27(1): 37-42.

AC C

Lawshe CH. A quantitative approach to content validity. Personnel Psychology 1975; 28: 563-75. Leaver R, Harris EC, Palmer KT. Musculoskeletal pain in elite professional musicians from British symphony orchestras. Occupational Medicine 2011; 61: 54955. Linton SJ, Nicholas M, MacDonald S. Development of a short form of the Örebro Musculoskeletal Pain Screening Questionnaire. Spine 2011; 36(22): 1891-5. Linton SJ, Boersma K. Early identification of patients at risk of developing a persistent back problem: the predictive validity of the Örebro Musculoskeletal Pain Questionnaire. Clinical Journal of Pain 2003; 19(2): 80-6. 22

ACCEPTED MANUSCRIPT

McDowell I. Measuring health: a guide to rating scales and questionnaires. 3rd ed. New York: Oxford University Press; 2006. p. 1-748. Melzack R. The short-form McGill Pain Questionnaire. Pain 1987; 30(2): 191-7.

RI PT

Melzack, R. The McGill Pain Questionnaire: major properties and scoring methods. Pain 1975; 1(3): 277-99. Mendoza T, Mayne T, Rublee D, Cleeland C. Reliability and validity of a modified Brief Pain Inventory short form in patients with osteoarthritis. European Journal of Pain 2006; 10(4): 353-61.

SC

Mendoza TR, Chen C, Brugger A, Hubbard R, Snabes M, Palmer SN, Zhang Q, Cleeland CS. The utility and validity of the modified brief pain inventory in a multiple-dose postoperative analgesic trial. Clinical Journal of Pain 2004; 20(5): 35762.

M AN U

Menezes Costa L da C, Maher CG, McAuley JH, Hancock MJ, de Melo Oliveira W, Azevedo DC, Freitas Pozzi LM, Pereira AR, Costa LO. The Brazilian-Portuguese versions of the McGill Pain Questionnaire were reproducible, valid, and responsive in patients with musculoskeletal pain. Journal of clinical epidemiology 2011; 64(8): 903-12. Mokkink LB, Terwee CB, Patrick DL, Alonso J, Stratford PW, Knol DL, Bouter LM, de Vet HCW. The COSMIN checklist for assessing the methodological quality of studies on measurement properties of health status measurement instruments: an international Delphi study. Quality of Life Research 2010; 19: 539‐49.

TE D

Ohlsson K, Attewell RG, Johnsson B, Ahlm A, Skerfving S. An assessment of neck and upper extremity disorders by questionnaire and clinical examination. Ergonomics 1994; 37(5): 891-7.

EP

Salaffi F, Stancati A, Grassi W. Reliability and validity of the Italian version of the Chronic Pain Grade questionnaire in patients with musculoskeletal disorders. Clinical rheumatology 2006; 25(5): 619-31.

AC C

Smith BH, Penny KI, Purves AM, Munro C, Wilson B, Grimshaw J, Chambers WA, Smith WC. The Chronic Pain Grade questionnaire: validation and reliability in postal research. Pain 1997; 71(2): 141-7. Streiner DL, Norman GR. Health measurement scales: a practical guide to their development and use. 3rd ed. Oxford: Oxford University Press; 2003. p. 1-283. Terwee CB, Bot SDM, de Boer MR, van der Windt DAWM, Knol DL, Dekker J, Bouter LM, de Vet HCW. Quality criteria were proposed for measurement properties of health status questionnaires. Journal of Clinical Epidemiology 2007; 60: 34-42. Von Korff M, Ormel J, Keefe FJ, Dworkin SF. Grading the severity of chronic pain. Pain 1992; 50(2): 133-49. 23

ACCEPTED MANUSCRIPT

World Health Organisation (WHO). Towards a common language for functioning, disability and health: the international classification of functioning, disability and health. Geneva: World Health Organisation; 2002. Wu SJ. Occupational risk factors for musculoskeletal disorders in musicians. Medical Problems of Performing Artists 2007; 22(2): 43-51.

RI PT

Yeung E, Chan W, Pan F, Sau P, Tsui M, Yu B, Zaza C. A survey of playing-related musculoskeletal problems among professional orchestral musicians in Hong Kong. Medical Problems of Performing Artists 1999; 14(2): 43-7.

SC

Zaza C. Playing-related musculoskeletal disorders in musicians: a systematic review of incidence and prevalence. Canadian Medical Association Journal 1998; 158(8): 1019-25. Zaza C, Charles C, Muszynski A. The meaning of playing-related musculoskeletal disorders to classical musicians. Social Science and Medicine 1998; 47(12): 2013-23.

M AN U

Zaza C, Farewell VT. Musicians’ playing-related musculoskeletal disorders: an examination of risk factors. American Journal of Industrial Medicine 1997; 21: 292300.

AC C

EP

TE D

Zetteberg C, Backlund H, Karlsson J, Werner H, Olsson L. Musculoskeletal problems among male and female music students. Medical Problems of Performing Artists 1998; 13(4): 160-8.

24

RI PT

ACCEPTED MANUSCRIPT

Table 1 Summary ratings of psychometric testing for selected instruments (adapted from McDowell, 2006, p.7) Instrument

Reliability Thoroughness

MSK Load and Physical Health Questionnaire for Musicians Ackermann and Driscoll (2010)

0

Musculoskeletal Pain Questionnaire for Musicians (MPQM) Lamontagne and Bélanger (2012)

*

Reliability Results

Validity Results

0

0

0

**

*

*

*

**

*

**

Nordic Musculoskeletal Questionnaire (NMQ) Kuorinka et al. (1987), Dickinson et al. (1992), Ohlsson et al. (1994), Baron et al. (1996)

*

*

*

*

NMQ – E (Extended version) Dawson et al. (2009)

*

**

0

0

**

**

**

**

**

**

**

**

M AN U

TE D

Chronic Pain Grade Questionnaire (CPGQ) Von Korff et al. (1992), Smith et al. (1997), Elliot et al. (2000), Salaffi et al. (2006)

EP

McGill Pain Questionnaire: LF-MPQ (Long-Form) and SF-MPQ (Short-Form) Melzack (1975), Melzack (1987), Holroyd et al. (1992), McDowell (2006), Menezes Costa et al. (2011)

AC C

Brief Pain Inventory (BPI) Cleeland et al. (1988), Keller et al. (2004), Mendoza et al. (2004), Mendoza et al. (2006), McDowell (2006), Krebs et al. (2010), Atkinson et al. (2011)

Ratings for thoroughness of reliability and validity testing: 0 No reported evidence of reliability or validity * Basic information only; information only by the original authors of the scale ** Several types of test, and several studies by different authors, have reported reliability or validity *** All major forms of reliability or validity testing reported in numerous studies

SC

Validity Thoroughness

Ratings for the results of the reliability and validity testing: 0 No numerical results reported * The evidence suggests weak reliability or validity ** Adequate reliability or validity *** Excellent reliability or validity: higher coefficients than those normally seen in other instruments

25

Table 2 Characteristics of selected instruments

RI PT

ACCEPTED MANUSCRIPT

Purpose

Summary / Description

Scoring Method

Cut-off Scores

MPQM Musculoskeletal (MSK) Pain Questionnaire for Musicians Lamontagne and Bélanger (2012)

Evaluative. To develop and validate a tool that evaluates MSK pain in musicians.

10-item self-report questionnaire specific to the measurement of PRMDs. 3 components included: pain intensity (4 items); disability associated with playing-related MSK pain (4 items); frequency (1 item), and duration of pain (1 item).

No indication given on how to obtain the overall score. Likert scales not given in article text.

N/A

CPGQ Chronic Pain Grade Questionnaire Von Korff et al. (1992)

Evaluative and predictive. To develop and validate a graded classification of the global severity of chronic pain.

7-item self-report questionnaire dealing with 3 anatomical sites (back/headache/TMD). 3 scale components: pain intensity (3 items), disability score (3 items), disability days (1 item).

Pain intensity score: converted to a 0-100 range.

Cut-off scores were determined by the goodness of fit of the Guttman scales.

Disability points (0-6 scale): sum of disability days (number of days converted to 0-3 scale) and disability score (0-100, converted to 0-3 scale).

Classification: Grade 0: pain free. Grade I: low disability (