Disability & Rehabilitation, 2012; Early Online: 1–9 © 2012 Informa UK, Ltd. ISSN 0963-8288 print/ISSN 1464-5165 online DOI: 10.3109/09638288.2012.702849
RESEARCH PAPER
I nternal consistency and construct validity of the Revised Illness Perception Questionnaire adapted for work disability following a musculoskeletal disorder Valérie Albert1, Marie-France Coutu2,3 & Marie-José Durand2,3 Disabil Rehabil Downloaded from informahealthcare.com by 67.68.212.164 on 08/14/12 For personal use only.
1
Department of Clinical Sciences, Faculty of Medicine and Health Sciences, Université de Sherbrooke, Longueuil, Canada, School of Rehabilitation, Faculty of Medicine and Health Sciences, Université de Sherbrooke, Longueuil, Canada, and 3Centre for Action in Work Disability Prevention and Rehabilitation, Hôpital Charles-LeMoyne Research Center, Longueuil, Canada
2
Purpose: To assess internal consistency and construct validity of the French version of the Revised Illness Perception Questionnaire adapted for Work Disability (IPQR-WD). Method: A cross sectional study was conducted in rehabilitation centers and private clinics in the Montreal region of Canada, involving 43 men and women, French speaking, absent from work between 3 months and a year due to musculoskeletal disorders. The 9 dimension IPQR-WD and the following eightrelated instruments for construct validity were administered: Tampa Scale for Kinesiophobia (TSK), Pain Catastrophizing Scale (PCS), Psychological Distress Index (PDI-14), Pain Disability Index (PDI), Self-Efficacy for Return to Work Scale (SERWS), Pain Beliefs and Perceptions Inventory (PBPI), Implicit Models of Illness Questionnaire (IMIQ) and a Visual Analog Scale for pain intensity (VAS). Results: Calculations of Cronbach’s α (from 0.58 to 0.87) revealed satisfactory internal consistency of the IPQRWD dimensions. Multiple regression analyses were performed with each IPQR-WD dimension and significant independent variables. Final models explained a good proportion of the variance (adjusted r2 = 0.33–0.70) for each dimension, except for the Cyclical timeline dimension for which associations became non-significant after adjusting for gender, age and length of sick leave. The SERWS and the VAS were not significantly associated to any of the IPQR-WD dimensions. Conclusions: Moderate to strong correlations were found with six-related instruments which support the multidimensional nature of the IPQR-WD and its unique contribution as one simple questionnaire that can assess representations related to work disability.
Implications for Rehabilitation • Certain unhelpful beliefs related to the worker’s own understanding (i.e. representation) of his/her current health condition may lead to behaviours that are poorly adapted to reducing the work disability following a musculoskeletal disorder. • The Revised Illness Perception Questionnaire adapted for Work Disability (IPQR-WD) is the only questionnaire available in French that can specifically assess the worker’s representation following a musculoskeletal disorder. • The use of this questionnaire will facilitate a systematic evaluation of the worker’s representation according to a biopsychosocial approach, which may contribute to a clinical practice based on the latest scientific evidences available in the work disability field.
Numerous psychosocial factors have been associated with work disability following musculoskeletal disorders (MSD). Recent systematic reviews indicate that fear of movement [1], recovery expectations [2,3], perception of incapacity [4], pain catastrophizing [5] and psychological distress [6] are among the psychological factors most commonly cited. However, a comprehensive understanding of the mechanisms by which these factors are interrelated for a particular worker in order to influence their work disability situation is still lacking. One possible avenue to reveal those links is the study of health and illness representations, which refer to how individuals understand their own health problem [7].
Keywords: musculoskeletal disorder, psychometrics, Revised Illness Perception Questionnaire adapted for Work Disability (IPQR-WD), representation, work
Correspondence: Valérie Albert, CAPRIT – Université de Sherbrooke – Campus de Longueuil, 150, Place Charles LeMoyne 9e étage, Longueuil (Quebec) Canada J4K 0A8. Tel.: (450) 463-1835 ext. 61881. Fax: (450) 463-6593. E-mail:
[email protected] (Accepted June 2012)
1
Disabil Rehabil Downloaded from informahealthcare.com by 67.68.212.164 on 08/14/12 For personal use only.
2 Valérie Albert et al. The Common Sense Model of self-regulation of health and illness (CSM) exposes how illness representation influences the strategies adopted by the individual to control or cure the health problem [8]. First, information gathered about the illness is organized together to generate the illness representation. Based on this comprehension, an action plan is elaborated and criteria are set to evaluate if the selected strategies are effective to control or cure the illness. The effect of these strategies gives feedback to the individual that may modify his/her representation according to this new information. The dynamic character of this process contributes to the individual ‘making sense of ’ his/her condition, which is named illness coherence [9]. According to the CSM, the representation has an emotional and a cognitive aspect with a mutual influence on one another, and both have an impact on the strategies selected by the individual to improve the health problem [8]. The emotional representation is generated to manage negative emotions related to the illness and if these emotions, for example fear or helplessness, appear too threatening to the individual, the strategies might only be directed towards distress management, rather than solving the health problem [10]. The cognitive representation is organised in five core components which consist of beliefs regarding: (1) symptoms and diagnosis (Identity), (2) elements that provoked illness (Causes), (3) negative effects related to illness (Consequences), (4) expected duration of illness (Timeline) and (5) effectiveness of treatment to recover from illness or possibilities of having personal control on illness (cure/control) [8]. Although these five components appear consistent among different cultures or health problems [8,11], beliefs related to each representation component may vary greatly from one individual to another and the representation is not always in accordance with biomedical reality [12]. More specifically, recent research in the work disability field showed that certain erroneous components of the worker’s representation may lead to behaviours that are poorly adapted to reducing the work disability [13]. The systematic evaluation of the worker’s representation should thus be part of an evidence-based practice in work disability management [13]. Several tools have been developed to assess the patient’s illness representation, but they are mainly found in a medical context rather than in a disability prevention paradigm [14]. Furthermore, qualitative studies among workers on sick leave related to persistent musculoskeletal pain revealed that a number of workers did not define themselves as being “ill”, but rather in an “intermediate state” between illness and health [15,16]. Questionnaires assessing “illness” representation must therefore be adapted prior to their use among this clientele. Interestingly, a critical review of assessment tools indicated that the Revised Illness Perception Questionnaire (IPQ-R) [17] appeared noteworthy because of its factorial structure that is highly similar to the CSM and its superior psychometric properties [14]. All the subscales demonstrated good internal consistency with Cronbach’s α coefficients ranging from 0.79 to 0.89. Test–retest reliability was assessed over a 3-week period and the dimensions generally showed good stability, with Pearson correlations ranging from 0.46 to 0.88.
Known group validity was assessed by comparing the illness beliefs of the acute and chronic pain patients. The groups were significantly different on all dimensions, thus indicating good construct validity [17]. A French translation of the original IPQ-R can be found on the Web (http://www.uib.no/ipq/pdf/ IPQ-R-French.pdf) and its psychometric properties suggest the equivalence of the translated version [18]. The IPQ-R has also been studied in medical contexts among patients with an MSD [19–21]. For these reasons, the IPQ-R was selected for an adaptation [22]. A validation study was subsequently needed to evaluate internal consistency and construct validity of the adapted questionnaire, named the Revised Illness Perception Questionnaire for Work Disability (IPQR-WD).
Methods Study design A cross-sectional study was conducted from December 2010 to July 2011 in the region of Montreal, Canada. The project was approved by the Research Ethics Committee of the Centre de Recherche Interdisciplinaire en Réadaptation du Montréal Métropolitain on 8 June 2010 and by Hôpital CharlesLeMoyne on 27 July 2010. Participants The target population is composed of workers on sick leave due to persistent pain from an MSD and participating in a work rehabilitation program. Inclusion criteria for the participants consisted of: being aged between 18 and 64 years, having between 3 months and a year of absence from work related to an MSD and being admitted in a rehabilitation program. Workers unable to understand and/or read French, suffering from an MSD related to a specific pathology, or presenting major psychiatric problems as indicated in their medical record were excluded. Sample size was calculated for a multiple regression analysis by an a priori power analysis performed with the G*Power 3.1.2 software [23]. In order to identify at least 3 significant associations (r2 > 0.35) with a power value of 80% and an α level of 5%, a minimum of 36 participants were required. Procedures Participants were recruited in rehabilitation centers and private clinics by convenience sampling. Those who agreed and signed the consent form were then asked to complete the adapted questionnaire, a sociodemographic questionnaire and eight-related instruments assessing separate components of the representation, which took approximately one hour. Representation of current health condition With the authors’ permission, the original French version of the IPQ-R was adapted for work disability using a group consensus technique [24]. The group was composed of eight experts, clinicians and researchers, all in the work disability field. As a result of the adaptation, eight items from the original IPQ-R have been removed due to a lack of relevance to work disability following an MSD: five items belonging to the Identity dimension (sore throat, breathlessness, sore eyes, wheeziness, upset stomach)
Disability & Rehabilitation
Disabil Rehabil Downloaded from informahealthcare.com by 67.68.212.164 on 08/14/12 For personal use only.
Work Disability Representations Assessment 3 and three to the Causes dimension (a germ or virus, pollution in the environment, altered immunity) [22]. Furthermore, the term “illness” was replaced by “current health condition” and 26 items were added in order to assess factors associated with work disability according to a biopsychosocial approach [22]. In order to assess item clarity, a qualitative pre-test was performed following the adaptation process among a sample of workers from the target population. Pretest results confirmed item clarity [22]. The Revised Illness Perception Questionnaire adapted for Work Disability (IPQR-WD) thus contains 88 items divided among the nine original dimensions of the IPQ-R: Identity, Causes, Consequences, Personal Control, Treatment Control, Acute/Chronic Timeline, Cyclical Timeline, Coherence and Emotional Representation. The new items, which have been carefully attributed to a particular dimension by the experts, are presented in Table I. As in the IPQ-R, items are rated on a 5-point Likert scale (strongly disagree to strongly agree), except for the Identity dimension which symptoms require a yes or no answer. A score is calculated for each dimension, except for the Causes dimension which is not an actual scale [17].
Variables and instruments assessing separate components of the representation The following variables (and instruments) were selected to assess the construct validity of the IPQR-WD: fear of movement (Tampa Scale for Kinesiophobia – TSK) [25–27]; pain catastrophizing (Pain Catastrophizing Scale – PCS) [28,29]; psychological distress (Psychological Distress Index 14 items – PDI-14) [30–32]; perception of incapacity (Pain Disability Index – PDI) [33–35]; self-efficacy with regard to work capacity (Self-Efficacy for Return to Work Scale – SERWS) [36,37]; pain intensity (Visual Analog Scale – VAS) [38–40]; mystery and pain duration (Pain Beliefs and Perceptions Inventory – PBPI) [41–43] and illness representation (Implicit Models of Illness Questionnaire – IMIQ) [44–47]. These instruments assessing separate components of the representation were selected according to their superior psychometric properties and their availability in a validated French version. An overview of the variables and the psychometric properties of the selected instruments is presented in Table II. Psychometric criteria (good, acceptable, OK, marginal) were set according to a critical review of psychosocial hazard measures [48]. As well, the correlations expected with the IPQR-WD dimensions are presented in the last column of Table II. Statistical analyses Statistical analyses were performed with the SPSS software version 17. Frequencies and central tendency measures were calculated to describe the sample’s profile. Furthermore, frequencies to which causes and symptoms were endorsed by participants were calculated in order to assess the validity of the range of items included in the Causes and Identity dimensions, as well as the relevance of the new items, as reported by Moss-Morris et al. [17] to assess validity of the original IPQ-R. Internal consistency was estimated by the calculation of Cronbach’s α for each dimension, with the exception of the Identity and Causes dimensions, which were analysed © 2012 Informa UK, Ltd.
Table I. IPQR-WD items added to the IPQ-R following the adaptation process. Dimension (Total number of items in dimension) New items Identity (16) Changes in physical sensations Decrease of physical capacity Mood changes Numbness Attention/memory problems Lack of endurance Fears Causes (20) Anterior physical condition Work environment Fate/destiny Another’s error or negligence Delays in healthcare system Acute/chronic timeline (7) I have no idea when my current health condition will end Consequences (9) My current health condition affects the way I see myself My current health condition is making me dependent My current health condition affects my social life (family, work, relationships) Personal control (8) I have no idea of what I should do to control my current health condition I can learn ways/strategies to improve my current health condition Treatment control (8) The effect of my treatment will vary according to how I participate in it My treatment can help me recover an ability despite the persistence of my symptoms The treatment will decrease the length of my actual health condition Coherence (7) I have contradictory information about my current health condition I don’t understand how my injury can be healed and remain painful Cyclical timeline (4) No item was added to this dimension Emotional representation (9) I feel punished by the pain My current health condition is a significant stress in my life My current health condition makes me feel old (dependent)
differently, with descriptive statistics. Furthermore, in order to judge the conceptual equivalence between the IPQR-WD and the original IPQ-R, a statistical test for independent samples was performed to compare the α coefficients obtained for this study to those obtained from a previous study on the IPQ-R French version [18], using the following formula [49]:
W=
(1 − α 2 ) (1 − α1 )
The values of αj are the α coefficients obtained from both studies, the larger coefficient being the numerator and W being tested as an F statistic with degrees of freedom of N2−1 and N1–1 [49]. In order to determine whether the new items were pertinent to the scales they were assigned to through the adaptation process, three methods were used. First, the variation of α coefficient if each item was removed was analysed, in order to identify if any item would deteriorate the scale’s reliability. Second, the item-dimension correlations were examined, where values superior to the threshold (0.30) were expected [50]. Finally, a T test for dependant samples was performed to compare α coefficients of each dimension of the IPQR-WD with and without the new items added through the adaptation process, with the following formula [51]: ( N − 2) ( α 1 − α 2 ) 2 t= 4(1 − α1 ) (1 − α 2 ) (1 − r12 )
1/ 2
The value of r12 is the correlation coefficient between α1 and α2 and the t statistic is evaluated with N−2 degrees of freedom [51]. To assess convergent validity, a Pearson correlation matrix was first obtained to identify the significant bivariate associations between dimensions of the IPQR-WD and the selected instruments measuring separate components of the representation (independent variables). Second, multiple regression analyses using a stepwise method were performed on each dimension with potentially confounding demographic variables and the significant independent variables.
Results
CFA, confirmatory factor analysis; EFA, exploratory factor analysis.
Variable (instrument) Fear of movement (TSK) Pain catastrop-hizing (PCS) Psychological distress (PDI-14) Perception of incapacity (PDI) Self-efficacy with regard to work capacity (SERWS) Pain intensity (VAS) Mystery (PBPI) and Pain duration (PBPI) Illness representation (IMIQ)
Psychometric properties (Rating: Good, Acceptable, OK, Marginal) Reliability Validity Internal Responsiveness Factor Expected correlations with Description Test–retest consistency to change analysis Concurrent IPQR-WD dimensions 11 items rated on a 4-point Likert scale assessing fear of movement/(re)injury Acceptable Acceptable Good [26] Good [26] Emotional representation and [25,26] [26,27] treatment control 13 items rated on a 5-point frequency scale (0 = not at all; 4 = all the time) describing Acceptable Acceptable Good [29] Emotional representation and thoughts and feelings that individuals may experience when they are in pain [28,29] [28,29] coherence 14 items using a 4-point frequency scale (1 = never; 4 = very often) which assess negative Acceptable Good [32] Emotional representation, reactions to stress (depressive and anxious symptoms, anger and attention deficits) [31] identity and causes 7 items assessing the extent to which pain interferes with daily activities, on Acceptable Good [35] (CFA) OK Consequences a scale from 0 (no disability) to 10 (total disability) [33,34] [33,34] 8 items rated on a scale ranging from 0 to 100% which assess the worker’s level of Acceptable (EFA) OK Personal Control confidence in his/her capacities to reintegrate work despite certain obstacles [36,37] [37] 10 cm line marked at each end with labels (no pain; worst pain imaginable) on which Acceptable Good [40] Good [39] Identity the participant is asked to indicate the point representing his/her pain intensity [39] 16 items rated on a 4-point Likert scale assessing beliefs regarding pain, distributed Acceptable (EFA) OK Coherence, A/C and among 3 dimensions: (1) Mystery, (2) Pain duration and (3) Self-blame [41,42] [42] cyclical timeline 45 items rated on a 5-point Likert scale assessing illness representation based on the CSM Marginal (CFA) OK Consequences, personal conand divided in 4 dimensions : consequences, responsibility, curability and variability [47] [44] trol and treatment control
Table II. Variables and instruments selected for the construct validity.
Disabil Rehabil Downloaded from informahealthcare.com by 67.68.212.164 on 08/14/12 For personal use only.
4 Valérie Albert et al.
Descriptive statistics Among workers who met the inclusion criteria, a total of 46 participants were approached by a research assistant and offered the opportunity to participate in this study. However, three workers refused to participate because of a lack of interest in the research. Therefore, complete data were obtained for 43 participants. The sample presented diverse clinical profiles and consisted of 23 women and 20 men, with an average age of 41 and an average duration of work absence of 8 months, as presented in Table III. Descriptive results for each dimension of the IPQR-WD are presented in Table IV. For the Causes dimension which is not an actual scale, 1 point was attributed to each Cause rated 4 (agree) or 5 (strongly agree), in order to identify the average number of Causes endorsed by the participants. Internal consistency As shown in Table V, Cronbach’s α coefficients calculated for each dimension indicate satisfactory internal consistency for each dimension, except for Cyclical Timeline (0.58), which contains only four items. Indeed, α coefficients for the
Disability & Rehabilitation
Disabil Rehabil Downloaded from informahealthcare.com by 67.68.212.164 on 08/14/12 For personal use only.
Work Disability Representations Assessment 5 Table III. Characteristics of participants (n = 43). Categorical variables Frequency (%) Gender Female 23 (53.5) Marital status Single 19 (44.2) In a relationship 24 (55.8) Educational attainment Secondary or inferior 19 (44.2) Post-secondary 24 (55.8) Type of work Manual 30 (69.8) Non-manual 5 (11.6) Mixed 8 (18.6) Site of injury Back/neck 21 (48.8) Upper extremity 4 (9.3) Lower extremity 5 (11.6) Multiple sites 13 (30.2) Previous attempt to return to work No 30 (69.8) Contestation No 39 (90.7) Claim Agent Public automobile insurance 19 (44.2) Worker’s compensation board 19 (44.2) Other/None 5 (11.7) Continuous variables* (units) Average (SD) [range] Age (years) (n = 41) 41 (12) [19;62] Length of sick leave (days) 237 (102) [86;436] VAS for pain intensity (/10) 5.2 (2.1) [0;8.3] PCS score (/52) 19 (12) [1;42] PDI score (/70) 33 (13) [8;54] SERWS score (/100) 27.6 (21) [0;81] TSK score (/44) 29 (6) [17;40] PDI-14 score (/100) 36 (20) [0;79] PBPI score (/60) 32 (6) [20;45] IMIQ score (/225)
116 (17) [76;144]
*All continuous variables are normally distributed (Kolmogorov-Smirnov Test > 0.05).
remaining dimensions were superior to the 0.60 threshold identified by Stafford and Bodson [52] and the majority were between 0.70 and 0.90, indicating good to excellent internal consistency among the items within the scale [53]. Furthermore, statistical comparisons [49] with the coefficients obtained for the French version of the original IPQ-R [18] indicated that IPQR-WD coefficients were comparable or superior, except for the Cyclical Timeline dimension, thus suggesting the conceptual equivalence of the adapted version [54]. All new items appeared pertinent to better assess the different dimensions which form the representation related to work disability. Indeed, variation of the Cronbach’s α if item is deleted showed no coefficient superior to the scale’s and item-dimension correlations were all superior to the 0.30 threshold identified by Field [50], suggesting that the new items effectively belong to their attributed dimension. © 2012 Informa UK, Ltd.
Table IV. Descriptive results for each dimension of the IPQR-WD (n = 43). Percentiles Dimension [Range 50th of possible scores] Min. Max. Mean* SD 25th (Median) Identity [0–16] 2 16 11.0 3.7 8 11 Acute/Chronic 8 29 19.7 4.8 17 20 Timeline [7–35] Consequences 14 40 29.9 5.9 26 30 [9–45] Personal Control 17 40 30.1 4.6 27 30 [8–40] Treatment Control 23 40 31.9 4.5 29 32 [8–40] Coherence [7–35] 11 35 24.0 5.7 20 24 6 20 14.2 2.8 13 14 Cyclical Timeline [4–20] 11 45 28.6 8.0 24 29 Emotional Representation [9–45] Causes [0–20] 0 12 4.9 2.8 3 4
75th 14 24 34 33 36 28 16 33
6
*Scores for each dimension are normally distributed (Kolmogorov-Smirnov Test > 0.05).
Table V. Internal consistency of the IPQR-WD dimensions compared to the original French version of the IPQ-R (al Anbar et al., 2010). IPQR-WD α IPQ-R α coefficient coefficient (NB items) (NB items) [n = 88] W p [n = 43] Dimension Acute/Chronic Timeline 0.81 (7) 0.69 (6) 0.61 0.04 Consequences 0.77 (9) 0.71 (6) 0.80 0.21 Personal Control 0.68 (8) 0.73 (6) 0.85 0.26 Treatment Control 0.77 (8) 0.62 (5) 0.61 0.04 Coherence 0.83 (7) 0.81 (5) 0.88 0.33 Cyclical Timeline 0.58 (4) 0.73 (4) 0.64 0.04 Emotional Representation 0.87 (9) 0.73 (6) 0.47 0.003
Table VI. Internal consistency of the IPQR-WD dimensions compared to the IPQR-WD dimensions without the new items (n = 43). IPQ-WD Dimension IPQR-WD without new items (α2) (number of new items) (α1) Acute/Chronic Timeline (1) Consequences (3) Personal Control (2) Treatment Control (3) Coherence (2) Cyclical Timeline (None) Emotional Representation (3)
r12
T
p
0.49
0.62
0.81
0.81
0.98
0.77 0.68 0.77 0.83 0.58
0.59 0.59 0.73 0.80 0.58
0.95 7.97 >0.000 0.97 4.73 >0.000 0.95 2.73 0.03 0.96 3.21 0.002 1.00 N/A N/A
0.87
0.81
0.97
7.88 >0.000
Moreover, statistical comparisons [51] presented in Table VI indicated that with the exception of the Acute/Chronic Timeline dimension, all remaining IPQR-WD dimensions showed a statistically superior Cronbach’s α coefficient when the new items were included. These results indicate that for a population of workers on sick leave due to MSD, the
6 Valérie Albert et al. IPQR-WD appears to better assess their representation, compared to the original IPQ-R.
Disabil Rehabil Downloaded from informahealthcare.com by 67.68.212.164 on 08/14/12 For personal use only.
Construct validity Validity of the identity and causes dimensions As reported by Moss-Morris et al. [17], validity of the Identity dimension was assessed by frequency calculations of symptoms declared by participants. Results presented in Table VII show that the seven symptoms added following the adaptation process were experienced by 50–95% of the sample, supporting their pertinence to the Identity dimension intended for a population presenting MSDs. Similarly, frequency calculations for the Causes dimensions revealed that the five new causes were also endorsed by a significant proportion of the sample, supporting their validity for a population in a situation of work disability following an MSD. Convergent validity of each dimension of the IPQR-WD Moderate to strong correlations were found for each dimension with six theoretically-related variables, as shown in Table VIII, suggesting good construct validity. The only exception is the Cyclical Timeline dimension, which bivariate associations with the IMIQ Curability dimension (r = 0.39; p = 0.009) and IMIQ-Variability (r = 0.36; p = 0.02) became Table VII. Endorsement frequencies of items added to the IPQR-WD Identity and Causes dimensions (n = 43). Identity Frequency (%) Causes Frequency (%) Changes in 26 (61) Anterior physical 12 (28) physical sensations condition Decrease of 41 (95) Work environment 18 (42) physical capacity Mood changes 29 (67) Fate/destiny 13 (30) Numbness 31 (72) Another’s error 14 (33) or negligence Attention / 22 (51) Delays in 5 (12) memory problems healthcare system Lack of endurance 41 (95) Fears 29 (67)
non-significant after adjusting for gender, age and length of sick leave. Furthermore, two of the selected variables were not significantly associated to any of the IPQR-WD dimensions following the regression analyses: pain intensity (VAS) and self-efficacy with regard to work capacity (SERWS). Indeed, bivariate analyses revealed that pain intensity (EVA) was correlated to the Acute/Chronic Timeline (r = 0.34; p = 0.02), Personal Control (r = −0.32; p = 0.04) and Coherence (r = −0.42; p = 0.005) dimensions, but these correlations became non significant when other independent variables were taken into account. Interestingly, the correlation matrix showed that self-efficacy with regard to work capacity (SERWS) was not related to the Personal Control dimension (r = 0.09; p = 0.58) as expected, but rather to the Identity (r = −0.40; p = 0.008) and Consequences (r = −0.35; p = 0.02) dimensions, and these correlations finally became non significant when other independent variables were taken into account.
Discussion This study presents for the first time certain psychometric properties of the IPQR-WD, which is a theory-based tool. Internal consistency is satisfactory and convergent validity analyses revealed a correlation structure that generally supported the hypothesized pattern, which contributes to the evidence of good construct validity [55]. The Cyclical Timeline dimension was the only one that was not significantly associated to any independent variable. However, this might be the result of a lack of statistical power, since the sample size was determined in order to identify strong associations (r2 > 0.35). Results from the construct validity hold interesting clinical implications in the work disability field. For the Identity dimension, results support that a high number of symptoms endorsed by participants is more strongly associated to psychological distress (PDI-14) than to pain intensity (EVA). Moreover, although an association between pain and psychological distress among people with chronic musculoskeletal pain has already been described [56,57], these two variables
Table VIII. Convergent validity of the IPQR-WD with related instruments (n = 43). Variable(s) included in the final Dimension Adjusted r2 (p) regression model* (instrument) Coherence 0.70 (
