Translation and validation of the Norwegian version of the Injustice Experience Questionnaire

Design

This study has a mixed method, cross sectional and explorative design. A forward-back-translation method was used to translate and culturally adapt the IEQ in to Norwegian. Individual cognitive debriefing interviews and a focus group interview were used to investigate patients’ understanding and experiences of the questionnaire. Registry data was used to investigate correlations with related health phenomena and factor structure of the IEQ.

Translation of questionnaire

The IEQ was translated using a modified version of the forward-back-translation, cultural adaptation and linguistic validation method as described by the ISPOR Task Force [24]. The translation was performed in a 10-step manner. Three investigators first discussed the meaning of content and concepts in the questionnaire with the author (M. Sullivan). Then three of the investigators (SER, L-PG, TML) separately translated the English version into Norwegian. After a thorough discussion of translational challenges, a reconciled Norwegian version of the IEQ was created. The Norwegian IEQ was then back-translated from Norwegian to English by a professional translator (with experience in health science document translation). The back-translated version was discussed in the research team, with the translator, and with the author in order to detect and correct misunderstandings and cultural/linguistic indifferences. Then an adjusted version of the Norwegian IEQ was created and tested in cognitive debriefing patient interviews (see below for an outline of these interviews). After adjustments based on the interviews and discussion among the researchers and the author, a final Norwegian version of the IEQ was created, proof read and made ready for trial in the local Oslo University Hospital Pain Registry (OPR) at the department of pain management at Oslo University Hospital. For more detailed information, see ReKS modified translation method for questionnaires: https://oslo-universitetssykehus.no/seksjon/Regional-kompetansetjeneste-for-smerte/Documents/ReKS%20modifiserte%20oversettelsesprosedyre%20for%20sp%C3%B8rreskjema%2018sept2015.pdf.

Cognitive debriefing interviews

The understanding of the IEQ was tested in individual cognitive debriefing interviews with seven patients with chronic pain conditions and experience of injustice. The interviews were conducted by two of the researchers (SER, TML), and took place at the pain clinic in 2015. The interviewers used a detailed cognitive debriefing guide based on discussions in the research team and the MAPI Research Institute procedure. The patients were asked about their understanding of the introduction, items, response options and layout, as well as whether the questionnaire was relevant to them, easy/difficult to understand, and evoked any emotions (positive/negative). Information from the interviews was collected as written notes and quotes.

Focus group interview

The understanding of perceived injustice was investigated in a focus group interview with nine patients with chronic pain conditions and previous experience of injustice. The interview was conducted by two of the researchers (HSB, TML), and took place at the pain clinic in 2017. The interviewers used the same detailed cognitive debriefing guide as described for the individual interviews. The focus group interview was sound-recorded, and transcribed verbatim.

Pain registry data

Data for the statistical analyses on sociodemographic and clinical information, injustice, pain catastrophizing, pain severity (i.e., intensity, bothersomeness) disability, psychological distress, and quality of life was retrieved from the OPR [25]. Data was retrieved from 3,068 patients in the period from March 2017 to February 2019.

Sample and recruitment

Psychologists at the pain clinic recruited the seven patients for the individual cognitive debriefing interviews and nine patients for the focus group interview. They chose patients who they believed would provide rich narratives and data (purposive sample). Inclusion criteria were age over 18 years, a persistent pain condition, fluency in spoken Norwegian, and experience with injustice related to their pain condition. Exclusion criteria were cognitive, affective and social problems prohibiting them from participating in an individual or group interview.

The 3,068 patients who provided data on the IEQ was recruited to the OPR. This sample and recruitment procedure is described in a previous article [25].

Measures

Ethics

The project was submitted to the Norwegian Regional Ethical Committee (2016/1942, REK sør-øst). As the study did not fall under the Health Research Act [40], no further ethical approval was required. The head of the department for data collection, and the data protection officer at Oslo University Hospital, approved of the cognitive debriefing interviews and the focus group interview. The OPR board and the local data protection officer also approved of the registry analyses of the study. All parts of the study adhered to the Helsinki declaration, and written and signed informed consent was provided from each participant in the qualitative interviews.

Analyses

The data from the individual and focus group interviews was examined by summarizing and comparing the patients’ reflections according to the already predefined areas of the cognitive debriefing interview guide.

The registry data was analyzed using SPSS for windows (version 25). Cronbach’s alfa was used to test scale reliability, and the internal consistency of the items in the IEQ. A level above 0.7 represents an acceptable internal consistency [41].

Factor analysis (principal component analysis) was used to test the construct validity of the IEQ-N. We sought to replicate the two factor structure (e.g., severity/irreparability of loss and blame/unfairness) that was presented in the original publication of the IEQ [2].

Pearson correlations were used to test construct validity of the IEQ-N against related health constructs. We hypothesized that more perceived injustice would correlate with more severe pain catastrophizing, pain intensity and bothersomeness, pain related disability, psychological distress as well as poorer quality of life. An alpha level of 0.01 was considered statistically significant. Cohen’s criteria was used to determine the strength of correlation between means (i.e., small=0.2; medium=0.5; large=0.8) [42].

Three separate hierarchical regression models were used to investigate the unique contribution of perceived injustice to the prediction of pain intensity, pain disability and psychological distress (dependent factors) beyond the variance explained by pain catastrophizing. In each regression model, age and gender were entered in the first step, followed by pain catastrophizing in the second step, and perceived injustice in the third step.

Sociodemographic and clinical information

Patients who participated in the cognitive debriefing interviews had a mean age of 47 years, 86% were women, and the majority had high school education (42.9%). Patients who participated in the focus group interview had a mean age of 50 years, and the majority (67%) were women. Their pain condition had lasted for a mean of 7.9 years. Most patients had high school (33.3%) or more than 4 years of university education (33.3%), were not working (66.6%), and lived with spouse/children (66.6%) (Table 1).

Patients in the pain registry had a mean age of 50 years, had their pain condition for a mean of 8.4 years, and a majority (57.9%) were women. Most patients had high school (42.9%) or university education shorter than 4 years (30.9%), were not working (64.4%), and lived alone (68.9%) (Table 1). The IEQ total score mean of 22.4 did not exceed the cut-off for clinically significant experiences of injustice. In general, patients reported strong pain intensity (mean 7.2), a high degree of pain bothersomeness (mean 7.7), considerable pain disability (mean 42.7) and psychological distress (mean 2.2), and poorer quality of life than the general population (mean 44.1) (Table 2).

Translation of the questionnaire

The researchers, original author, and professional translator discussed a number of challenges throughout the translation process. We present only substantial challenges here.

In the forward translation, the authors replaced the original term “injury” with “pain condition” in order to better represent a hospital pain clinic population. “Scale” was changed to “questionnaire” since this word is more commonly used in Norwegian for clinical screening and evaluation tools. The expression “being robbed of” (statement 10) was altered to “taken away from”. “Being robbed” is a dramatic experience and an intentional action, while “taken away from” is a more commonly used and neutral expression in Norwegian.

In the individual cognitive debriefing interviews, patients pointed out that the language was too formal, and “designed” and “indicate” were foreign words. We therefore made the language less formal by replacing the aforementioned verbs with “made” and “specify”.

A number of common issues were raised and discussed by the patients in both the cognitive debriefing interviews (2015) and focus group interviews (2018). The majority of patients found the introduction of the IEQ-N easy to understand and informative, but a descriptive title for the questionnaire was requested. All items of the IEQ-N were easy to understand, and relevant to the feeling of injustice.

Some patients highlighted that item 3 “I am suffering because of someone else’s negligence” is relevant to only those who have pain due to an accident or medical malpractice, and not to those who’s pain condition is due to incidence (i.e., genetic conditions), or own negligence (i.e., life style disease). Some patients also suggested that “uaktsomhet” (negligence) is a foreign word often used in terms of law/justice, and may not be understood by everyone. However, for the majority of patients “uaktsomhet” was comprehendible.

Some patients scrutinized the general approach of the questionnaire items. They suggested that the items should address more specific factors related to the feeling of injustice, such as experiences with the health care and welfare system, as well as pain interference with family- and social life. In addition, a number of patients claimed that the questionnaire evoked both positive feelings (i.e., they felt understood, difficult thoughts were validated, hope) and negative feelings (i.e., depressed, angry, provoked, suspicious, self-pity, shame).

Registry analyses

Internal consistency of the scale was assessed and showed a Cronbach alpha coefficient of 0.92.

The 12 items of the IEQ were further subjected to factor analysis (principal components analysis, PCA). Prior to performing PCA the suitability of data was assessed. We found no outliers or extreme values of importance for the eight variables. Inspection of the correlation matrix revealed that all coefficients but one (for item 3 and 5: 0.29) were of 0.3 and above. The Kaiser–Meyer–Oklin value was 0.95, exceeding the recommended value of 0.6 [43, 44] and the Bartlett’s Test of Sphericity [45] reached statistical significance, supporting the factorability of the correlation matrix.

The PCA revealed the presence of one component with eigenvalues exceeding 1, explaining 54% of the variance. An inspection of the screeplot supported a one-factor structure (Figure 1). This was further supported by the results of Parallel Analysis, which showed only one component with eigenvalues exceeding the corresponding criterion values for a random generated data matrix of a similar size (12 variables × 2,500 respondents). The component matrix (Table 3) also shows that all items load strongly (all above 0.56) on only one factor.

Figure 1:

Screeplot showing a one-factor structure of the Norwegian Injustice Experience Questionnaire (IEQ-N).

Oblimin rotation with a forced two-factor structure showed that this model explains 62% of the variance. Most items making up the blame/unfairness subscale (items 7, 9, 10, 11, 12, except 3) load on the same factor, while items making up the severity/irreparability loss subscale loads on both factors. These results support a one-factor structure. All communality values exceed 0.3 indicating that all items fit well with the other items (Table 4).

Correlations between experienced injustice and concurrent health measures were moderate to strong (see Table 5). The strongest correlation was with pain catastrophizing (r=0.71, p≤0.001), followed by psychological distress (r=0.61, p≤0.001).

The three hierarchical regression models showed that perceived injustice contributed with significant unique variance to pain related factors, but the additional contribution beyond pain catastrophizing was small. The additional contributions of IEQ-N were 0.2% (p=0.02), 2.8% (p<0.01), and 6.7% (p<0.01) for pain intensity, pain-related disability, and psychological distress, respectively (Table 6).

Translation of the IEQ-N

The first aims of the study were to translate the IEQ from English to Norwegian, and evaluate the experience and understanding of the questionnaire in patients with persistent pain.

In general, patients reported that the questionnaire was “spot on”, and that the 12 items were relevant for persons with chronic pain conditions. We believe that the thorough discussions of issues and meaning of words and concepts among the actors (i.e., authors, original author M. Sullivan, professional translator, patients) at every step of the translation process contributed to this sound Norwegian version of the IEQ.

Few minor but important changes were made to the questionnaire, and mainly involved replacing single words with synonyms more commonly used in Norwegian. Some suggested alterations from the patients were considered, but not met. For instance, patients pointed out that item 3 “I am suffering because of someone else’s negligence” is relevant to only those who have pain due to an accident or medical malpractice, and not to those who’s pain condition is due to incidence (i.e., genetic conditions) or own negligence (i.e., life style disease). When it comes to blame, perceived injustice is particularly prominent when another person can be held responsible for the situation [2]. The relation between other sources of blame and the perception of injustice is to our knowledge not known. Therefore, further qualitative studies should explore the relevance of various sources of blame for perceived injustice, before considering changing item 3 to encompass blame directed toward additional sources.

One important issue discussed in the patient interviews was that items of the IEQ were phrased too generally. The patients suggested that some of the items should more directly address the unfair challenges experienced with the Norwegian welfare and healthcare system, as well as the unjust situation of pain interference with personal and social life. These suggestions were thoroughly discussed, but we decided to keep the general approach for several reasons. Making the questionnaire more specific, would involve changing existing items as well as adding new items to the questionnaire. Such an operation would change the content of the IEQ-N substantially, and make it difficult to compare experience of injustice across countries and cultures [24]. Additionally, we believed that the general approach has an important function by making the IEQ more applicable to different patient groups and various circumstances for experiences of injustice.

Another major issue raised in the patient interviews, was that the questionnaire evoked negative feelings such as depression, self-pity, and shame. This phenomenon was particularly prominent in patients with higher IEQ scores. In addition, some became angry and provoked by the questions, and suspicious of the clinic’s intention of distributing the questionnaire. One may argue that a questionnaire that evokes negative emotions, makes the patients project their emotions into their answers, and as a consequence make an exaggerated picture of their situation. However, research shows that assessment of negatively loaded constructs does not worsen the patients’ health condition or psychological wellbeing [46, 47]. The questionnaire may evoke negative emotions and cognitions then and there, but do not make them stronger on a permanent basis or create new ones. Note that the patients in the present study also emphasized that the IEQ evoked positive emotions. The questionnaire addressed their difficult thoughts regarding experiences of injustice, and thereby the patients felt understood and that their emotions were validated.

A substantial change proposed in the patient interviews, was to make the questionnaire’s title more informative. Some suggested mentioning explicitly “perceived injustice” in the title in order to highlight the concept assessed by the questionnaire. However, after thorough discussions between the researchers and original author, we decided to keep the title “IEQ” to avoid potential bias from too much disclosure of the intended construct being assessed.

Factorial structure of IEQ

The third aim of the present study was to test the factorial structure of the IEQ. The theoretical construct of perceived injustice compromises two major aspects, (1) attribution of blame; and (2) magnitude and irreparability of loss. However, our factor analysis supported a one-factor structure. Interestingly, some validations of the IEQ showed similar results [11, 13, 15] while other validations indicated at two-factor or three-factor structure [2, 9, 12]. However, the original author suggests that perceived injustice is a complex but unitary construct due to a number of factors: The high internal reliability, high correlation between the two components of the factor analysis, and the fact that several items load on both factors of the factor analysis [2, 5]. Our findings support these arguments. Although the forced two-factor solution showed strong factor loadings on blame/unfairness, half of the items on the other factor (irreparability/magnitude of loss) also loaded substantially on the first factor. While the two subscales of the IEQ are plausible from a theoretical point of view, the 12 IEQ items seem to represent a unitary construct of perceived injustice.

Validity and reliability of IEQ

The fourth aim of the present study was to investigate the validity and reliability of the IEQ-N. The construct validity of the Norwegian version was good, as shown with the moderate to strong significant univariate correlations between perceived injustice and related constructs such as pain catastrophizing and disability, psychological distress, and quality of life. These findings are supported by previous studies on validation of the IEQ [2]. In addition, injustice was significantly, but to a smaller extent, related to pain severity. This finding may be explained by the crude measures of pain severity used in the present study; two single 0–10 NRS for pain intensity and bothersomeness. Similar findings are found in studies also using unidimensional measures of intensity for pain severity in patients with chronic pain conditions [3, 4, 9, 10, 14], but not in trauma patients [8, 15]. Interestingly, studies that used multidimensional measures of pain severity [2, 3, 7, 10] showed higher correlations between the constructs. An explanation may be that the experience of injustice is more strongly correlated to the affective, cognitive, and reactive dimensions of chronic pain, than the sensory dimension of pain intensity.

The internal reliability of the IEQ-N was high, with a Cronbach’s alpha of 0.92. Previous studies also show high degrees of internal reliability with values between 0.89 and 0.93 [2, 9], [10], [11], [12]. In fact, values exceeding 0.90–0.95 [48] indicate redundancy among the items of the questionnaire, and calls for further work on reducing the number of items. In order to overcome overlap of items and provide a questionnaire more suitable for screening purposes, Sullivan and colleagues developed the IEQ short form in 2016 [49]. This shorter version only retains five of the original items (items 1, 2, 4, 5 and 6), and shows acceptable, yet a little poorer, validity and reliability than the original 12 item IEQ. Further research should therefore investigate redundant items of the IEQ-N, and evaluate the need of item reduction.

Like previous research [2, 7, 9], this study shows that perceived injustice and pain catastrophizing are closely related. The correlation between the two concepts was strong (r=0.71, p<0.01), and regression analyses (n=2,700) showed statistically significant but small (0.2–6%) unique contribution of perceived injustice above pain catastrophizing on the prediction of variance in pain intensity, pain disability, and psychological distress. Note that previous studies [7, 9] show greater magnitude of the unique contribution of perceived injustice on pain related outcomes. The use of different measures of pain outcomes, as well as cultural differences in the perceptions of the concepts under study, may partially explain this discrepancy of findings. However, although research shows that there is a considerable statistical overlap between perceived injustice and catastrophizing, the resemblance between the two concepts and the questionnaire items also needs to be considered. In fact, the “irreparability of loss” subscale of the IEQ does resemble the “exaggeration of threat” dimension reflected in several items of the PCS (e.g., items 6, 7, 13) [2], however the IEQ items of blame do not reflect the concept of pain catastrophizing nor any of the items or subscales of the PCS [2, 7]. Further studies should investigate the conceptual relation and cultural understandings of perceived injustice and pain catastrophizing in order to reveal the uniqueness of the concepts.

Strengths and limitations of the study

The present study’s major strengths are the mixed-method design where both quantitative and qualitative data are integrated and used to address the research questions, as well as the large sample size in the quantitative data and heterogeneous study population of patients with chronic pain. These strengths allow us to explore perceived injustice across various pain diagnoses, unlike most previous research that focuses on specific pain diagnoses. In addition, the generalizability of findings is good, but one need to consider that the sample is drawn from a rather homogenous Norwegian population. Note also that the registry data was not adequate to explore test–retest reliability of the IEQ-N. Although the Canadian and Spanish versions of IEQ [2, 9] show that the questionnaire measures stable experience of injustice, further studies are needed on the ability to detect relevant changes with appropriate interventions over time.

Conclusions

The present study used mixed research methods in the translation and validation process of the Norwegian version of the IEQ (IEQ-N). Qualitative analyses show that the patients in general find the questionnaire relevant for their situation, and easy to understand. Quantitative analyses indicate good criterion validity with high correlations to similar construct of pain catastrophizing and psychological distress, and high reliability in terms of internal consistency. The one factor structure of the IEQ found in the present study is supported by previous research. Although overlap between concepts, perceived injustice showed a small unique contribution above pain catastrophizing on the prediction of pain outcomes. The questionnaire is a valid and reliable tool for assessing experience of injustice in Norwegian patients with various persistent pain conditions.