DOI: https://doi.org/10.21203/rs.3.rs-48391/v1
Background: Catastrophizing has been recognized as an important contributor to chronicity in individuals with chronic pain syndromes including low back pain (LBP). The Pain Catastrophizing Scale (PCS) is perhaps the most widely used questionnaire to evaluate the degree of pain catastrophizing. However, its use is limited in Hausa-speaking countries due to the lack of a validated translated version.
Objective: To translate and cross-culturally adapt the PCS into Hausa (Hausa-PCS) and evaluate its psychometric properties in mixed urban and rural patients with chronic LBP.
Methods: The original PCS was translated and cross-culturally adapted into Hausa in accordance with established guidelines. To evaluate psychometric properties, a consecutive sample of 200 patients with chronic LBP was recruited from urban and rural Nigerian hospitals. Validity was evaluated by exploring content validity (skewness and item-total correlation), factorial structure (exploratory factor analysis [EFA] and confirmatory factor analysis [CFA]), concurrent validity (Spearman’s rho) and known-groups validity. Reliability was evaluated by calculating internal consistency (Cronbach’s α), intraclass correlation coefficient (ICC), standard error of measurement (SEM), minimal detectable change (MDC) and limits of agreement with 95% confidence interval (LOA95%).
Results: The Hausa-PCS demonstrated good face and content validity. Both EFA and CFA confirmed a 3-factor structure similar to the original English version. The concurrent validity was supported as 83% (5/6) of the a priori hypotheses were confirmed. Known-groups comparison showed that the questionnaire was unable to differentiate between male and female or urban and rural patients (p > 0.05). Internal consistency and ICC was adequate for the Hausa-PCS total score (α = 0.84; ICC = 0.90) and the subscale helplessness (α = 0.78; ICC = 0.89) but for the subscales rumination (α = 0.69; ICC =0.68) and magnification (α = 0.41; ICC = 0.43). The LOA95% for the Hausa-PCS total score was between –8.10 and +9.75, with SEM and MDC of 3.47 and 9.62 respectively.
Conclusion: The Hausa-PCS was successfully developed and psychometrically adequate in terms of factorial structure, concurrent validity, internal consistency and test-retest reliability when applied in mixed urban and rural patients with chronic LBP. However, the internal consistency and reliability coefficients (ICC) for the individual subscales are questionable. Thus, we support the use of the total score when evaluating pain catastrophizing for clinical or research purposes.
Low back pain (LBP) remains the most common painful musculoskeletal disorder affecting the population indiscriminately across the world [1]. It is the greatest contributor to years lived with disability worldwide [2] and imposes a sizable economic, societal and health impact [3, 4]. Though LBP is considered to be a multifactorial disorder associated with numerous possible etiologies and diverse interpretations of the underlying mechanisms [3, 5], in most cases, it is non-specific, signifying that the cause of the pain cannot be reliably identified [6].
Although it is commonly believed that most people experiencing a new episode of LBP recuperate within a few weeks, reoccurrences are possible and some fractions may go on to develop chronic LBP [7]. The development of non-specific chronic LBP is believed to be multifaceted with biomechanical and psychosocial factors being implicated [5, 8]. However, while biomechanical factors appear to have a greater impact on the occurrence of LBP episodes, psychosocial factors seem to have a major impact on its chronicity, as the latter predicts the transition to and maintenance of chronic LBP [6, 8–10].
One important psychological factor linked with the chronicity of LBP is pain catastrophizing. Catastrophizing is a maladaptive coping strategy defined as an exaggerated negative mental state related to an actual or anticipated painful experience [11]. Pain catastrophizing has been recognized as an important mediator to pain behavior and pain-related fear in individuals with chronic pain conditions [12, 13]. It is closely related to fear-avoidance beliefs [14] as the former is thought to be a precursor of the latter [15]. According to the fear-avoidance model, when pain is interpreted as threatening, it influences the use of a catastrophizing pain coping style which in turn may influence pain-related fear to produce avoidance and hypervigilance to bodily sensation that is followed by physical disuse, functional disability, depression and pain chronicity [16]. There is evidence to suggest that pain catastrophizing predicts persistent pain and chronic LBP disability [17–20] as well as a mediator and arbitrator of treatment effectiveness among sufferers of chronic LBP [21]. Furthermore, evidence suggests that catastrophizing increases with treatment failure, and improves significantly with effective treatments [22]. Thus, pain catastrophizing is an important predictor of patient outcomes [14], and evaluating this psychological construct in patients with suspected risk is essential to guide the choice of a therapeutic intervention targeting such factor.
The Pain Catastrophizing Scale (PCS) developed by Sullivan et al.[23] is perhaps the most widely used tool to evaluate the degree of pain catastrophization in clinical practice and research. The PCS is a valid and reliable measure of how catastrophizing impact on pain experience [23, 24]. Essentially, it has been proved to be a useful measure of pain catastrophizing in various pain conditions such as chronic LBP [25], chronic neck pain [26], anterior knee pain [27], neuropathic pain [28], postsurgical pain [29], soft tissue injuries [30], respiratory tract illness [31] and dental procedures [32]. Furthermore, the scale has been culturally adapted and validated into many languages/cultures such as the Arabic [33, 34], Afrikaans [35], Brazilian Portuguese [36], Catalans [37], Chinese [38], German [39], Italian [40], Korean [41], Malay [42], Norwegian [43], Simplified Chinese [44], Sinhala [45], Swedish [46], Spanish [47], Turkish [48] and Xhosa [35] versions.
Chronic LBP appears to be a major cause of disability in Nigeria with an estimated annual prevalence of 33–74% [49]. The burden, however, is unduly greater in rural areas compared to urban areas as the one-year prevalence rate of 74% found in rural Nigeria is higher than the 44% found in urban Nigeria [50, 51]. In the same vein, maladaptive beliefs including pain catastrophizing have been reported to be associated with chronic LBP disability in both urban and rural Nigeria [52, 53] similar to that found for western nations [54]. Despite the greatest burden of chronic LBP in Nigeria, self-report outcomes to evaluate cognitive or maladaptive beliefs are generally lacking in Nigeria indigenous languages.
There are over 500 native languages spoken in Nigeria, with English being the official language of communication. However, it has been reported that a significant number of patients cannot speak or write in English [55, 56]. The Hausa language is one of the three major native languages spoken in the country particularly in the northern part. The other two languages are Igbo and Yoruba. Although the Hausa language is also commonly spoken in other West African countries including Cameroon, Chad, Ghana, Benin Republic, Ivory Coast, Mali, Niger and Sudan [57] and estimated to be spoken by 100–150 million people worldwide, the most important dialect is generally regarded as that spoken in Kano, northwestern Nigeria. This language is the standard variety used for official purposes. Therefore adapting the PCS into Hausa in this context will facilitate its use not only in Nigeria but also in other Hausa-speaking countries. The purpose of this study was to translate and cross-culturally adapt the PCS into Hausa and evaluate its psychometric properties in mixed urban and rural patients with chronic LBP.
This study got ethical approval from the Health Research Ethics Committee, Ministry of Health Kano State (Ref: MOH/Off/797/T.I./651). Written permission (via email) to translate the PCS into Hausa language was obtained from the original developer (Prof. Michael J. Sullivan) and copyright holder (MAPI Research Trust) of the PCS. Written informed consent was obtained from all participants prior to their involvement in the study.
Translation, cross-cultural adaptation, test-retest and cross-sectional study of psychometric analysis of the Hausa version of the PCS.
The original PCS consists of 13 items, with each item rated using a 5-point Likert scale ranging from 0 (not at all) to 4 (all the time) [23]. Each item is rated according to respondent’s perceived thoughts and feelings while experiencing pain. The total score ranges from 0 to 52 with higher scores indicating more catastrophic thoughts [23]. The PCS has three dimensions; rumination (4 items), magnification (3 items) and helplessness (6 items). The scale has been shown to have strong construct validity, reliability and stability [23, 24, 58].
The Hausa version of the NPRS [59] was used to evaluate levels of the pain intensity. The scale consists of an 11-point numerical pain rating scale with a horizontal bar or line of 0 to 10 cm anchored on the left with the phrase ‘‘No Pain’’ and on the right with the phrase ‘‘Worst Imaginable Pain. The respondents were asked to mark the value that best reflects their current pain at rest. The Hausa NPRS is a valid, reliable and responsive measure of pain intensity in patients with chronic LBP [59].
The Hausa version of the ODI 2.1a [55] was used to evaluate levels of functional disability. It consists of 10 topics concerning pain intensity, personal care, lifting, walking, sitting, standing, sleeping, sex life, social life, and traveling. Each question has six statements scored from 0 to 5. Scores obtained for each topic are summed and divided by the number of answered topics to give a final score out of 100 which indicates the respondent’s percentage perceived level of disability (0–100), with higher scores indicating greater disability [60]. The Hausa-ODI 2.1a was found to be a valid and reliable measure of functional disability in chronic LBP patients [55].
The Hausa version of the FABQ [61] was used to evaluate fear-avoidance beliefs. It consists of 16 items, with each item scored using a Likert scale ranging from 0 (completely disagree) to 6 (completely agree). The instrument consists of two subscales: a 4-item physical activity subscale (FABQ-physical activity) and a 7-item work subscale (FABQ-work). Each subscale scores are summed to give a total score with the FABQ-physical activity subscale having a score ranging from 0 to 24 and the FABQ- work subscale having a score ranging from 0 to 42. Summing the two subscale scores gives a total FABQ score of 66, with higher scores indicating greater fear-avoidance beliefs [62]. The Hausa FABQ is a valid and reliable measure of fear-avoidance beliefs in patients with chronic LBP [61].
The Hausa version of the SF-12 [63] was used to evaluate mental well-being. It consists of 12-item, and evaluates two global health constructs: the physical component summary (PCS-12) and the mental component summary (MCS-12). Each item of the questionnaire has response categories which vary from 2 to 6 and raw scores for items ranging from 1 to 6. To calculate the PCS-12 and MCS-12 sores, a web-based scoring tool (www.orthotoolkit.com/sf-12/) was used. Higher scores indicate better health status. The Hausa SF-12 was shown to be a valid and reliable measure of health-related quality of life in patients with chronic LBP [63].
The translation procedure was conducted according to the guidelines published by Beaton et al.[64]. The translation included 5 steps as follows:
Forward translation: The PCS was translated from English into Hausa by two independent bilingual translators (Hausa and English, with Hausa as their first language). The first translator was a clinical physiotherapist and familiar with the concept of the questionnaire. The second translator was a professional translator and unaware of the concept being examined. The translators produced two forward-translated versions (T1 and T2).
Synthesis of forward translations: The two forward translated versions were then synthesized to one version (T3) following consensus between the two forward translators and coordination of the primary author (AAI).
Backward translation: The synthesized version (T3) was then back-translated into English by two independent bilingual translators (Hausa and English) who had no medical background and knowledge of the original English version. The translators produced two backward-translated versions (T4 and T5).
Expert committee review: An expert committee consisting of all forward and backward translators, a methodologist and two of the study authors (AAI and BK) reviewed all the translated versions and reached a consensus on any discrepancy with the aim of achieving semantic, idiomatic, experiential and conceptual equivalences between the original and targeted versions. A prefinal version of was then produced.
Pilot testing: The prefinal version was tested in a group of 20 patients with chronic LBP to evaluate comprehensibility and acceptability. Problematic items and how they were resolved were documented. This stage confirmed face validity.
Proofreading: A professional translator independently proofread the final version for any minor errors that may have been missed in the previous stages. The final version (see Additional file 1) was then produced and sent to MAPI Research Trust.
The procedure used throughout this section has been used in the cross-cultural adaptation of other Hausa self-report measures as described elsewhere [55, 63].
Generally, there is no clear consensus on the required sample size for a validation study. However, “The quality criteria for measurement properties of health status questionnaires” suggest that a sample size of ≥ 50 would be adequate for reliability, construct validity, and ceiling/floor effects analyses whereas 4 to10 subjects per variable (Rules-of-thumb) would be adequate for factorial structure analysis. Based on these recommendations, 200 participants were recruited to study the psychometric properties of the Hausa version of the PCS (Hausa-PCS).
The study was carried out purposely in a selected tertiary health facility (Murtala Muhammad Specialist, and three secondary health facilities (Dawakin-Kudu General Hospital, Wudil General Hospital and Kura General Hospital), all in Kano State, Northwestern Nigeria. These hospitals were chosen to recruit both urban and rural patients to have broader applicability of the questionnaire in these settings. The participants were recruited from the physiotherapy out-patient unit of the selected hospitals, from February to May 2018. They were included if they were males or females, aged 18 to 70 years, suffering from LBP for 3 months or greater and fluent in Hausa language. Participants with previous spine surgery, evidence of serious spine pathology for example infection, malignancy, fracture, osteoporosis or inflammatory disease, cognitive impairment or impaired capacity to be interviewed, and pregnancy were excluded.
Four physiotherapists (with clinical experience between two to five years) were recruited from the selected hospitals and received a one-day training session on the study procedure including interviewer-administration of measures as many Hausa patients especially rural dwellers are not literate (ability to read and write in Hausa). The training was conducted by the primary author. The physiotherapists in each of the selected hospitals were responsible for assessing patients’ eligibility which involves medical history taking, screening of ‘red flags’ (using simple questions about the presence of red flags) to rule out evidence of serious spine pathology, and obtaining patients’ informed consent.
The participants’ socio-demographic information (age, gender, marital status, education level, occupation and habitation) and data on duration of pain, height, weight and body mass index were obtained and documented for each participant. The Hausa-PCS along with the NPRS, ODI, FABQ and SF-12 were administered using interviewer-administration or self-administration method where applicable. The Hausa PCS was re-administered among 100 participants, 2 weeks after the first measurement to assess test-retest reliability.
The normality of the data was tested using visual (normal distribution curve and Q-Q plot) and statistical methods (Kolmogorov-Smirnov and Shapiro-Wilk’s test). Descriptive statistics of mean, standard deviation (SD), frequencies and percentages were applied to summarize the data. The following statistical approaches were used in evaluating the psychometric properties of the Hausa-PCS.
Ceiling and floor effects: Ceiling or floor effects are considered if more than 15% of respondents scored the maximum or minimum possible score (Table 1). Potential ceiling or floor effects of the Hausa-PCS were investigated by estimating the percentage of respondents indicating the maximum or minimum possible score in all the 13 items of the questionnaire.
Content validity: Content validity refers to the degree to which an instrument is relevant and representative of the construct it is designed to measure. Content validity of the Hausa-PCS was evaluated by examining response trend (using skewness) and Item-total correlations. Items with a skewness > 1.96 suggest a response trend that deviated from a normal distribution pattern [44]. Item-total correlations were examined to identify if the items were appropriate for the construct. An item-total correlation coefficient < 0.40 suggests that the items did not measure the same construct [44].
Factorial validity: Factor structure of the Hausa-PCS was examined by first performing exploratory factor analysis (EFA) using principal component analysis with Promax rotation to verify the original three-factor structure of the PCS (i.e. rumination, magnification and helplessness) extracted by Sullivan et al. [23]. Subsequently, a confirmatory factor analysis (CFA) using maximum likelihood estimates was then carried out with the three-factor structure. Additionally, CFA was performed with the one-factor and two-factor structure extracted by Chibnall and Tait [65]. Modification indices were applied to observe for item’s redundancy or those with low factor loadings, and correlation of error terms to improve model fit [66]. The model fit was assessed with four goodness-of-fit indicators including the ratio of chi-square to degrees of freedom (χ2/df), comparative fit index (CFI), Tucker-Lewis index (TLI), standardized root mean square residual (SRMR) and root mean square error of approximation (RMSEA) [63]. The following cut-off criteria were considered adequate for model fit; χ2/df of ≤ 2.0, CFI of ≥ 0.95, TLI ≥ 0.90, SRMR ≤ 0.08, and RMSEA ≤ 0.06 [67, 68].
Construct validity: Construct validity was evaluated through concurrent validity (i.e. how well a new instrument correlates with well-established instruments) and known-group validity (how well an instrument discriminates between relevant known or extreme groups). Concurrent validity was evaluated by correlating the Hausa-PCS with measures of pain intensity (NPRS), functional disability (ODI), fear-avoidance beliefs (FABQ) and mental well-being (MCS-12). Spearman’s correlation coefficients (rho) were used and considered as being strong (rho = > 0.60), moderate (rho = 0.30–0.60) and weak/low (rho = < 0.30) [69]. The expected direction and magnitude of the correlations were formulated a priori as shown in Table 1. According to Terwee et al.[70], the construct validity is supported when at least 75% (≥ 5) of the predefined hypotheses are confirmed (Table 1). Known-groups validity was evaluated by comparing the Hausa-PCS total score and its subscales with gender and habitant groups using independent t-test. We hypothesized that female and rural respondents would have higher pain catastrophization [71, 72].
Internal consistency: Internal consistency of the whole questionnaire and its subscales was evaluated with Cronbach’s alpha (α). A Cronbach’s α values of 0.70–0.95 indicate acceptable internal consistency [70].
Test-retest reliability: Test-retest reliability was assessed by calculating intraclass correlation coefficient (ICC) for agreement using a two-way random effects analysis of variance (ANOVA) model, with a coefficient value ≥ 0.70 indicating acceptable reliability [70]. As per the recommendation of the Consensus-based Standards for the selection of health Measurements Instruments (COSMIN) [73], the standard error of measurement (SEM) and minimal detectable change (MDC) at 95% confidence interval (CI) were calculated to compliment the test-retest reliability. The SEM is computed as the square root of the mean square error term from the reliability ANOVA table. The MDC was then calculated by multiplying the SEM by 2.77 to indicate the minimum amount of change that needs to be observed for it to be considered a true change above measurement error [74, 75]. Additionally, 95% limits of agreement (LOA95%) were evaluated with Bland-Altman plots by plotting the difference between baseline and follow-up Hausa-PCS total scores against the mean of NPRS-H total scores at baseline and follow-up. A priori hypotheses for the internal consistency (Cronbach’s α), ICC, SEM, MDC and LOA95% for the Hausa-PCS are presented in Table 1.
| Psychometric properties | Hypotheses |
|---|---|
| Floor and ceiling effects | |
| 1. Ceiling effects | 15% of the respondents having the maximum score (0) [70] |
| 2. Floor effects | 15% of the respondents having the minimum score (52) [70] |
| Reliability | |
| 1. Internal consistency | Cronbach’s α = 0.70–0.95 [70] |
| 2. Test-retest reliability | Intraclass correlation coefficient = ≥ 0.70 [70] |
| 3. Standard error of measurement | 1.6–4.6 [35, 39, 41, 43, 76, 77] |
| 4. Minimal detectable change | 4.0–13.0 [35, 39–41, 43, 76, 77] |
| 5. 95% limits of agreement | –15.1 to + 16.0 [43, 44, 76, 77] |
| Concurrent validity | |
| 1. PCS vs FABQ-total, FABQ-physical and FABQ-work | Significant moderate to strong positive correlation (rho; 0.34–0.61) [39, 43, 48] |
| 2. PCS vs NPRS | Significant moderate to strong positive correlation (rho; 0.31–0.64) [40, 41, 43, 76] |
| 3. PCS vs ODI | Moderate to strong positive correlation (rho; ≥ 0.30) [76] |
| 4. PCS vs MCS-12 | Significant moderate negative correlation with (rho; ≥ − 0.30) [42] |
| PCS: Pain Catastrophizing Scale; FABQ: Fear-Avoidance Beliefs Questionnaire; NPRS: Numerical Pain Rating Scale; ODI: Oswestry Disability Index; rho: Spearman correlation coefficient | |
There were no major disagreements between the forward and back translations of the Hausa-PCS. The phrase “I feel I can’t go on” in item 2 was somewhat difficult to translate into Hausa. The translators, however, decided to use the phrase “carry on” in place of “go on” for easy understanding. The translators ensured that standard Hausa was used to attain equivalence between the original English questionnaire and the Hausa version. None of the respondents reported any difficulty with comprehension of the questionnaire items during the pilot testing. Thus, no further modification was carried out and all the items were retained by the expert committee.
Of the 200 participants recruited, the response rate was 100%. There were 123 (61.5%) males and 77 (38.5%) females. Their age ranged between 18–70 years (mean age 45.5 ± 14.5 years). The majority of the participants were living in rural areas (60%). Slightly over half of them were illiterate in Hausa (55.5%) and self-employed (mainly farmers and traders). The socio-demographic and clinical characteristics of the participants are fully presented in Table 2.
| Characteristics | N = 200 |
|---|---|
| Age, years, mean ± SD | 45.5 ± 14.5 |
| Gender, n (%), male: female | 123 (61.5), 77 (38.5) |
| Habitation, n (%), urban: rural | 80 (40.0), 120 (60.0) |
| Marital Status, n (%), married: unmarried | 157 (78.5), 43 (21.5) |
| Educational Status, n (%) | |
| None | 66 (33.0) |
| Completed Primary Education | 30 (15.0) |
| Completed Secondary Education | 41 (20.0) |
| Completed Tertiary Education | 63 (31.5) |
| Literacy (ability to read and write), n (%) | |
| Illiterate (inability to read and write) | 111 (55.5) |
| Hausa | 89 (44.5) |
| Occupational Status, n (%) | |
| Paid work (government or private) | 49 (24.5) |
| Self-employed (farming and trading) | 112 (56.0) |
| Student | 17 (8.5) |
| Unemployed | 16 (8.0) |
| Retiree | 6 (3.0) |
| PCS (score range 0–52) | 30.0 ± 8.21 |
| FABQ-total (score range 0–64) | 36.4 ± 11.4 |
| FABQ-physical activity (score range 0–42) | 13.1 ± 5.81 |
| FABQ-work (score range 0–24) | 23.3 ± 7.74 |
| NPRS (score range 0–10 cm) | 36.1 ± 12.6 |
| ODI (score range 0–100) | 36.0 ± 10.8 |
| SD: standard deviation; PCS: Pain Catastrophizing Scale, FABQ: Fear-Avoidance Beliefs Questionnaire; NPRS: Numeric Pain Rating Scale; ODI: Oswestry Disability Index | |
All the respondents completed the Hausa-PCS without missing values. Ceiling effects were found for items 1, 3, 6, 7, 8, 10, 11 and 13 whereas floor effects were found for items 1, 11 and 13. No ceiling or floor effects were seen in the Hausa-PCS total score or subscales (Table 3).
| Range | Mean (SD) | Highest score | Lowest score | Ceiling effects n (%) | Floor effects n (%) | ||
|---|---|---|---|---|---|---|---|
| Total score | 0–52 | 30.0 (8.21) | 52 | 1 | 1 (0.5) | 3 (1.5) | |
| Rumination subscale | 0–16 | 10.3 ( 3.20) | 16 | 4 | 7 (3.5) | 11 (5.5) | |
| Item 8 | I anxiously want the pain to go away | 0–4 | 2.86 (1.10) | 4 | 0 | 72 (36.0) | 0 (0.5) |
| Item 9 | I can’t seem to keep it out of mind | 0–4 | 2.33 (1.09) | 4 | 0 | 31 (15.0) | 2 (1.0) |
| Item 10 | I keep thinking about how much it hurts | 0–4 | 2.41 (1.07) | 4 | 0 | 38 (19.0) | 1 (0.5) |
| Item 11 | I keep thinking about how badly I want the pain to stop | 0–4 | 2.81 (1.13) | 4 | 1 | 76 (38.0) | 36 (18.0) |
| Magnification subscale | 0–12 | 6.99 (2.55) | 12 | 2 | 6 (3.0) | 1 (0.5) | |
| Item 6 | I become afraid that the pain may get worse | 0–4 | 2.26 (1.10) | 4 | 0 | 31 (15.5) | 1 (1.0) |
| Item 7 | I think of other painful experiences | 0–4 | 2.21 (1.12) | 4 | 0 | 33 (16.5) | 3 (1.5) |
| Item 13 | I wonder whether something serious may happen | 0–4 | 2.51 (1.17) | 4 | 1 | 54 (27.0) | 59 (29.5) |
| Helplessness subscale | 0–24 | 12.7 (4.20) | 24 | 2 | 1 (0.5) | 1 (0.5) | |
| Item 1 | I worry all the time whether the pain will end | 0–4 | 2.40 (1.04) | 4 | 1 | 37 (18.5) | 47 (23.5) |
| Item 2 | I feel I can’t go on | 0–4 | 1.96 (0.96) | 4 | 0 | 14 (7.0) | 3 (1.5) |
| Item 3 | It’s terrible and I think it’s never going to get any better | 0–4 | 2.21 (1.08) | 4 | 1 | 31 (15.5) | 1 (0.5) |
| Item 4 | It’s awful and I feel that it overwhelms me | 0–4 | 1.92 (1.07) | 4 | 0 | 22 (11.0) | 5 (2.5) |
| Item 5 | I feel I can’t stand it any more | 0–4 | 2.15 (1.07) | 4 | 0 | 29 (14.5) | 1 (0.5) |
| Item 12 | There is nothing I can do to reduce the intensity of the pain | 0–4 | 2.08 (1.06) | 4 | 0 | 24 (12.0) | 1 (0.5) |
| SD: standard deviation | |||||||
Table 4 shows that the scores for each item of the Hausa-PCS were normally distributed, and none of the items exhibited a low item-total correlation < 0.40. Thus, none of the items was excluded in the Hausa-PCS.
| Item | Skewness | Corrected item-total correlation |
|---|---|---|
| Item 1 | 0.121 | 0.400 |
| Item 2 | 0.458 | 0.460 |
| Item 3 | 0.271 | 0.508 |
| Item 4 | 0.573 | 0.405 |
| Item 5 | 0.372 | 0.548 |
| Item 6 | 0.106 | 0.579 |
| Item 7 | 0.209 | 0.446 |
| Item 8 | –0.596 | 0.472 |
| Item 9 | 0.189 | 0.586 |
| Item 10 | 0.028 | 0.513 |
| Item 11 | –0.389 | 0.461 |
| Item 12 | 0.415 | 0.468 |
| Item 13 | –0.071 | 0.488 |
The Hausa-PCS total score demonstrated significant strong positive correlation with NPRS (rho = 0.74, p < 0.001) and significant moderate positive correlation with FABQ-total (rho = 0.42, p < 0.001), FABQ-physical activity (rho = 0.32, p < 0.001), FABQ-work (rho = 0.36, p < 0.001) and ODI (rho = 0.35, p < 0.001) as hypothesized (Table 1). However, the correlation between the Hausa-PCS and MCS-12 was weakly negative (rho = − 0.20, p < 0.05) (Table 5). Overall, 83% (5/6) of the a priori hypotheses were confirmed (Table 1).
| Pain Catastrophizing Scale | |||
|---|---|---|---|
| Measures | rho | P-value | Hypothesis confirmed (Yes/No) |
| FABQ-total | 0.42 | 0.000 | Yes |
| FABQ-physical activity | 0.32 | 0.000 | Yes |
| FABQ-work | 0.36 | 0.000 | Yes |
| Numeric Pain Rating Scale | 0.74 | 0.000 | Yes |
| Oswestry Disability Index | 0.35 | 0.000 | Yes |
| MCS-12 | –0.20 | 0.004 | No |
| All outcomes are in Hausa. FABQ: Fear-Avoidance Beliefs Questionnaire; rho: Spearman correlation coefficient; MCS-12: mental component summary | |||
Known-groups comparison of the Hausa-PCS with regard to gender and habitation groups showed no significant differences in the questionnaire total score and its subscales (p > 0.05) (Table 6).
| Gender | Habitation | |||||||
|---|---|---|---|---|---|---|---|---|
| Male | Female | Urban | Rural | |||||
| Mean (SD) | Mean (SD) | t-cal | p-value | Mean (SD) | Mean (SD) | t-cal | p-value | |
| Total score | 30.2 (8.29) | 29.8 (8.14) | 2.559 | 0.056 | 30.2 (7.96) | 29.9 (8.41) | 0.106 | 0.745 |
| Rumination | 10.5 (3.26) | 9.34 (3.53) | 2.453 | 0.068 | 9.93 (3.57) | 10.1 (3.29) | 0.126 | 0.723 |
| Magnification | 7.30 (2.56) | 6.15 (2.42) | 2.593 | 0.057 | 6.62 (2.24) | 7.05 (2.78) | 0.672 | 0.414 |
| Helplessness | 13.1 (4.51) | 12.4 (4.47) | 1.550 | 0.207 | 12.7 (4.04) | 13.0 (4.82) | 0.065 | 0.799 |
| SD: standard deviation | ||||||||
The results of the EFA with Promax rotation and Kaiser criterion of eigenvalues > 1.0 revealed a three-factor solution explaining 54.38% of the total variance. Factor 1 included items 8, 9, 10 and 11 which reflect rumination subscale. Factor 2 included items 6, 7 and 13 which reflect magnification subscale. Factor 3 included items 1, 2, 3, 4, 5 and 12 which reflect helplessness subscale. The first factor accounted for 34.1% of the total variance, whereas factor two and three accounted for 12.5% and 7.69%, respectively. The item-factor loadings (λ) for factor 1 ranged .47–.80, and for factor 2 and 3 ranged .31–.61 and.31–.77, respectively.
Table 7 shows the results of CFA for the one-factor, two-factor and three-factor structures of the Hausa-PCS with and without modifications. All the tested models demonstrated poor fit as indicated by the fit indices except the three-factor structure after modifications. Modifications of the three-factor structure were done by allowing 5 error terms to covary (e1–e4, e3–e4, e8–e12, e9–e10, and e10–e11) (Fig. 1).
| Model and modifications | Confirmatory factor analysis | |||||
|---|---|---|---|---|---|---|
| χ2 (df) | χ2/df | CFI | TLI | SRMR | RMSEA (95%CI) | |
| 1. One-factor structure | ||||||
| a. No modifications | 194.9 (65) | 3.00 | 0.796 | 0.755 | 0.083 | 0.100 (0.084–0.117) |
| b. With modifications | 116.5 (60) | 1.94 | 0.911 | 0.884 | 0.065 | 0.069 (0.050–0.087) |
| 2. Two-factor structure | ||||||
| a. No modifications | 166.9 (64) | 2.60 | 0.838 | 0.803 | 0.080 | 0.090 (0.073–0.107) |
| b. With modifications | 111.5 (60) | 1.85 | 0.919 | 0.895 | 0.066 | 0.066 (0.046–0.084) |
| 3. Three-factor structure | ||||||
| a. No modifications | 132.4 (62) | 2.13 | 0.889 | 0.861 | 0.724 | 0.076 (0.058–0.093) |
| b. With modifications | 86.8 (57) | 1.52 | 0.953 | 0.936 | 0.056 | 0.051 (0.028–0.072) |
| χ2: chi-square; df: degrees of freedom; CFI: comparative fit index; TLI: Tucker-Lewis index; SRMR: standardized root mean square residual; RMSEA: root mean square error of approximation; CI: confidence interval | ||||||
As shown in Table 8, the internal consistency as measured by the Cronbach’s α, if item deleted was high (0.837) for the Hausa-PCS total score. Also, adequate internal consistency was found for the subscale helplessness (α = 0.78) but for the subscales rumination (α = 0.69) and magnification (α = 0.41) (Table 8).
The ICC for the Hausa-PCS total score was good (0.90; CI: 0.85–0.93). Similarly, acceptable ICC was found for the helplessness subscale (0.89; CI: 0.83–0.93) but for the subscales rumination (0.68; CI: 0.52–0.78) and magnification (0.43; CI: 0.16–0.62) (Table 8). The SEM for the Hausa-PCS total score and its subscales are presented in Table 8. The Bland and Altman analysis showed a mean difference between test and retest of 0.87, with LOA95% of − 8.10 to + 9.75 (Fig. 2).
| Internal consistency | Test-retest (repeatability) | SEM | MDC | |||||
|---|---|---|---|---|---|---|---|---|
| Cronbach’s α | Mean (SD) test | Mean (SD) retest | t1-t2 | p-value for test-retest | ICC (95% CI) | |||
| Total (0–52) | 0.84 | 29.8 (8.46) | 28.5 (8.19) | 1.25 | 0.012* | 0.90 (0.85–0.93) | 3.47 | 9.62 |
| Rumination (0–16) | 0.69 | 10.0 (3.40) | 9.40 (3.11) | 0.67 | 0.028* | 0.68 (0.52–0.78) | 2.24 | 6.20 |
| Magnification (0–12) | 0.41 | 6.87 (2.56) | 7.05 (3.54) | –0.18 | 0.630 | 0.43 (0.16–0.62) | 2.63 | 7.29 |
| Helplessness (0–24) | 0.78 | 12.9 (4.48) | 12.1 (4.17) | 0.76 | 0.005* | 0.89 (0.83–0.93) | 1.88 | 5.21 |
| SD: standard deviatuon; t1-t2: mean values at test subtracted from retest; ICC: intraclass correlation coefficient; CI: confidence interval; SEM: standard error of measurement; MDC: minimal detectable change | ||||||||
| *P < 0.05 | ||||||||
To be able to assess pain catastrophizing and designed interventions targeting this important factor in Hausa LBP population, this study described the development of the Hausa-PCS through translation and cross-cultural adaptation of the original PCS into Hausa, and finally validation of the translated version in mixed urban and rural individuals with chronic LBP. The results of the study suggested that the Hausa-PCS was comprehensible, acceptable, valid and reliable when evaluating catastrophic thinking related to pain in Hausa-speaking individuals with chronic LBP.
The PCS was fairly simple to translate as there were no serious translation issues encountered. The items of the questionnaire were easily understandable during field verbal pretesting with urban and rural participants. The translators ensured that standard Hausa wordings and phrases were used for easy understanding in both urban and rural contexts with the goal of achieving conceptual equivalence rather than literal translation. The PCS-Hausa appears to be acceptable as all the respondents completed the questionnaire without missing values. Although no ceiling or floor effects were observed in the total score or subscales similar to reports of previous studies [43, 78], however, ceiling effects were seen in 8 out of the 13 items whereas floor effects were seen in only 3 items. In line with our findings, ceiling effects in more than half of the PCS items were also reported in the Norwegian validation [43]. In contrast, respondents exhibiting floor and ceiling effects were removed in the validation of the Simplified Chinese PCS among chronic pain patients [44].
The mean total score of the Hausa-PCS was 30.0 and comparable to both the urban (30.2) and rural (29.9) respondents, indicating that the studied population experienced a high level of catastrophizing considering the report that pretreatment score of greater than 24 was associated with high follow-up pain outcomes. Thus, it can be inferred that individuals with catastrophizing scores greater than 24 as in the case of our sample may warrant intervention targeting catastrophization. Similar to the Simplified Chinese version of the PCS [44], the content validity of the Hausa-PCS was acceptable as the items were normally distributed and none of them demonstrated a low item-total correlation.
The PCS has been widely reported to poses a three-factor structure consisting of the rumination, magnification and helplessness subscales following EFA [24, 37, 39–42, 44, 45, 47, 65, 76] even though minor differences exist regarding how the PCS items loaded onto factors. A two-factor structure has been also reported in the literature [24, 48, 77, 79–82]. In the present study, the result of the EFA with Promax rotation suggested the same three-factor structure as found in the original English version [23]. Additionally, the CFA suggests that the three-factor structure had the best fit for our sample compared to the one-factor or two-factor structure obtained for the English version as indicated by a low SRMR and RMSEA and high CFI and TLI. This finding is similar to the reports of many validations conducted in different samples of individuals with chronic pain [39, 79, 80, 82]. On the contrary, other validations found the two-factor structure of the PCS to exhibit adequate model fit [48, 77]. In another vein, the Huijer et al.[34] found the one-factor, two-factor (based on the authors’ EFA), and Sullivan’s original three-factor structures all exhibited adequate fit to the Arabic population. However, it is important to note that the differences in the factor structure of the PCS across studies may be attributed to cultural differences in different countries.
According to the recommendations of the quality criteria for measurement properties of health status questionnaires [70], construct validity of a measurement is supported when at least 75% of the predefined hypotheses are verified. Based on our a priori hypotheses that the Hausa-PCS total scores would correlate moderately to strongly and significantly with that of the criterion variables, the concurrent validity was supported as 83% (5 out of 6) of the hypotheses were confirmed. The questionnaire demonstrated strong positive correlation with NPRS (rho = 0.74) comparable to that obtained for the Hindi version (rho = 0.65) [76] and higher than that (rho range = 0.19–0.52) reported by many other adapted versions [33, 36, 39–41, 43, 44, 82]. The moderate correlation obtained with the FABQ-physical activity (rho = 0.32) and FABQ-work (rho = 0.36) subscales were smaller compared to that obtained for the German (FABQ-physical activity; rho = 0.51 and FABQ-work; rho = 0.61) and Turkish (FABQ-physical activity; rho = 0.49 and FABQ-work; rho = 0.47) [48] versions but comparable to the Norwegian version [43] (FABQ-physical activity; rho = 0.34 and FABQ-work; rho = 0.25) except for the FABQ-work subscale which was found to be very low in the later version. Similarly, the moderate correlation obtained between our questionnaire and the ODI (rho = 0.35) coincides with the 0.35 obtained in the Hindi version [76] but slightly lower than the range of 0.40–0.57 obtained by other versions using the Roland Morris Disability Questionnaire [39, 40, 78]. This variation in correlations values across studies could be explained for the different questionnaires used in the assessment of functional disability. In another vein, the Hausa-PCS correlated weakly with the MCS-12 scores (rho = − 0.20) contrary to the Malay version which demonstrated moderate significant correlation with the MCS-12 (rho = − 0.38).
The result of the known-group validity of the Hausa-PCS revealed that the questionnaire and its subscales are not influenced by socio-demographic variables in terms of gender and habitation. Although this aspect of validity may require further investigation, it can be deducted based on the sample studied that male and female as well as urban and rural patients are likely to experiences the same level of pain catastrophization as a result of chronic LBP. In contrast, the Persian version demonstrates its ability to differentiate male and female patients with non-malignant musculoskeletal pain [72].
Regarding internal consistency, the Hausa-PCS total score exhibited adequate internal consistency (α = 0.84) consistent with the original English version (α = 0.87) [83] and the range of 0.84–0.93 reported by many validation studies [36–38, 40–43, 45, 84, 85]. However, we obtained lower alpha coefficients for the rumination (α = 0.69) and magnification (α = 0.41) subscales but sufficient for the helplessness subscale (α = 0.78). Consistent with our findings, most previous studies [37, 40, 43, 76, 82] found lower alpha coefficients for the magnification subscales, which could be attributed to the small number of items peculiar with the three-factor structure. It is important to note that increasing the number of scale items typically increases the Cronbach's alpha [86]. Thus, caution should be exercised when considering the magnification as independent subscale in computing catastrophization. Consequently, the two-factor structure of the PCS may be considered but warrants further investigation.
The test-retest reliability of the Hausa-PCS total score was highly adequate (ICC = 0.90) suggesting excellent reproducibility. Our value is higher than the original English version (ICC = 0.73) [23] and the range of 0.76–0.85 obtained by several language versions [33, 37, 39–43, 45, 47], consistent with the 0.90 obtained for the Afrikaans [35], Japanese [84], Nepali [77] and Xhosa [35] versions but slightly lower than the range of 0.92–0.97 obtained by other language versions [36, 38, 44, 76]. However, for the Hausa-PCS subscales, the ICC was only adequate for the helplessness subscale (ICC = 0.89). The magnification (ICC = 0.68) and rumination (ICC = 0.43) subscales had insufficient test-retest reliability which is consistent with the findings of previous validations demonstrating smaller ICC values for these subscales compared to the helplessness subscale [39–41]. These findings, thus, suggest that further investigation into the factorial structure of the Hausa-PCS may be useful.
The SEM and MDC at 95% CI were computed in this study to supplement the test-retest reliability since ICC does not account for the size of measurement error that is clinically meaningful [63]. The smaller the SEM the better the reliability (precision) of the measure whereas the smaller the MDC the more sensitive is the measure [87]. In the present study, the SEM (3.47) and MDC (9.62) values calculated for the Hausa-PCS total score were comparable to the values for the Afrikaans (SEM = 3.30; MDC = 9.00) [35] and Xhosa (SEM = 3.30; MDC = 9.30) [35]; lower than the values calculated for the Korean (SEM = 3.72; MDC = 10.3) [41], German (SEM = 4.6; MDC = 12.8) [39] or Norwegian (SEM = 4.60; MDC = 12.8) [43] versions; but higher than the values for the Hindi (SEM = 1.90; MDC = 5.26) [76] and Nepali (SEM = 2.52; MDC = 6.98) [77] versions. Compared to the SEM and MDC values of the Hausa-PCS total score, the three subscales of the questionnaire demonstrated lower values consistent with the reports of prior studies [35, 43, 76]. Regarding our SEM for the Hausa-PCS total score (3.47), it can be interpreted that if an individual has a baseline total score of 29, we can be 95% confident that the true score lies between 25.5 and 32.5. As for the MDC (9.62), a change of 9.7 can be considered as a true change in the total score above measurement error. Additionally, the result of the Bland-Altman plot for the Hausa-PCS total score showed minimal bias as the mean difference (0.87) calculated was close to zero, with LOA95% of − 8.10 to 9.75 which lies within the range of − 15.1 to 16.0 reported in the literature [43, 44, 76, 77]
One strength of this study is that the translation and cross-cultural adaptation was conducted as per the recommendation of standard guidelines outlined by Beaton et al.[64]. Additionally, the psychometric evaluation was conducted and reported in line with the COSMIN guidelines [73] even though we did not use the global rating of change scale to confirm the respondents’ stability for reliability assessments. However, one potential limitation of this study is that the correlations of Hausa-PCS with the criterion variables used were based on cross-sectional data. Thus, any causal conclusion concerning the influence of catastrophizing on pain, disability, fear-avoidance beliefs and mental health could not be drawn. Another potential limitation is that we were unable to evaluate responsiveness. Furthermore, divergent validity, which is another essential measure of construct validity, was not evaluated in the present study. Studies are needed to examine the causal relationships between Hausa-PCS and the aforementioned criterion measures in this population. Moreover, future researches evaluating responsiveness and divergent validity of the Hausa-PCS would be useful.
The Hausa-PCS was successfully developed and proved to be comprehensible, acceptable and psychometrically adequate in terms of factorial structure, construct validity, internal consistency and test-retest reliability when applied in mixed urban and rural patients with chronic LBP. However, the internal consistency and reliability coefficients (ICC) for the individual subscales are questionable, thus warranting further investigation. The tool can be used especially when considering the total score to evaluate pain catastrophizing for clinical or research purposes.
LBP: Low back pain; PCS: Pain Catastrophizing Scale; ODI: Oswestry Disability Index; FABQ: Fear-Avoidance Beliefs Questionnaire; SF-12: Short-form Health Survey; PCS-12: Physical component summary; MCS-12: Mental component summary; SD: Standard deviation; EFA: Exploratory factor analysis; CFA: Confirmatory factor analysis; CFI: Comparative fit index; TLI: Tucker-Lewis index; SRMR: Standardized root mean square residual; RMSEA: Root mean square error of approximation; ICC: Intraclass correlation coefficient; ANOVA: Analysis of variance; COSMIN: Consensus-based Standards for the selection of health Measurements Instruments; SEM: standard error of measurement; MDC: minimal detectable change; CI: confidence interval; LOA: limits of agreement.
This study was approved by the Health Research Ethics Committee of the Ministry of Health Kano State, Nigeria (Ref: MOH/Off/797/T.I./651). Written informed consent was obtained from all participants before their participation in the study.
The data that support the findings of this study are available from the corresponding author upon reasonable request.
The authors have no conflicts of interest to declare.
No funds were received in support of this work.
AAI and MOK conceptualized and design the study. AA and BK were responsible for data acquisition and analysis. AAI and MKO were responsible for drafting the final manuscript. BK was responsible for reviewing the final manuscript. All the authors read and approved the final manuscript.
The authors would like to acknowledge all the translators who translated the PCS into Hausa, the patients who participated in the study, and the physiotherapists who assisted in the validation process.