The aim of this study was to cross-culturally adapt and validate the PCS for use in English-, Afrikaans- and Xhosa-speaking patients with fibromyalgia living in the western parts of South Africa. Modifications to the wording of the items and scoring system were required to ensure that the PCS would be applicable within a South African context. The study results show that the English, Afrikaans and Xhosa SA-PCS are valid and reliable tools for administration within the public health sector among South African patients with fibromyalgia living in the Western Cape (South Africa).
As anticipated, cross-cultural adaptation and validation of the PCS for the South African context was complex because of the cultural and linguistic variability evident in various areas of this diverse country. Although the SA-PCS was cross-culturally adapted and validated in three of the most predominant languages of the western part of South Africa, the translated and validated instruments resulting from this study may not be applicable to Zulu- or Sesotho-speaking patients with fibromyalgia living in the northern and eastern parts of South Africa. The instrument adaptations from this study may also not be applicable for, or accepted by, other English-, Afrikaans- and Xhosa-speaking ethnic and culture groups living in the other parts of South Africa. Careful consideration for diversity is therefore required when applying any health outcome measure among various languages, cultural or ethnic groups uniquely found in South Africa. Further cross-cultural adaptation and validation of the SA-PCS in other South African language and ethnic groups is therefore recommended.
Face and content validation identified that the English, Afrikaans and Xhosa SA-PCS were acceptable, applicable and easily comprehended by the included subjects. The English, Afrikaans and Xhosa SA-PCS were also simple to complete as it took subjects less than five minutes to complete the questionnaire. Furthermore, the adaptation and application of the scoring system proved to be easier for the subjects to understand what was expected. We anticipated that complicated scoring systems for outcome measures would be ineffective if the target group does not fully understand how the system works, and how the system should be applied. As a result, incorrect responses and inaccurate study results and conclusions may be obtained. In South Africa, clinicians working in the public health sector particularly have limited time to consult with individual patients due to limited resources and staff. The time taken and the ease of completing a questionnaire, in addition to the cultural applicability of an outcome measure, therefore need to be considered. Accurate measurement is essential before, after and during all management programs for determining the progress of management and the effectiveness of a treatment. It is as important to ensure that acquiring these measures from patients is not frustrating for the health professional and the patient and that clinicians do not neglect assessing outcomes on a regular basis due to time constraints. The English, Afrikaans and Xhosa SA-PCS are therefore simple, efficient, easy to understand, easy to complete and valid tools to use among South African patients with fibromyalgia receiving services the public health sector in the Western Cape (South Africa). Further validation is however required for application of the English, Afrikaans and Xhosa SA-PCS in the South African private sector and in research studies.
Internal consistency for the English, Afrikaans and Xhosa SA-PCS, as a whole, was excellent (α = 0.98, 0.98 and 0.97 respectively). These estimates are higher than the original English PCS (α = 0.87) as well as the Spanish PCS (α = 0.79), Dutch PCS (α = 0.85), French PCS (α = 0.85); Singalese PCS (α = 0.89), Catalan PCS (α = 0.89); Italian (α = 0.92), Chinese PCS (α = 0.93) and German PCS (α = 0.94)[7, 34–42]. However, the evaluation of internal consistency of the entire PCS is theoretically incorrect since, by definition, Cronbach’s alpha “indicates the correlation among items that measure one single construct”. The PCS contains three dimensions; hence evaluation of the internal consistency of each of the three subsections is required. The internal consistency for all subsections (rumination, helplessness and magnification) of the SA-PCS was also found to be excellent and considerably higher than previously reported ICC’s for subsections of the PCS. Nevertheless, the high internal consistency for the subsection magnification of the English, Afrikaans and Xhosa SA-PCS (α = 0.96; 0.99 and 0.93 respectively) found in this study is contradictory to the majority of validation studies which have previously reported that internal consistency for the magnification subsection, in particular, is usually unsatisfactory[7, 34, 37, 38, 40, 41]. The internal consistency reported for the subsection magnification of the original, French, French-Canadian, Catalan, Italian and German PCS ranged between α = 0.56 to 0.67. It has been postulated that the low internal consistency found for the subsection magnification may relate to the few items contained to this subsection and that it should be reconsidered if this subsection can be reliably used as an independent instrument. The higher internal consistency reported in this study may be due to the fact that subjects took more time to answer each question and may have considered each question more carefully, increasing the internal consistency for this subsection. The Chinese PCS reported an internal consistency of α = 0.77, which was closest to that of this study. Further validation of the psychometric properties of the English, Afrikaans and Xhosa SA-PCS among larger sample groups is however warranted.
The SA-PCS showed excellent stability (test-retest reliability) as a whole with no significant difference between test and retest scores, for one month correlation: English SA-PCS (ICC = 0.90), Afrikaans SA-PCS (ICC = 0.91) and Xhosa SA-PCS (ICC = 0.89) (Table 4). These results were higher than the original English (ICC = 0.73), French (ICC = 0.73) and the Catalan versions (ICC = 0.76); were comparable to the Spanish (ICC = 0.84), German (ICC = 0.83), Italian (ICC = 0.84) and French-Canadian (ICC = 0.85) versions; but were lower than those reported for the Dutch (ICC = 0.92) and the Chinese (ICC = 0.96) versions of the PCS[7, 34, 36–42]. Excellent stability was also found for the subsections (rumination, helplessness and magnification) of each version of the SA-PCS with no significant differences between test and retest scores, which is analogous with previously reported ICCs for the subsections of the PCS[7, 34, 36–42]. The ICC values obtained for an outcome measure is however largely dependent on variance of disease patterns between subjects and it is acknowledged that the time period between the test and retest influences the size of this variance[48, 57]. The longer the time period between the test and retest, the more likely variance between subjects may occur and the lower the ICC value. Conversely, if the period between the test and retest is too short, there is a possibility that recall bias may occur, resulting in a higher test-retest correlation. In chronic pain studies, there is also a good possibility that the results obtained will differ between individuals who are experiencing pain or symptoms at the time of the testing and those who are symptom-free. According to Lamé et al. (2008), the latter group often responds to questions by trying to remember how they feel when they are actually experiencing pain or symptoms than what they are feeling at the time of inquiry. However, due to the chronicity of fibromyalgia, rapid changes in general health, pain/symptom patterns and disability are usually not expected, and the timing and experiences of pain and symptoms naturally vary. The period between the test and retest should therefore be based on the usual clinical practice of most outpatient public health facilities in South Africa where patients often have to wait a few weeks to months between treatments[48, 57]. Since the included subjects with fibromyalgia were believed to not vary significantly in general health, pain/symptom patterns or disability within a short time period, the one month period used between the test and retest was deemed appropriate to ascertain reproducibility of the English, Afrikaans and Xhosa SA-PCS.
Sensitivity-to-change was also satisfactorily demonstrated in the English, Afrikaans and Xhosa SA-PCS. At a 95% confidence level, the MDC of the English, Afrikaans and Xhosa SA-PCS, as a whole, indicates that a change of more than 8.8, 9.0 and 9.3 points after a given intervention, respectively, would not be due to measurement error. These values are slightly lower than those reported for the Italian (10.5) and German PCS (12.8)[34, 37].
It is acknowledged that a major limitation to this study was that the SA-PCS should have been correlated with a ‘gold standard’. However, since no such measure currently exists for pain catastrophization, the PCS is usually correlated with related outcome measures such as intensity of pain/symptoms, disability, fear-avoidance behaviours or depression, to establish cross-sectional convergent validity. However, a number of related outcome measures had to be cross-culturally validated prior to their use within the South African context. The cross-cultural adaptation and validation of the PCS, TSK and FIQR for a South African fibromyalgia population was therefore concurrently conducted. It was therefore decided that for this study it would be appropriate to correlate the scores of the English, Afrikaans and Xhosa SA-PCS to the English, Afrikaans and Xhosa versions of the TSK and FIQR. Nevertheless, criticism may be justified as to why the SA-PCS was not correlated with intensity of pain/symptoms, such as the Numerical Pain Rating Scale or the Visual Analogue Scale which are widely-accepted valid outcome measures. Instead, in the current study severity of pain/symptoms was measured and subjects were required to report on the everyday activities which increased their pain and symptoms. To defend our research approach, at the time of conceptualizing the study, it was understood that in chronic pain patients, severity of pain and symptoms, and activities which increase pain and symptoms, are of more use as this potentially reflects the patients’ perceptions of his/her conditions, rather than quantifies pain and symptoms which may not always be present or present at the time of testing. Scores obtained from the SA-PCS, can therefore be related to the severity of the subject’s pain and symptoms, and to the difficulty subjects may experience in performing daily activities. The relationship between catastrophization, avoidance of a particular activity and the influence fibromyalgia has on an individual’s life, may therefore be more natural determined than momentarily-based experiences. The results of this study show that the SA-PCS related with pain severity, fear-avoidance behaviours and impact of fibromyalgia in an expected manner. However, the results for the concurrent validation of the SA-PCS in relation to the SA-TSK and SA-FIQR in this study should therefore be viewed with caution, as the SA-TSK and SA-FIQR were validated at the same time as the SA-PCS.
Another limitation to this study is that the SA-PCS was validated among English, Afrikaans and Xhosa-speaking patients with fibromyalgia living in and around the western parts of South Africa and not the entire South Africa. For this reason, the current validated SA-PCS may not be applicable for patients with fibromyalgia residing in other parts of South Africa due to the vast differences in cultures and languages between various provinces of South Africa. Further cultural and linguistic validation of the SA-PCS for other areas of South Africa is therefore required.
Factor analysis was not performed and may be deemed as a limitation. However, according to DeCoster (1998) the objectives for doing exploratory and confirmatory analysis may not actually apply to this study. Basically, exploratory factor analysis should be used when one is interested in making statements about the factors that are responsible for a set of observed responses. The primary objectives of an exploratory factor analysis are to determine: 1) the number of common factors influencing a set of measures; and 2) the strength of the relationship between each factor and each observed measure. Some common uses of exploratory factor analysis are to: 1) identify the nature of the constructs underlying responses in a specific content area; 2) determine what sets of items \hang together" in a questionnaire; 3) demonstrate the dimensionality of a measurement scale. Researchers often wish to develop scales that respond to a single characteristic; 4) determine what features are most important when classifying a group of items; and 5) generate\factor scores representing values of the underlying constructs for use in other analyses. Since this was not the case in this study, exploratory factor analysis was not conducted. On the other hand, confirmatory factor analysis should be used when one has large numbers of data. The primary objective of a confirmatory factor analysis is to determine the ability of a predetermined factor model to an observed set of data. Some common uses of confirmatory factor analysis are to: 1) establish the validity of a single factor model; 2) compare the ability of two different models to account for the same set of data; 3) test the significance of a specific factor loading; 4) test the relationship between two or more factor loadings; and 5) test whether a set of factors are correlated or uncorrelated. Since items were not added or removed, the factors within the outcome measures essentially remained the same, hence factor analysis was not deemed necessary in this study.
Lastly, there are always queries regarding the appropriate sample size for validation studies, particularly when it is inappropriate to estimate the extent of cultural differences on instrument construction. Therefore, despite the encouraging results of this study, further validation of the SA-PCS should include larger samples per language and ethnic group, and chronic pain conditions other than fibromyalgia.