Validation of the Indonesian version of multiple sclerosis quality of life-54 (MSQOL-54 INA) questionnaire

Inclusion criteria were (1) MS patients fulfilling the 2010 McDonalds diagnosis criteria, (2) age > 18 years old, able to read and write in Indonesian; (3) providing written consent to be recruited into the study. Exclusion criteria were subjects undergoing relapse in the previous month or having psychiatric disorders or other chronic diseases (diabetes mellitus, hypertension, etc.).

The QOL in this study was assessed using MSQOL-54 questionnaire consisting of fifty-four questions. This questionnaire comprised SF-36, the instrument for general QOL assessment, and the additional eighteen specific questions for MS patients. The examination result contained two domains of composite scores, reflecting the physical health (PH) and mental health (MH), independent to each other. The separation of domains and components were adapted from the original MSQOL-54 version proposed by Vickrey et al. [14].

Due to the availability of the validated Indonesian version of the SF-36 questionnaire, translation and transcultural adaptation processes were only performed on eighteen specific questions for MS patients [7, 21]. Standardized forward-backward translation method was performed for the production of the MSQOL-54 INA. The steps of the translation process were (1) forward translation of the original MSQOL-54 questionnaire to Indonesian by two groups of professional translators blinded to each other; (2) reconciliation of the translated questionnaire by a panel team, consisting of an Indonesian neurologist and psychiatrist expert in their field, to compose conceptually and semantically similar questionnaire with the original version; (3) back-translation of reconciled questionnaire to English by two different translators blinded to each other; (4) reconciliation of the translated result to produce the targeted version. The panel team then created agreement regarding the culture of the source and targeted instrument following four aspects, which is semantic, idiomatic, daily experience, and conceptual. This step produced a questionnaire which was then tested to MS patients (expert committee); (5) the reconciled questionnaire was then tested to 10 MS subjects consecutively to gain inputs from the subjects’ perception regarding the grammatical clarity, easiness to be understood, and any suggested alternate words during the test. (6) composing the last version of the questionnaire [17, 22, 23].

The samples meeting the inclusion criteria were asked to fill the MSQOL-54 INA questionnaire directly attended by medical staff as necessary. During the process, the time needed to complete the questionnaire was documented as well as whether the patient needed help in reading and writing the answers were documented. Simultaneously, Expanded Disability Status Scale (EDSS) score was measured to assess the patient’s disability level.

Psychometric analysis for MSQOL-54 INA questionnaire was performed, consisting of reliability, internal validity, and external validity tests. Reliability of internal consistency was measured by using the Cronbach alpha coefficient (α) with a range of 0–1. The higher the result, the better the reliability. An instrument was defined to have consistent internal reliability if α was ≥0.7 [7, 24]. The Cronbach alpha value from the original research by Vickrey et al. was also used as a comparison.

Internal validity was assessed with content validity and construct validity. Content validity had been performed during the transcultural adaptation process by the neurologist and the psychiatrist. Construct validity was performed by executing computation of correlation coefficients, which was by measuring the correlation coefficients of each domain and its original domain (item internal consistency (IIC)) compared with the opposing domain (item discriminant validity (IDV). The IIC of > 0,4 and higher than the IDV proved the construct validity. In addition, the principal component analysis using direct Oblimin method was performed to assess whether the separation of domains in the original study by Vickrey et al. suited the condition in Indonesia.

External validity was assessed by measuring the association of each component and the composite scores with EDSS score. The score of each dimension and domain of MSQOL-54 was expected to be negatively correlated with EDSS. Pearson (r) and Spearman (p) correlation coefficients were used for measuring the association of composite scores with the clinical and demographic variables [7].

Almost all patients could complete the questionnaire within the recommended range of 11–18 min [14], which described that the questions were easy to understand and manageable. The missing data in this study could be minimized as little as possible with the presence of clinicians who could help the subjects as needed during the test. However, missing data were found in sexual function and satisfaction with sexual function components since only those who were married were willing to answer those components. This condition was closely related to local cultural perception which believed that sexual matters were taboo, especially for unmarried individuals. For the married individuals, the questions regarding sexual were distressing, too private or sensitive to be answered. This statement was in conjunction with the good result of sexual function and satisfaction with sexual function components in this study.

In general, this study found high reliability value, which showed that this questionnaire was internally consistent. The low Cronbach α scores on health perception and social function components could result from the less number of questions assessing those components; for example, the social function component consists of only three questions.

From the internal validity assessment, satisfactory results were obtained. Observing the internal validity using principal component analysis, there were several discrepancies between the domain analysis and the proposed domain by Vickrey, et al. These findings could still not be explained in this study, but a larger sample was still needed to observe the consistencies of these findings as these two components were those having the least reliability score.

Assessing the external validation using EDSS scores, the negative correlation was found at almost all components, except for sexual function and satisfaction with sexual function components. This might be due to the perception of MS patients’ population in Indonesia who perceived questions regarding sexual problems as taboo, leading to potentially unreported issues. Physical health-related questions were more statistically significant in correlation with EDSS since EDSS was the most common clinical assessment used for the physical disability assessment of MS patients [5, 25]. Similar results were also shown by the previous study by Brunet et al. which found that the disability assessed with EDSS only had the correlation with physical function domain at RAND-36 questionnaire [26].

In accordance with the studies by Idiman E et al. in Turkish [16], Catic T et al. in Bosnia and Herzegovina [18], Pekmezovic T et al. in Serbian [19], Stern B et al. in Slovenian [20], and Solari A et al. in Italian [15], the questions regarding sexual function and satisfaction with sexual function were the most potential missing data. This problem arose due to the feeling of embarrassment in admitting the issues regarding sexual, which is in accordance with the respective culture. However, different from the other studies, this study could not collect many subjects to fill this part due to the small MS population in Indonesia, most of which were still unmarried.

Compared to the original study conducted by Vickrey et al. in the United States, there were similarities that health perception and social function domain had the lowest Cronbach α [14]. However, the value in this study was less than 0.7, which was similar with the studies conducted in Serbia, Slovenian, Israel, and Iran [5, 19, 20, 27]. Stern et al. explained that the low Cronbach α value at health perception component might be caused by the nature of the questions which included wide and unspecific health evaluation aspects which led to relatively lower consistency [20].

Regarding the analysis of IIC and IDV, this study found that energy and role limitation due to emotional problems component were more suitable to show MH domain rather than the PH domain. These results were also reported by previous psychometric studies [5, 20, 27]. This described that in those population, as well as in Indonesian MS population, low energy score was more perceived as mental problems whereas role limitation due to emotional problems was more perceived as physical problems in conjunction with role limitation due to physical problems.

In view of the analysis of principal component analysis, there was only a study focusing on this analysis. In this study by Stern B, et al. in Slovenian [20], the discrepancies between domains and components were only observed in energy and role limitation due to emotional problems. In contrast, besides those two components, the component of health perception and social function was also assigned into different domain compared with the original domain proposed by Vickrey et al. [14]. This result was not found on a similar study by Stern B et al. The small sample size and the lower reliability number may explain this discrepancy, but further research was needed.

The study of Solari A et al. in Italia, Idiman E, et al. in Turkish, and Catic C, et al. in Bosnia supported the correlation of EDSS and PH composite score in this study [15, 16, 18]. These studies also found that EDSS was negatively correlated with the PH composite score, which supported the fact that EDSS described the health in physical aspects.

In this study, the correlation between composite scores and other clinical and demographic factors, such as age, duration of disease, and the number of relapses in the past one year did not show significant correlation. Similar results were found in a study in Turkey in which there was no significant correlation between composite scores and age, gender, education level, marital status, and health insurance [16].

This study provided a readily-used of MSQOL-54 INA which can be implemented for the Indonesian population. The study results also proved that this instrument was valid and reliable, which can be trusted by potential users. In addition, this study also provided psychometric properties of MSQOL-54 INA questionnaire in Indonesia MS population. However, this study had a small sample size, which may impact on the reliability score for some components. This limitation was inevitable because this disease was still rare even though the study was conducted in the national referral hospital in Indonesia. Additionally, confirmatory factor analysis was not applied in this study. Similar studies with more subjects in the future were needed to prove the consistency of this study result.

Another limitation of this study was associated with the content validity process, in which an independent third party did not supervise the translation process as proposed by Acquardro C et al. [17].

This is the first study focusing on the specific QOL instrument for MS patients. This instrument can generally be used for Indonesian MS patients with good validity and reliability. Besides, this validated questionnaire may become one of the tools to evaluate the success of the therapy as well as the baseline for the clinician to focus not only on the patients’ clinical aspects but also on the medical rehabilitation, cognition, and psychiatric issues.

This study had produced an adapted MSQOL-54 INA questionnaire with well-considered content validity and high reliability in the aspect of internal consistency on most components. Further studies with larger sample size are needed to prove the internal consistency of some components, particularly health perception and social function that had lower Cronbach alpha value and sexual function that had a smaller sample size. In addition, this instrument can also be used as a measure for assessing the success of therapy, which has not been analyzed in this study. Further additional analysis with confirmatory analysis method was also recommended to be performed with larger sample size.