Cross-cultural adaptatiion and validation of the stroke specific quality of life 2.0 scale into Hausa language

Stroke is the leading cause of adult long-term disability and death in the world [1] with a worldwide population prevalence of 0.5% [2]. Two thirds of all deaths related to stroke worldwide occur in low and middle-income countries [3]. Continued advances in medical interventions have increased the survival rate of patients who suffer strokes which has in turn increased the number of stroke survivors with chronic disability. Out of the people who suffer from a stroke, 30% die, 30% are left functionally disabled and 40% have a successful recovery with minor to no disabilities [4]. Over 50% of the stroke survivors return to their homes after hospital discharge and the impacts of stroke are considerable on them and their informal caregivers [5]. They have to deal with the long term consequences of stroke, which were seen in physical, psychological and social areas of functioning [4].

Stroke survivors have impaired or decreased quality of life (QoL) on long term basis even among those who have no post-stroke disability [6, 7]. With ongoing rehabilitation, however, improvements in functional status are possible [8] and contribute to improved QoL for stroke survivors. Hence, significant interest has arisen in health-related quality of life (HRQoL) as a tool to assess changes in patient health throughout his lifetime. Assessment of outcomes of stroke rehabilitation should therefore include disability and QoL scales to assess the domains which are directly impacted by the disease [9].

Cross-cultural adaptation encompasses a process that looks at both translation and cultural adaptation issues in the process of preparing a questionnaire for use in another setting [10] and ensures the attainment of semantic, idiomatic, experiential and conceptual equivalence between the source and the target questionnaires. It is the deliberate modification of some features of a questionnaire to better fit a particular target population [11]. It is inappropriate to simply translate and use a questionnaire in another linguistic context [12, 13]. Studies may have a comprehensive linguistic translation process, but this still do not ensure construct validity and reliability [10, 13]. Most questionnaires that clinicians use were originally developed in the English language. Cross-cultural adaptation of existing English language questionnaires would enable comparisons of different populations and permit the exchange of information across cultural and linguistic barriers [14].

Patient-reported outcome measures have been used to supplement clinician-based outcome measures [15]. This is because information from the patients’ perspective on the consequences of disease and the therapeutic benefits is considered critical in the evaluation of health care. Various specific instruments for assessing post-stroke HRQoL assess domains relevant to stroke, such as vision or language. These, however, are not available in all languages. The Stroke-Specific Quality of Life Scale (SS-QoL 2.0) developed by Williams et al. [16] is one of the most comprehensive [17] and frequently used patient-reported stroke-specific outcome measure [6, 18, 19]. The language dimension and the patient approach used in the process of item development in the SS-QoL 2.0 scale also make it a scale of choice [6]. It comprises 49 items in 12 domains of self-care, vision, language, mobility, work, upper extremity function, thinking, personality mood, family roles, social roles and energy. The scale is s suitable for mild to moderate stroke of 1 to 6 months’ duration according to previous reports [20–22]. The SS-QoL 2.0 has been cross-culturally adapted to many languages. These include Spanish, German, Danish, Greek, Brazilian, Chinese, Turkish, Kannada (India), Persian, Malayalam (India), Hindi (India) and Yoruba.

Nigeria is a multilingual nation with Hausa, Igbo and Yoruba languages as the three major languages. Hausa is the Chadic language of Afro-asiatic descent with the largest number of speakers. It is spoken as first and second languages by an approximate total of 43 million people (National Population Commission) [23, 24]. Hausa language is spoken across southern Niger and Northern Nigeria and it has developed into lingua franca across most parts of West Africa for trading purposes. There are more than twelve Hausa dialects across the northern part of Nigeria with Kano dialect; Kananci as the standard [24].

The SS-QoL 2.0 has been cross-culturally adapted into Yoruba, the indigenous language of the people of South-western Nigeria using the guidelines of American Association of Orthopaedic Surgeons (AAOS) by Akinpelu et al. [25], and validated by Odetunde et al. [26]. The AAOS guidelines have been used by other authors to conduct cross-cultural adaptation of SS-QoL 2.0 to other languages. These include German [27], Danish [6], Brazilian [28], Spanish [29] and Dutch [30] among other languages. Non-availability of other Nigerian language versions of SS-QoL 2.0 other than the Yoruba version might have limited its utilization in Nigeria. The availability of psychometrically sound translated versions of standardized health measuring scales such as the SS-QoL 2.0 in indigenous Nigerian languages may promote the use of the scales in Nigeria and ensure that Nigerians who are not literate in English are not excluded from measurement of health variables that are important and meaningful to them. This study was therefore aimed at cross-cultural adaptation and validation of SS-QoL 2.0 into Hausa in order to promote its wide availability and utility all over the country and other places where the Hausa language is spoken.

Descriptive statistics of mean and standard deviation were used to analyze domains and overall scores on FHV. Convergent, discriminant and construct (comparing scores on FHV and English version of SS-QoL 2.0) validity of the FHV were tested by using the Spearman’s correlation method. Known groups validity by comparing domain scores with gender and age groups was tested using independent t-test and One-way ANOVA respectively. Intra-class correlation (ICC) was used to determine the test-retest reliability and Cronbach’s alpha values were determined for internal consistency of the FHV. SPSS version 16.0 (Chicago IL SPSS Inc.) was used to analyze data. Level of significance was set at p < 0.05.

Hausa is one of the three major and most common Nigerian languages spoken by 37%, of the population [23]. Hausa is also spoken by about 43 million people as a first and second language. The language is spoken in Niger, Nigeria, Benin, Ghana, Cameroon and Sudan [24]. The process of cross-cultural adaptation of SS-QoL 2.0 into Hausa language in this study was according to the guidelines recommended by American Association of Orthopaedic Surgeons (AAOS) [10]. In the forward translation of the SS-QoL 2.0 to Hausa language, some words such as ‘zip’ and ‘button’ which have no equivalent in Hausa language had to be used as ‘loan’ or ‘borrowed’ words. Zip was written as ‘zif’ in Hausa language. This was also the case in cross-cultural adaptation of SS-QoL 2.0 into Yoruba language in which the words ‘zip’ and ‘button’ were written as ‘siipu’ and ‘bọtini’ [26]. Back-translation of the Hausa version of SS-QoL 2.0 scale revealed discrepancies, inconsistencies and omission in twelve items/expressions. These discrepancies were resolved by the expert committee (details in Table 1).

This study assessed the construct, convergent, discriminant and known group validity, test-retest reliability and internal consistency of the FHV. The mean age of stroke survivors who participated in this study was 60.00 ± 12.13 years with median time since stroke of 12 months. The mean age of participants at onset of stroke was 58.33 ± 11.54 years. These findings support the earlier reports that stroke occurs more frequently in middle age [21, 22, 26, 28, 38–41]. Strokes can occur at any age, though its risk increases with age, [42] while stroke duration varies extensively among the survivors. This may be attributed to factors such as presence of co-morbidities, quality of care and access to care and rehabilitation, racial and ethnic disparities, perceived social support and psychological state among other factors [43]. Sex distribution of participants in this study also indicated a male preponderance, consistent with findings from previous studies [21, 22, 25, 27],[38, 39, 43]. The distribution of participants by hemiparetic sides was about equal for the right and left sides. Fifty-one percent of the participants had right hemiplegia/hemiparesis while 49% had left hemiplegia/hemiparesis. Left hemispheric ischaemic stroke presenting with right hemiplegia has been reported to be more frequent than right hemispheric cases [44]. The close proportion reported for right and left hemiplegia/hemiparesis reported in this study may be due to the sample size and the use of only two hospitals for data collection. A similar study by Odetunde et al. [26] involved 100 stroke survivors across 8 different hospitals.

Hausa version of the SS-QoL 2.0 demonstrated good to excellent test-retest reliability and high internal consistency or homogeneity of the domains in agreement with homogeneity of the domains of SS-QoL 2.0 scale reported by Hsueh et al. [45]. The findings from this study are further supported by findings from previous studies on different versions of SS-QoL 2.0 scale. Williams et al. [16] and Boosman et al. [21] reported excellent internal consistency for ischaemic stroke survivors and patients with aneurysmal subarachnoid haemorrhage respectively. Similarly, Kerber et al. [46] reported excellent test retest reliability and internal consistency for SS-QoL 2.0 scale in English speaking Mexican American population. The Danish version was reported to have moderate to excellent test retest reliability and excellent internal consistency by Muus et al. [20]. The Brazilian version of SS-QoL 2.0 was reported by Lima et al. [28] to have excellent test-retest reliability while the Spanish version of SS-QoL 2.0 showed good to excellent test retest reliability and internal consistency [47, 48]. Harkverdioglu and Khorshid [49] reported excellent test-retest reliability and internal consistency for Turkish version of SS-QoL 2.0, while the Chinese version of SS-QoL 2.0 demonstrated good to excellent internal consistency according to Wong et al. [50].

The Hausa version of the SS-QoL 2.0 demonstrated evidence of same construct as the English version. This is further supported by the finding that participants’ overall and almost all (ten out of twelve) domains scores on the English and Hausa versions of the SS-QoL 2.0 did not differ significantly. Differences or lack thereof could however be interpreted several ways, and imply that the adaptation was successful or not successful entirely based on how differences are interpreted (e.g. very similar scores therefore an excellent translation; or conversely, a failure to adequately capture differences). Coincidentally, mean domain scores were highest on mood domain and lowest on work domain on both versions of SS-QoL 2.0 scale. Our findings may imply that the FHV is a valid translation of the English version of the SS-QoL 2.0 and provides an evidence of construct validity of FHV. Similar findings were reported by Akinpelu et al. [25] and Odetunde et al. [26] for Yoruba version of SS-QoL 2.0 scale.

The result of test of known-group validity of the FHV indicated that there was no significant gender difference between the domains and overall scores. This is consistent with report of Xie et al. [43] and Zalihic et al. [51]. Male participants had higher mean score in vision, language, mobility, work, thinking and mood domains while female participants had higher mean score in self-care, upper extremity, personality, family role, social role and energy domains and the overall QoL score. Conversely, previous studies reported significantly lower post-stroke QoL in females and that females were more negatively affected in their QoL [1, 52, 53]. In these studies, QoL was assessed in different settings from the current study using different QoL scales, mode of administration and stroke patients, and these factors may be responsible for the observed differences in QoL and gender. Nevertheless, the finding of this study should be treated with caution as it might be underpowered to detect a difference in domain scores based on gender. Given the difference scores between males and females on some domains like SR which may be interpreted as inconclusive.

Our study also showed that the domains of FHV were associated with age. The mean domain scores differ significantly across the four age groups on self-care, vision, language, mobility, upper extremity and overall score. Significant influence of age on QoL scores in the self-care, mobility and upper extremity functions domains is not surprising, as these domains are the underlying component of the physical domain sub-scales of SS-QoL2.0. Left hemispheric stroke which manifests as right hemiparesis/hemiplegia would most likely cause communication difficulties and may explain the reason for significant influence of QoL scores on Language domain of FHV. This is similar to the report of Nichols-Larsen et al. [54]. Significant difference across age groups and vision domain may be explained in terms of visual impairments that are sometimes associated with stroke, effect of which might not have been overcome by the stroke-survivors. In a like manner, post-stroke difficulties in visual function has been reported to cause significant impact to the quality of life of stroke survivors [55]. The reported differences in the QoL scores among participants is not significant on the remaining 7 domains of FHV but has demonstrated that age of stroke survivors was an important factor that determines their HRQoL [54, 56–58].

All items in the FHV demonstrated strong correlations with its own domain than with domains measuring other concepts. For example, item M7 ‘Did you have trouble with needing to stop and rest when walking or using a wheelchair?’ had r value of 0.969 with its own domain, while the same item had r value of 0.247 with thinking domain. This finding of strong correlations of items with its own domain is consistent with reports of Muus et al. [20] and Cruz-Cruz et al. [43] which describes the relevance of the items in their respective domains and the discriminant validity of the FHV.

The WHOQoL-BREF was the only Hausa quality of life scale available to the researcher as at the time of data collection for this study. The finding of moderate to good correlation of FHV with all domains of Hausa version of WHOQoL-BREF indicates that the overall score of the FHV represents the overall status of HRQoL and is consistent with the report of Hsueh et al. [46] and Williams et al. [16]. This link between overall SS-QoL score and the overall status of HRQoL of stroke survivors is useful in capturing the multiple impacts of stroke from the patients’ point of view. Correlation between similar domains on the two scales can be explained in terms of correlation between related domains where in most cases domains of SS-QoL 2.0 are components of domains of WHOQoL-BREF. Boosman et al. [21] reported similar findings of significant correlation between the Cognitive Failures Questionnaire (CFQ), Life Satisfaction-9 (LiSat-9) and Hospital Anxiety Depression Scale (HADS) and corresponding SS-QoL 2.0 domain scores. The SS-QoL 2.0 scale has been compared with various other measures with varying findings. Muus et al. [20] reported moderate to excellent correlation (r ranging from 0.37 to 0.88) between the domains of Danish version of SS-QoL 2.0 and the Barthel Index and National Institute of Health Stroke Scale. Similar to the findings of the present study, Boosman et al. [21] reported that SS-QoL 2.0 scores showed weak to moderate correlations (0.24–0.32) with the Glasgow Outcome Scale and moderate to strong correlations (0.35–0.72) between SS-QoL 2.0 scores and CFQ, LiSat-9 and HADS. Lin et al. [22] also reported weak to moderate correlations (r ranging from − 0.04 to 0.52) between the domains of the SS-QoL 2.0 and Fugl-Meyer Assessment, Functional Independent Measure, and Frenchay Activities Index with excellent correlation (r = 0.65) only between the SS-QoL 2.0 Self-Care domain and Functional Independent Measure. The findings in the present study are consistent with findings in the reported studies. The fact that the correlations between similar domains of Hausa versions of SS-QoL 2.0 and WHOQoL-BREF are statistically significant may be evidence that the Hausa version of SS-QoL 2.0 scale capture relevant QoL items which are also assessed by the WHOQoL-BREF. Low correlation coefficient between some similar domains on the WHOQoL-BREF and SS-QoL 2.0 may be due to certain factors. Previous studies that compared the SS-QoL with generic scales used more than one scale for comparison such that most related domains on the SS-QoL and the other scales were compared [20, 27, 48, 49, 59]. In the current study however, the domains of WHOQoL-BREF only were used to compare the domains of SS-QoL 2.0 and this may not allow for comparison with most related domains as obtained in previous studies. The process of social adaptation has been reported to be a possible factor that can cancel out supposed differences in QoL. This fact can also explain in like manner the apparent weak correlation between the generic measure, WHOQoL-BREF and a specific measure SS-QoL 2.0 [20, 49]. The items in each domain of SS-QoL 2.0 were structured in a specific and simplified manner that addresses stroke-related problems, but this is not so for the generic WHOQoL-BREF. The differences in the contents of the two scales may also explain the weak correlations between these compared domains. The FHV 2.0 demonstrated no significant floor or ceiling effects on any of its domains or the overall scores. Absence of floor effect implies that FHV can discriminate between varying degrees of stroke severity and can identify stroke survivors with severe stroke. Absence of ceiling effect implies that the scale can also identify any level of improvements in condition of stroke survivors in any of the QoL domains. This is in line with findings of Williams et al. [16] and Cruz-Cruz et al. [43].

This study has some limitations that may need to be considered in interpreting and generalizing its findings. First, with regards to the revision made to item SC1, preparing food and going to the market to buy foodstuffs are not equivalent. They are different activities requiring different amounts of effort or exertion. This revision was however done to make the item applicable to both male and female participants. Future study should consider what else could be done to address this very interesting cultural issue/gender difference. Second, participants were 20.0 ± 20.4 month post-stroke. This suggests that the results may have limited generalizability to patients early in the stroke recovery, patients at different phases of post-stroke recovery may have different HRQoL issues and it is not clear that the FHV has demonstrated that validity. Future research is recommended to address this limitation. Third, participants in this study received different rehabilitation programs throughout the duration of the study; further research is needed to assess the psychometric properties of the FHV for specific treatment programs on larger samples to provide further insights into the psychometric properties of the FHV in particular situations. Fourth, there are possible differences in psychometric properties in patient-reported QoL outcomes due to the modes of administration [60], thus further research may be needed to study psychometric properties of the FHV using different modes of administration such as paper-and-pencil administration at home, via the mail and telephone interview. Lastly, in this study, the FHV was validated using the Hausa version of WHOQoL-BREF, other Hausa outcome measures with the advantage of assessing domains relevant to stroke, should be used to compare the domains of FHV in order to further ascertain its validity with other established outcome measures.

The English version of SS-QoL 2.0 was successfully cross-culturally adapted to Hausa language. The FHV captured the concept of interest as the English version. Test-retest reliability, internal consistency, discriminant, construct, convergent and known-group validity of the FHV are adequate. Hausa version of the SS-QoL 2.0 is therefore a valid and reliable measure with no floor or ceiling effects and recommended for assessing health-related quality of life among Hausa stroke survivors.