Introduction

The World Health Organization (WHO) defines quality of life (QoL) as an “individual’s perception of their position in life in the context of their culture and value system in which they live, and in relation to their goals, expectations, standards, and concerns”[1]. This definition underscores the subjective nature of QoL as it reflects individuals’ overall perception of and satisfaction with their different experiences. Many have noted the difficulties associated with interpreting the subjective meaning of QoL [2,3,4,5]. The WHO definition further implies that meaning and levels of QoL are likely to be different across the globe. Research has shown, for example, significant associations between QoL and societal conditions [6].

The consequences of spinal cord injury or disease (SCI/D) include issues of functional capacity, personal and social relationships, financial loss, loss of bowel and bladder function, sexual function, chronic pain, and lack of accessible environments [7]. Unsurprisingly, these issues impact QoL for individuals with SCI/D [8, 9]. International comparative studies on QoL among individuals with SCI/D are sparse but show converging results, suggesting similarities in what makes life meaningful that transcend national boundaries [10]. Yet, differences in QoL between countries that could not be completely explained by differences in demographic and SCI/D characteristics point to the potential impact of social and cultural differences [11, 12].

The demand is increasing for QoL research of cross-national comparisons or aggregation of data across societies [13]. Such research is important to inform the development of national policies and international principles to deter stigmatization and discrimination of people with SCI/D. Therefore, it is important to ensure that QoL measures developed in one society are equally valid for use in others [14]. Subjective concepts, such as QoL, require similar understanding and interpretation of meaning to ensure comparability of results. The process of achieving cross-cultural validation is often complex and includes congruency of conceptualization and item content, accuracy in translation, and establishment of key metrics such as reliability, validity, and responsiveness to change [15].

Conceptual equivalence (CE) is a key concept in assessing cross-cultural validity of QoL measures [16]. CE refers to whether or not a given domain has similar meaning and salience across different cultures. For instance, emotional and relational domains may have a higher salience for overall QoL in one country than another [17]. This issue is particularly relevant to the International Spinal Cord Injury Quality of Life Basic DataSet (SCI QoL-BDS), developed with the purpose of standardizing the collection and reporting of a minimal amount of information necessary to merge and compare results of studies on QoL across the globe [18]. Preliminary evidence suggests cross-cultural validity of the SCI QoL-BDS [10], but the equivalence of meaning of the SCI QoL-BDS is still in need of research. This study was designed to address cross-cultural CE, while examining response differences across international sites. The purpose of this article is to describe the results of this inquiry.

Methods

Design

This study utilized a mixed methods approach guided primarily by qualitative methods and using cognitive interviews (CI). CI are used to improve reliability and validity of new measures by showing how respondents interpret concepts and understand the survey questions they are asked [19]. CI are routinely administered as part of questionnaire design, piloting and refinement, and can be particularly helpful when attempting to measure abstract concepts like QoL [20]. Quantitative analyses were complementary. The study was conducted in 2017–2019 as part of a larger validation of the SCI QoL-BDS across five sites across four countries: Ann Arbor and Denver (United States), Melbourne (Australia), São Paulo (Brazil), and Utrecht (the Netherlands) [21]. Each of the five sites recruited 8–12 participants living in the community utilizing a convenience sampling approach. Eligibility criteria included diagnosis of SCI/D, minimum of 18 years of age, and completion of informed consent. Exclusion criterion was the inability to complete the interview due to cognitive or psychiatric limitations. Trained research interviewers assessed participants’ cognitive ability during the recruitment and consent process to assure ability to understand instructions, attend to the questions, and their ability to express their thoughts accordingly. Participants were recruited mainly through affiliated hospitals and clinics at the respective sites. Medical information was made available by these facilities.

CI were conducted to examine similarities in the meaning of QoL and the interpretation of the SCI QoL-BDS questions across sites [22]. CI is a semi-structured method for determining how people understand and respond to questions [23,24,25,26]. CI were conducted in person or via phone. Interviews lasted 30–40 min, were audio recorded, and transcribed. Interviews were conducted in the native languages. Responses to the research questions were transcribed and, for the São Paulo and Utrecht sites, translated into English by authors who are fluent in both English and their relevant language. Researchers from each site participated in the design of the interviews and training on qualitative methods used for CI.

Measure

Participants were asked the opening question, “How would you define the concept of QoL?” to elicit their definitions of QoL. This question was followed by the three items of the SCI QoL-BDS. The three items are: (1) Thinking about your own life and personal circumstances, how satisfied are you with your life as a whole during the past 4 weeks? (2) How satisfied are you with your physical health in the past 4 weeks? (3) How satisfied are you with your psychological health, emotions, and mood during the past 4 weeks? All items were rated on a scale of 0 (completely dissatisfied) to 10 (completely satisfied). Each of these questions was followed by semi-standard probes based on the context of respondents’ answers in order to gain a better understanding of their perspectives. This process allowed for determination of whether the SCI QoL-BDS items made sense to respondents, if the items were relevant and understood consistently, and if there were circumstances under which it was difficult to respond to them appropriately. These follow-up discussions allowed respondents to identify key areas within the QoL construct that impacted their QoL, providing detailed contexts to their numerical responses. Lastly, they were asked if there were any other issues that were not mentioned by the items that influenced their QoL.

The SCI QoL-BDS was translated using back translation methods recommended by the International DataSets Committee using two native speakers of the required languages (Dutch and Brazilian Portuguese) and English [21, 27, 28]. Translations were adjusted to be contextual and idiomatically meaningful to each site and language.

Analyses

Thematic content analysis was used to uncover patterns within the data, following established approaches in thematic analysis [29, 30]. Four principle authors (EJ Rohn, MWM Post, AJ Hakbijl, and DG Tate) read the interview transcripts to gain an overall sense of the content and quality of the data. These four authors independently identified and coded topics in the data using NVivo, a qualitative data analysis software (NVivo 12) [31]. This software helps organize and find insights in qualitative data using coding functions that link text to codes, facilitating the identification of patterns within the data. Inductive open coding was used to allow meaning to emerge from the participants’ own words. Redundant topic codes were collapsed together, resulting in 1067 discrete topics across all the text data. Topics with similar content or expressed ideas were combined further into 39 larger categories of related topics. Of these topical categories, 17 included by far the most topics (912 topics in the top 17; 155 topics across the bottom 22), and were treated as qualitative themes [32, 33]. These themes represented repeated, patterned expressions across multiple respondents, suggesting salient, self-defined feelings, experiences, and values therein [34, 35]. Consensus was reached at each step by teams of two principle authors working together, reviewing the work of the others on the same task, and followed by repeated discussions among all four principle authors [36].

In order to check the frequencies and distributions of themes, NVivo was used to identify the relative percentages of where the text coded at each of the 17 themes intersected with each site and each question item. This percentage represents the amount of text dedicated to discussions of all the topics combined in that theme, across each of the 17 themes. We grouped these themes further by broad components of participants’ lives (i.e., physical, environmental, social, personal, and emotional factors) into theme categories to facilitate CE comparisons of broadly related ideas [37]. The frequencies of the themes and theme categories in response to each question (i.e., QoL definition and the three SCI QoL-BDS items) were treated as indicators of salience and the participants’ understanding of these questions, allowing comparisons across sites. CE was suggested in cases where similar percentages of themes were found in response to an item, across sites, by each of the items in the measure. The senior authors held additional lengthy consensus-building discussions to confirm patterns in and across the theme categories, explore meaning in those responses, and examine CE. Those interpretations most strongly supported by the data were deemed credible and were included [32, 36]. As a last step, we examined differences in theme distribution in relation to participant demographic and injury characteristics, to examine whether differences in CE were explainable by these factors.

Quantitative analyses included calculating descriptive statistics for the sample. To assess differences in participants’ characteristics as a function of site, queries compared the five sites with analyses by age groups (<50; 50 + ), non-traumatic vs. traumatic SCI/D, level of injury (paraplegia; tetraplegia), and time since injury (<10 years; 10+ years). Chi-square tests, ANOVAs, or independent t-tests were used, and significance was assessed using a p value < 0.05.

Results

Demographic and SCI/D characteristics

The sample included 51 participants, with the number per site ranging from 8 to 12. Most participants were male (60.8%). The mean age was 55.8 and the mean time since injury was 14.2 years. Of the total sample, 64% had traumatic injuries and 60.4% had paraplegia (see Table 1).

Table 1 Demographics, neurological data, and scores on the Spinal Cord Injury Quality of Life Basic DataSet (SCI QoL-BDS).
Full size table

Themes related to QoL definition and SCI QoL-BDS

Analysis resulted in 17 themes. These themes, the number of topics in each, their definitions, and corresponding exemplar quotes appear in Table 2. Of the 17 themes, coping and adjustment, complications and comorbidities, and purpose and identity were the most frequently cited across the five sites. In general, the QoL definitions and the three items of the SCI QoL-BDS showed CE across sites in the distribution of themes. We detail specific findings for each of the four items below.

Table 2 Theme definitions and examples.
Full size table

QoL definitions

The first row of Table 3 shows the percentages of the top three themes, in descending order, appearing in the responses to the QoL definition question by site. The percentages of all 17 themes for this item (summing up to 100%) appear in Supplementary Table 1. Overall, the top three themes per site show partial CE in the topics used to define QoL, with only seven of the 17 themes appearing in the top three positions across all five sites. Ability and autonomy was cited in the top three by all five sites, suggesting self-efficacy and the ability to do things freely in one’s life was a shared common denominator defining QoL. Further, coping and adjustment was frequently referenced by three sites and purpose and identity was heavily cited by both USA sites. Physical ability and limitations (18.2%) and ability and autonomy (20.4%) were most emphasized by participants in Melbourne. Finally, Table 4 provides exemplar QoL quotes of each of the top-three themes by site, further illustrating commonalities in how QoL was defined through the exact language used by participants.

Table 3 Highest three % referenced themes across the Quality of Life Definition item, each Spinal Cord Injury Quality of Life Basic DataSet (SCI QoL-BDS) item, and all items combined by sitea.
Full size table
Table 4 Exemplar quotes for the Quality of Life Definition item for the three highest referenced themes by site.
Full size table

In looking at the larger theme categories (see Supplementary Table 1), personal factors dominated these definitions across all five sites, albeit with a spread from 33.6 to 65%. Physical factors were highly factored in Melbourne (26.6%) and Utrecht (23.3%) responses, emotional factors in Ann Arbor responses (19.8%), social factors in Denver (20.1%), and environmental factors in São Paulo (20.4%).

Satisfaction with life as a whole—item 1

Only six of the 17 themes appeared in the top three positions of responses to the Life as a Whole item across all five sites (see second row of Table 3 and Supplementary Table 2). Coping and adjustment, loss, and purpose and identity each appeared in three of the sites. Overall, the notions of working through SCI/D challenges, mourning the losses from SCI/D, and finding purpose post-SCI/D (successfully or not) were key factors shared across all the sites. That this item was designed to garner a more holistic perspective on life satisfaction (as compared to specific QoL domains in the other two items), it allowed for the most salient items in personal QoL meaning to come to the fore.

The theme categories again show that personal factors were high in the definitions across sites, ranging from 15 to 35%, with high instances of emotional factors in Ann Arbor (22.1%), Denver (28.1%), Melbourne (21%), and São Paulo (17.9%); and high instances of physical factors across all the sites (between 29% in Utrecht to 18% in Denver). Together, similarities in themes suggest a close equivalence in framing responses to this question. However, some variety (for example, Utrecht’s low rate of emotional factors, 3.8%) suggests local differences in what is included in “life as a whole” definitions.

Satisfaction with physical health—item 2

The second item of the SCI QoL-BDS revealed seven of the 17 themes appearing in the top three positions across the five sites (see third row of Table 3 and Supplementary Table 3). Physical health and wellbeing was cited in the top three themes by all five sites, and complications and comorbidities by four of the five sites. Themes included in the personal and emotional factors categories appear in the top three in three sites (Denver, São Paulo, Utrecht). Of note, themes that fall within the social and environmental factors categories were relatively low or absent.

Returning to the theme categories, the highest percentages appear in physical factors, though personal factors and emotional factors also were cited frequently. Overall, however, this item showed a strong pattern toward the intended focus of the question—physical health and wellbeing—suggesting good understanding cross-culturally, while leaving room for other meaningful issues.

Satisfaction with psychological health—item 3

The third item of the SCI QoL-BDS shows eight of the 17 themes appearing in the top three position across the five sites (see fourth row of Table 3 and Supplementary Table 4). Emotions (positive or negative) were cited in the top three in every site but Utrecht, where it was fourth with 13.9%. Purpose and identity was cited in three sites. Emotions (positive) was cited frequently only in Ann Arbor.

All five sites frequently cited themes within the larger emotional factors theme category. Further, personal factors accounted for over a third of all references in this category across sites. Melbourne and Utrecht referenced social factors often, while Denver referenced environmental factors and São Paulo referenced physical health with surprising frequency.

Demographic and injury characteristics

In an attempt to interpret differences in theme percentages between sites in terms of cross-site sample differences, the same NVivo percentage analyses were conducted for key demographic and injury characteristics. Table 5 shows percentages for these comparisons by theme category (see Supplementary Table 5 for percentages of all 17 themes). No notable differences were apparent, suggesting rough equivalence across the sample.

Table 5 Frequency of theme categories across all items by demographic and injury characteristicsa.
Full size table

Discussion

The results of this CI study suggest satisfactory CE of the SCI QoL-BDS, as demonstrated by the largely similar distribution of percentages of themes and theme categories in the responses to the QoL definition and SCI QoL-BDS items across sites. We used CI to establish CE as measured by the SCI QoL-BDS items. By using qualitative methods, our work further provides a unique in-depth perspective on the experiences of QoL across sites reflecting both differences and similarities in the interpretation of the meaning of this concept as defined by those with SCI/D [12, 13, 21].

Our findings suggest that in spite of contextual differences by sites, there are important shared ways of defining QoL and the concept is contextually understood and defined. Results depicted in Tables 3–5 reflect both conceptual similarities and consistency/percentage of common themes used to describe QoL and items, showing a good degree of CE. That a relatively circumscribed group of salient issues emerged across all four interview items suggests key components of QoL transcend some cultural boundaries by virtue of shared experience [10, 11]. Even with the third item (satisfaction with psychological health, emotions and mood) showing the highest variation in themes, still less than half appeared in the top three positions across all five sites. This spread, however, may suggest something about the contextual variance of individual or cultural definitions of psychological health in particular. Nonetheless, taken as a whole, this study confirms and extends preliminary evidence on the cross-cultural validity of the SCI QoL-BDS [12].

In research on cross-cultural validity of QoL measures, CI are mostly used to test the clarity and appropriateness of a translation for the target audience. Our study was novel in that it attempted to define CE in greater depth and made use of CI to explore similarities and differences in personal meaning of QoL concepts and domains. We could find only one previous study using qualitative analyses to compare the importance of QoL domains across five European countries [38]. That study also showed more similarities than differences across countries but only at a level comparable to our theme categories, without more detailed results. Our in-depth qualitative data allows the participants’ own words to demonstrate understanding of the questions. The concepts that emerged in our study were remarkably similar across sites. Further, our data allowed us to demonstrate, rather than assume, that translating these questions into participants’ languages is enough to reach equivalence.

Our findings point to the rich complexity of the QoL concept, in ways similar to other studies, and beyond the similarities in theme frequencies across our sites. For example, Rohn, et al.’s recent work on QoL shows similar richness of data. In that work, the paths toward acceptance and coping with SCI/D-related challenges showed similar concerns with self-efficacy, shifting perspectives, physical wellbeing, loss, and social factors [9]. Further, similarities within our themes are seen across Hammell’s meta-synthesis of qualitative SCI QoL research [3]. Hammell’s synthesis focused on the factors that contribute or detract from QoL across 68 studies and found thematic concepts very similar to our own, though at times with different labels (e.g., body problems, loss, relationships, occupation, and self-worth, to name a few). The overlap between our findings and these studies is considerable. That we endeavored to disarticulate the challenges of the concept from the challenges of cross-cultural understanding now allows the complexity of QoL to come through with less concern that this complexity is the result of cultural differences in understanding the concept.

However, we did expect some differences across sites due to language, socio-economic, and geographical factors influencing participants’ interpretation of this concept. In this respect, we expected to see the highest similarities between the two USA sites. While the data for these two sites were quite similar, the similarities do not appear to be greater than those with the other sites. We also did not see large differences between the English and non-English speaking sites. Further, we show that differences in theme distributions and percentages were not explained by the demographic or injury/disease characteristics collected as part of the study. For example, though São Paulo participants were significantly younger than the rest of the sample, and likewise less educated, neither of these demographic factors revealed differences in theme distribution. The distribution of themes within São Paulo, rather, were likely a factor of their experiences with SCI/D and their unique expectations in life as well as the living environment. We note that access and socio-economic were mentioned by this group only when asked to define QoL. Similarly, only they mentioned opportunities for work/school when asked about life as a whole. This pattern is suggestive of the different socioeconomics of São Paulo, where the other sites are all in more developed countries, with full access to rehabilitation.

In general, the patterns of differences around each item were minor and taken together showed good CE across sites. Overall, the themes expressed were congruent with the item domain reflecting themes pertinent to that domain. Most ideas expressed in relation to physical health included physical health, ability, comorbidities, losses (referring to functional losses) while emotions and coping were the main themes used to define psychological health. An intriguing development was the emergence of the range and frequency of social factors in the themes. This emergent factor informed the subsequent adoption of a fourth item in the SCI QoL-BDS focused on social QoL, which we began collecting subsequently in the larger study. The importance of social health as a domain of QoL is not an explicit focus of the current SCI BDS-QoL and future work will elucidate this domain further.

There were several limitations to this study. While CE was found to be good, further research is needed to widen the sample and deepen comparisons across more populations engaged in using the SCI QoL-BDS. Except for the USA, there was only one site per country. This is an especially important limitation in São Paulo, where there is greater access to rehabilitation and other SCI/D resources than much of the rest of Brazil. In addition, the small sample size limits our ability to generalize about potential cross-cultural differences. Finally, in qualitative research, findings are contextual and more highly dependent on the researchers involved. Although consensus was reached, individual idiosyncrasies of training, bias, and insights cannot be ruled out entirely.

In conclusion, we feel confident in recommending the use of the SCI BDS-QoL, with appropriate translation, to collect comparable QoL data across sites internationally, given the overall good CE across the five sites. Having a standard way to assess QoL consistently across different settings, cultures and environments will allow comparison of research results and clinical data worldwide. Our findings shed promising light on the similar conceptualizations of QoL cross-culturally and the sensitivity of this measure to accurately assess QoL for persons with SCI/D around the world.