Background
Significant racial and ethnic health care disparities, such as those experienced by Hispanic children with special health care needs (CSHCN), indicate gaps in enacting culturally competent health care, including within pediatric rehabilitation. Cultural competence is the ability to effectively and appropriately interact cross-culturally [
1]. Health disparities for Hispanic CSHCN are often described in three ways: (a) access to services, (b) quality care, and (c) service-related outcomes. This study focuses on culturally informed assessment of health care needs, which is essential to reducing each type of health disparity.
Racial and ethnic disparities in service access have been identified for Hispanic children according to their insurance coverage [
2], health screenings [
3], and community-based services (e.g., early intervention, pediatric rehabilitation). For CSHCN who have access to a primary care provider, nearly 22.5% of Hispanic CSHCN report having ‘moderate or big’ problems obtaining special health care services as compared to 13% of non-Hispanic White children. Further, non-White families, including Hispanic families, have more difficulty obtaining referrals when in need of specialty care [
4] and are less likely than non-Hispanic White children to have a usual or specific source of ongoing health care [
5].
Among those who obtain access, disparities exist with regards to quality care and service-related outcomes. Caregivers of Hispanic CSHCN appear to be less satisfied with their children’s services when compared to other major racial and ethnic groups [
3], as evidenced by poor provider-patient communication, care coordination, and family-centered care [
5]. In comparison to other major racial and ethnic groups, caregivers of Hispanic CSHCN also report higher rates of unmet therapy needs (i.e., physical therapy, occupational therapy, speech therapy) [
3]. Further, race and ethnicity are strong predictors of health status outcomes (e.g., medical status) for children participating in Part C early intervention (EI) where non-Hispanic White children are healthier than minority children [
6]. Moreover, health disparities between White and non-White children become larger between EI entry and exit at 36 months after controlling for health status at EI entry [
6]. Additionally, families of minority children are approximately twice as likely to have less positive family outcomes (e.g., perceived family quality of life) upon their discharge from EI than families of White children [
6]. Similarly, Hispanic preschoolers who had received EI services were more likely to experience community participation difficulties, when compared to non-Hispanic White preschoolers [
7].
One way to minimize health care disparities is to ensure that practitioners can enact culturally competent care during assessment, intervention planning, and intervention. As the first of these tasks, culturally competent assessment practices are especially critical to initial provider-client interaction and care plan development. Practitioners need to be able to effectively and efficiently conduct quality baseline assessment using culturally valid and reliable approaches to assessment [
8]. Due to time and costs associated with measure development, it is often more feasible to culturally adapt and validate an assessment from one culture for use in a different culture than to create new measures for specific cultural groups [
9].
Cultural adaptation involves establishing the equivalency and relevancy of an assessment from its source language and culture to the target population [
9‐
11]. Most efforts to increase the accessibility of measures for use with diverse clients typically emphasize language translation [
9,
12]. However, language translation alone is insufficient, because it captures only the etic perspective; however, further work is needed in order to capture the emic perspective to account for cultural differences that might alter the content of an assessment (i.e., instructions, questions, examples, scales) and how the assessment is administered (e.g., questionnaire, interview) [
9,
12,
13]. To our knowledge, there is no gold standard cultural adaptation framework, and a comparison of existing frameworks reveals several limitations: (a) process guidelines lack agreement; (b) discrepancies exist in what is required to achieve cultural equivalence; and (c) no known frameworks address cultural adaptation without the need for translation. The present study aims to address each of these limitations.
Cultural adaptation commonly involves forward and back translation and committee review [
9,
10,
14,
15]. Discrepancies exist around (a) the number of translators used and their qualifications, (b) whether to perform back translation processes on each forward translation or a synthesized version, (c) recruiting additional personnel for synthesis processes needed to produce a single translated version, (d) committee review composition and size, and (e) sample size recommendations for pilot testing [
9,
10,
14,
15]. These discrepancies may be related to feasibility, such as the lack of available qualified translators and cost.
Current frameworks and culturally adapted assessments have proposed different equivalence requirements and utilize different terms to assess for cultural equivalency. Cultural equivalency is “the extent to which an instrument is equally suitable for use in two or more cultures” [
11], (p1257). To achieve cultural equivalency, frameworks emphasize assessing for semantic, idiomatic, item (also called experiential and content), conceptual, measurement, and operational equivalencies (also called technical) [
9‐
11]. As shown in Table
1, semantic, item, and conceptual equivalencies are most commonly emphasized in current frameworks.
Table 1
Cultural equivalency dimensions across cultural adaptation frameworks
| ✓ | ✓ | ✓ | ✓ | | |
| ✓ | ✓ | ✓ | ✓ | | |
Sousa & Rojjanasriat, [ 14] | ✓ | | ✓ | ✓ | ✓ | |
Stevelink & van Brakel, [ 11] | ✓ | | ✓ | ✓ | ✓ | ✓ |
Similarly, semantic, item, and conceptual equivalencies are most commonly assessed in culturally adapted pediatric assessments with outcomes relevant to pediatric rehabilitation (i.e., quality of life, participation, performance, functional skills) as evaluated through reported assessment of equivalency types as well as comparison of equivalency definitions and the reported processes used (see Table
2) [
16‐
22]. Pediatric assessments commonly address semantic equivalence during translation. However, while these adapted children’s assessments emphasize item and conceptual equivalencies, there is less clarity about the methods used to achieve item and conceptual equivalencies. Only the cultural adaptation of PEDI for use with children in Puerto Rico outlined methods for assessing item and conceptual equivalencies; however, limitations exist in the methods employed [
19]. Limitations include: (a) suggestions for changes and additional feedback were not elicited from experts included in the study; and (b) no feedback was elicited from caregivers or children.
Table 2
Cultural equivalency dimensions addressed in culturally adapted assessments of children
Youth Quality of Life Instrument - Research Version (YQOL-R): Spanish version [ 16] | ✓ | | ✓ | | | ✓ |
Children’s Assessment of Participation and Enjoyment and Preferences for Activities of Children (CAPE/PAC): Spanish version [ 17] | ✓ | | ✓ | ✓ | ✓ | |
Perceived Efficacy and Goal Setting System (PEGS): German version [ 18] | ✓ | ✓ | ✓ | ✓ | | |
Pediatric Evaluation of Disability Inventory (PEDI): Puerto Rican version [ 19] | ✓ | | ✓ | ✓ | ✓ | ✓ |
Preschool Activity Card Sort (PACS): Spanish version [ 20] | ✓ | | | | | |
Paediatric Asthma Quality of Life Questionnaire (PAQLQ): Spain version [ 21] | ✓ | | ✓ | ✓ | ✓ | |
CAPE/PAC: Swedish version [ 22] | ✓ | | ✓ | ✓ | ✓ | |
Due to increasing prevalence of multiple cultural groups sharing a common language (e.g., Spanish-speaking clients may represent a wide range of cultural groups, including individuals from Mexico, Spain, and Venezuela), the development of a cultural adaptation framework that does not require translation is needed to facilitate culturally competent assessment practices. In the absence of such a framework, the Young Children’s Participation and Environment Measure (YC-PEM) is a pediatric functional assessment that was culturally adapted for use in Singapore, a primarily English-speaking country [
23]. Semi-structured interviews with providers and cognitive interviews with caregivers were used to inform content revisions related to item, semantic, and operational equivalencies.
The most common approaches to culturally adapting an assessment in preparation for psychometric testing involve assessing for (a) semantic and idiomatic as well as (b) item and conceptual equivalencies. Semantic equivalence is examined during the translation process by assessing the transfer of word meanings across cultures. Although idiomatic equivalence appears to be less commonly addressed, it likely gets assessed within semantic equivalence as both equivalencies address the transferability of words and phrases across cultures. Item and conceptual equivalencies are typically evaluated in terms of how relevant and appropriate the items and assessment concepts are to the end-user. Although prior literature lacks clarity on how to carry out these processes, they often involve focus groups, expert panels, or smaller samples of individuals from the target population and employ cognitive testing methodology [
16‐
19,
22,
24]. Lack of clarity in how to address item and conceptual equivalencies poses two potential problems: (a) it potentially limits the trustworthiness of existing culturally adapted measures; and (b) it limits the advancement of cultural adaptation processes, thereby limiting the development of high quality culturally adapted assessments. Although utilized in cultural adaptations, operational equivalence is less frequently addressed; however, it typically involves identifying accessible formats (e.g., interview, paper, or online questionnaire) based on participant characteristics [
11].
This study seeks to address these limitations in the context of culturally adapting the YC-PEM [
25], a validated patient-reported functional outcome measure, in English (no translation) and Spanish (translation) for potential use by caregivers of Hispanic CSHCN ages 0 to 5 years old living in the United States. To our knowledge, this is the first study to systematically develop and compare multiple versions of a culturally adapted questionnaire for potential use by a Hispanic population of young CSHCN.
The purpose of this study is to (a) compare the English and Spanish pilot YC-PEM versions to identify similarities and differences of culturally adapting an instrument with and without translation, and (b) examine the feasibility of developing culturally adapted English and Spanish versions of the YC-PEM for potential use by a Hispanic population of CSHCN. We had three specific aims:
1.
To identify revisions required to achieve semantic and idiomatic equivalencies when developing culturally adapted versions of the YC-PEM. We hypothesize that a greater number of changes are required to achieve semantic and idiomatic equivalencies of the culturally adapted and translated version of the YC-PEM than the non-translated version.
2.
To identify revisions required to achieve item and conceptual equivalencies when developing culturally adapted versions of the YC-PEM. We expect a similar number of changes required to achieve item and conceptual equivalencies for the English and Spanish YC-PEM versions.
3.
To examine the feasibility of developing culturally adapted English and Spanish pilot versions of the YC-PEM.
Study results will provide evidence to inform continued use of contemporary guidelines for culturally adapting patient-reported functional outcome measures with and without language translation.
Discussion
Existing racial and ethnic heath care disparities illuminate gaps in enacting culturally competent care. To enhance cultural competence within assessment practices, cultural adaptation can be used to ensure that assessments adequately capture outcomes for clients across cultures.
To our knowledge, this is the first study to examine similarities and differences in culturally adapting an instrument with and without language translation. Additionally, this study generates new knowledge surrounding the feasibility of carrying out cultural adaptation processes (i.e., language translation and cognitive testing). Study findings provide preliminary evidence to help guide decision making regarding cultural adaptation processes and the relative costs and benefits of cultural adaptation with and without language translation. Throughout the remainder of this section, each set of study findings is discussed in detail.
Semantic and idiomatic equivalencies
Results of this study support the hypothesis that more revisions are required to achieve semantic and idiomatic equivalencies of the YC-PEM Spanish version as compared to the YC-PEM English pilot version. These findings are congruent with the emphasis on using language translation to achieve sematic and idiomatic equivalence in cultural adaptation frameworks [
9,
10,
14,
15]. Additionally, results suggest that most, but not all, relevant revisions are detected during language translation, which suggests a potential benefit to pursuing cognitive testing following language translation. Although Lim and colleagues [
23] identified the need for multiple revisions without language translation, we identified only one revision to achieve semantic and idiomatic equivalencies when language translation was not required.
There are several ways to interpret the differences in results. One possible explanation is that Lim and colleagues [
23] pursued a transnational cultural adaptation without translation (i.e., from North America to Singapore), whereas this study focused on cultural adaptation without translation for use within the same country in which the instrument was originally developed. It is possible that a transnational context resulted in a greater number of revisions required to achieve semantic equivalence without translation. Alternatively, differences in results may be attributed to use of questionnaire versus caregiver interview for cognitive testing, as was used by Lim and colleagues [
23]. Questionnaires afforded feasible data collection but may have limited opportunities to ask clarifying and probing questions. Thus, results of this study may underestimate the revisions required in order to achieve semantic and idiomatic equivalencies for the non-translated (English) version.
Item and conceptual equivalencies
Caregivers proposed a greater number of revisions for item and conceptual equivalencies in the Spanish YC-PEM pilot version. Study results lend preliminary evidence that is contrary to the hypothesis that a similar number of revisions would be required for both versions.
Acculturation and language considerations
Group differences in the amount of feedback provided may suggest that fewer revisions are required to culturally adapt a measure without translation, particularly for use with a cultural group residing in the country in which the instrument was developed; however, this finding should be interpreted with caution due to the potential confounding effect of acculturation status. Skewed distributions for acculturation status in each group of English versus Spanish speakers may suggest that there is a potential effect of acculturation status on the number of revisions needed to achieve item and conceptual equivalencies. Thus, results may underestimate the impact of culture on cultural adaptations without language translation.
To our knowledge, no prior studies have examined the effect of acculturation status on the number of revisions required to achieve cultural equivalence of a measure. However, prior studies have examined the relation between acculturation and how caregivers conceptualize child development [
36]. Study findings indicated that concepts of child development vary across Mexican-American acculturation levels as well as between highly acculturated Mexican-American caregivers and Anglo-American caregivers when controlling for socioeconomic status (SES) [
36]. These findings suggest that Mexican culture may potentially influence caregiver perspectives about the concept of young children’s participation regardless of acculturation level, and thus revisions may be required across acculturation levels when culturally adapting measures. Additionally, concepts of child development vary across acculturation levels for high-SES participants, but not for low-SES participants [
36]. Hence, future studies might sample across acculturation levels and control for income and education levels in order to further examine the influence of acculturation on revisions required to achieve cultural equivalence of measures, such as the YC-PEM. Employing online YC-PEM completion (e.g., via personal computer or iPad during service visits) followed by caregiver group interviews might increase feasibility and improve data quality.
Language may also influence cultural expression. For example, Arcia, Reyes-Blanes, and Vazquez-Montilla [
37] found that participants placed emphasis on different Mexican cultural values depending on the particular interview language (i.e., Spanish or English). Niemann, Romero, Arredondo, and Rodriguez [
38] suggested that language preference may be indicative of in-group or out-of-group discrimination based on language, and that as a result of discrimination, different cultural values may be emphasized. These findings suggest that cultural adaptations of instruments with and without translation will likely result in different types of revisions. Thus, future research should examine the impacts of language on cultural expression and the interaction between language and acculturation level on culturally adapting measures. Future studies might add a measure to capture discrimination, such as the Hispanic Stress Inventory [
39], in order to examine this interaction effect.
Item equivalence
Feedback on the Spanish YC-PEM indicated that revisions necessary to achieve item equivalence primarily related to the addition or deletion of activity examples listed for each activity type. Given that the original YC-PEM activity types are fairly broad in nature, these categories may be deemed applicable across multiple cultural contexts due to their more generic nature. Common suggestions for revisions to activity examples pertained to self-care, educational activities, celebrations, and religious gatherings. These findings are consistent with identified Mexican cultural values pertaining to the values of responsibility, education, celebration, and familialism [
37,
38,
40].
Additionally, participant feedback emphasized the social and emotional aspects of their child’s participation, which is consistent with established Mexican cultural values [
37,
38,
40]. Because these features are not clearly captured in original YC-PEM activity types and examples, the emphasis on these qualities may warrant the addition of examples to further operationalize the dimension of involvement or the reframing of category descriptions in order to better capture these elements when assessing for a child’s participation in activities.
Conceptual equivalence
Conceptual equivalence concerns were identified among most participants, lending preliminary support for addressing conceptual equivalence regardless of language and acculturation level. Participants in this study commonly conceptualized “involvement” as requiring skills or some level of independence by the child. Hence, participants often indicated that items were not relevant based on the child’s young age or disability status. In a study by Arcia, Reyes-Blanes, and Montilla [
37] that examined the impacts of disability on cultural values, caregivers of children with disabilities placed higher value on “being independent.” These findings contrast with Mexican cultural values for interdependence [
37,
38,
40]. However, Arcia et al. [
37] noted that caregivers used labels of “independence” to indicate internalization of caregiver values (e.g., respect, strong ties to caregivers), which contrasts with common definitions of independence pertaining to autonomous child behaviors. Therefore, the notion of “independence” may be more in line with caregiver values typically associated with Mexican culture including familialism (e.g., respect, strong ties to caregivers), work ethic, responsibility, and education [
37,
38,
40]. Thus, findings from Arcia et al. [
37] may be reflected in caregiver feedback from this study, which indicates that the concept of participation is associated with independence. Participant feedback further supports this through the emphasis placed on social relationships and emotional sharing, which aligns with values of familialism and responsibility.
Therefore, study findings and prior literature may indicate that framing participation more explicitly in terms of co-occupation is more in line with the conceptualization of participation in Mexican culture. Co-occupation involves shared engagement in daily activities resulting in shared meaning [
41]. Although the YC-PEM implies co-occupation as young children typically participate in activities with a caregiver, this idea may need to be made more explicit throughout YC-PEM instructions and participation sections (e.g., by providing examples and reframing participation category descriptions).
Feasibility of cultural adaptation
As noted in previous studies [
9,
11], one of the greatest barriers to producing culturally adapted measures is the resource-intensive nature. Despite resource requirements, cultural adaptation of existing measures remains more time and cost effective than creating new measures specifically for the targeted culture [
9,
11]. Thus, examining the feasibility of the processes used in this study can help inform decision making about how to pursue future cultural adaptation work in ways that minimize cost and maximize quality.
Language translation
Language translation is a costly phase. In this study, costs were mitigated by recruiting bilingual, but not professional, translators. Reduced costs ensured that resources were available to undertake the full process and make enhancements to improve rigor.
During language translation, we found that synthesis and committee review phases were the most time intensive; however, examination of the discrepancy rate among translators demonstrates that these processes serve to systematically reduce discrepancies, thus suggesting an increase in translation quality. High translation quality is further supported by Aim 1 results, which show low revision rates related to semantic and idiomatic equivalencies following the translation period. Thus, for cultural adaptations when language translation is required, the language translation phase is critical to ensuring semantic and idiomatic equivalencies.
We recruited three forward translators, which exceeds the minimum of two translators recommended by most established translation guidelines [
9,
10,
14,
15]. Although the addition of a third translator increased baseline costs, it potentially reduced the time investment with respect to language translation. Specifically, less time was needed for synthesis and committee review processes, because principles of majority rules could be applied when issues of language preference arose. Therefore, priorities of time or cost may help inform decisions surrounding the numbers of translators used in future cultural adaptations with language translation.
Cognitive testing
The use of early intervention service providers as the primary recruitment method increased access to eligible families; however, recruitment and data collection time increased due to provider constraints (e.g., limited time within therapy sessions, cancelled visits) in comparison with direct participant interaction by research staff [
42].
Additionally, most eligible families who declined participation cited being “too busy/stressed.” Given that all eligible caregivers had children with a disability or delay, this may have occurred due to caregiver burden associated with caring for a child with a disability or delay. Alternatively, data collection occurred in the winter season when caregiver demands are high due to holidays and illness. Thus, future research should consider expanding the targeted population to include children with and without disabilities and delays and/or sampling during different seasons in order to increase recruitment rates and ultimately feasibility.
Due to resource constraints, PDF and paper versions of study materials were issued in lieu of pursuing online data collection. Consequently, errors in YC-PEM completion (e.g., incomplete or incorrectly completed items) occurred in both PDF and paper formats and across all participant cases. These errors did not occur in prior studies when the YC-PEM was administered online, because the online versions include automated prompts to guide participant completion [
28]. Thus, feasibility and data quality can be enhanced with online data collection (e.g., via personal computer or iPad during service visit). However, cognitive testing may be enhanced via interviews as the questionnaire format used in this study provided limited opportunities to pose clarifying or follow-up questions. Thus, to balance feasibility with data quality, future studies might employ online YC-PEM administration followed by group interviews to confirm findings.
Study limitations
Study results are preliminary and should be applied with caution due to several study limitations. First, small sample size and homogeneity did not permit parametric testing to compare language subgroups according to acculturation status and number of required revisions to achieve cultural equivalence. Cognitive testing using questionnaire versus interview format limited the ability to reach saturation with respect to participant feedback. Additionally, although participants completed cognitive testing on the daycare/preschool section of the YC-PEM, none of the children reported on for this study were enrolled in a center-based daycare/preschool program at the time of study enrollment. Thus, participant feedback for this section may have been restricted due to limited exposure to the daycare/preschool setting. Finally, reliable data on precise cost estimates for this study were not accessible and thus limited specificity in the examination of feasibility. Thus, more robust cost-benefit analyses are recommended for future studies.