How Do You Translate "Peek-a-Boo" When "Peek-a-Boo" Doesn't Exist? Cultural Adaptation of a Parent-Report Screening Tool

Background: Parent-report screening and assessment tools are frequently used to identify children showing behaviors associated with autism spectrum disorder. Most available tools were created and validated with non-minority English-speaking individuals in western countries, then translated for diverse populations. A basic forward-back methodology has traditionally been used (Brislon, 1970). However, this methodology does not sufficiently prevent translation errors and fails to account for cultural differences (Beaton et al., 2000; Gjersing et al., 2010). Without appropriate translation methods, actual rates of over- or under-identification may differ between instrument versions, possibly resulting in a poor-quality assessment tool (Soto et al., 2015). This could contribute to disparities in identification among non-English-speaking communities.

Objectives: This presentation will describe the application of a rigorous translation and cultural adaptation process for a parent-report infant screening tool, including broader implications for Latin-American Spanish-speaking populations.

Methods: Three native-Spanish translators, from a variety of linguistic and professional backgrounds, collaboratively translated the instructions, items, and answer choices of the First Years Inventory (Baranek, et al., 2013). The quality of this forward translation was checked through preliminary pretesting and back translation. All translators, original instrument creators, and research team members developed a pre-final translated version by reviewing data collected during each of these phases. The pre-final version was tested with twenty target population members using cognitive interview strategies. Identified discrepancies in semantic, idiomatic, experiential, or conceptual equivalence between each item’s original intent and its meaning as understood by participants revealed areas in need of additional cultural adaptation. A final revision was developed based on qualitative and quantitative data from throughout the process.

Results: Pre-testing results indicate that Spanish-speaking parents had difficulty choosing from frequency-based answer choices. When narrating their decision-making process for selecting answer choices, parents largely considered which answer would describe typical development. This was evident in two patterns: First, parents assumed all items described typical behaviors and tended to respond positively, even to items describing atypical behaviors (i.e. reverse-scored items). Second, parents tended to adjust their answers based on their child’s age. If parents felt that an item described a behavior or skill that was developmentally advanced for their child, they adjusted the meaning of the item to describe a similar but more age appropriate behavior, and answered the item according to their modification. Parents also had difficulty understanding items with descriptions of actions (e.g. gestures, pretend play, motor milestones, motor-based RRBIs), some adjectives, and complex grammatical structures. Back-translation methods yielded some false impressions of the translation.

Conclusions: When translating a tool for use with a new population, significant revisions to instructions, answer choices, and item wording may be necessary to maintain conceptual equivalence across cultures. Results illustrate the importance of pre-testing as a quality checking procedure. Without data from pre-testing, significant issues identified would likely have influenced the way Spanish-speaking parents responded to items, in turn inflating or minimizing risk-scores inappropriately. Similar patterns may be relevant for other parent-report tools and other cultural groups. Additionally, some difficulties experienced by pre-testing participants may be relevant cross-culturally.