Cross-cultural measurement development: The case of the Hispanic population

Purpose: Culturally and linguistically appropriate measurements of health care quality are needed to eliminate racial/ethnic disparities in the US. The healthcare experiences of Hispanics vary considerably by country of origin, generation, language ability and preference. Yet, intra group variations are typically ignored, and may lead to measurement errors. Applying these considerations to rigorous cross-cultural instrument development has been limited to date. This study aims to describe a) the systematic process for the development of a bilingual measurement of patient reported experiences of discrimination in inpatient health care settings, and b) the lessons learned engaging Spanish speaking participants.

Methods: A decentering approach was utilized for the simultaneous and iteratively development of an English- Spanish bilingual measurement. A racially/ethnically, and linguistically diverse research team was involved in: 1) Item candidate identification, 2) translation by committee, 3) expert panel review, 4) cognitive testing, 5) quality control, and 6) Item review analysis. A convenience sample of recently hospitalized adults was recruited during a follow-up outpatient visit at 4 urban clinics in Connecticut. Each subject participated in a face-to-face, audio recorded 50-minute qualitative interview in English or Spanish (based on language preference) to assess item comprehension, relevance, and appropriateness. Each item was evaluated by 3 subjects in each language. Bilingual data available was reviewed by the entire team to build the item pool for the metric testing phase. Results: Of the 173 items that were initially considered, 43 items were dropped after expert panel assessment and 24 after cognitive testing. The concurrent development of bilingual items helped achieved linguistic/conceptual equivalency as both versions aimed for simple language, active voice, and the elimination of jargon. All Spanish-speaking participants (n=50) self identified as Hispanic, mostly from Puerto Rico (84%), female (73%), and older than 51 years (61%). Only 16% reported speaking English very well and 20% not at all. 55% reported feeling most comfortable communicating with their medical care providers in Spanish. Spanish speakers were most effective in assessing specific word choices, but general comprehension probes were initially problematic. Participants had the most difficulty assessing comprehension of response options. Discussion: Engaging community collaboration in the early phases of cross-cultural instrument development enhances team diversity and acknowledges within group variations. This phenomenon is not unique to Hispanics, so it is hoped that raising awareness of cross-cultural and intra-cultural considerations will stimulate further discussion regarding best practices on measuring patient experiences, both locally and globally.