Translation and validation of the COPD Patient Reported Experience Measure (PREM-C9) in Spanish and Catalan

The study addresses the need for validated patient-reported experience measures (PREMs) to evaluate healthcare quality from the patient perspective, specifically in chronic obstructive pulmonary disease (COPD). While PREMs are distinct from satisfaction measures and from PROMs, few COPD-specific PREMs exist. The PREM-C9, originally developed and validated in the UK, assesses patient experience across three conceptual areas (everyday life with COPD, usual care, and exacerbations). Given cross-cultural differences in healthcare delivery and language, the authors aimed to translate and culturally adapt the PREM-C9 for Catalan and Spanish and to assess its psychometric properties (reliability, structural validity, and known groups' validity) in these populations. They also explored factor structure for potential subscales and examined associations between demographic/clinical variables and PREM-C9 scores.

The paper outlines the increasing international emphasis on patient experience as a component of healthcare quality, referencing the OECD PaRIS initiative and literature differentiating PREMs from patient satisfaction and PROMs. It notes that while generic PREMs exist, disease-specific PREMs provide targeted insight. Prior work on COPD measurement has focused on health-related quality of life, with fewer tools capturing patient experience. The PREM-C9 was rigorously developed in the UK (including Rasch analysis and convergent validity with CAT and HADS), but further construct validation (e.g., known groups' validity and factor structure supporting subscales) had not been reported. This context motivates the present translation and extended validation.

Design: Observational study conducted March–June 2022 in two Catalonia sites (Hospital Clínic de Barcelona and Hospital Mútua Terrassa). Inclusion criteria: age ≥18, confirmed COPD diagnosis with spirometric obstruction, able to consent and follow instructions in Catalan or Spanish. Exclusions: asthma/pulmonary fibrosis, end-of-life, significant comorbidities such as severe heart failure, and discharge within the previous 15 days. Sample size: Targeted ≥90 participants (10 per 9 items) to support CFA and regression. Data collection: Self-completed PREM-C9 via paper or electronic (LimeSurvey); optional language choice (Catalan/Spanish). Sociodemographics: sex, age, education (not completed, primary, high school, technical apprenticeship, university), personal support at home. Clinical: years since COPD diagnosis (>10, 5–10, 1–5, <1), number of chronic diseases (0, 1–2, 3–4, ≥5), hospital COPD admissions per year (none, 1, ≥2). Dyspnoea: Medical Research Council (MRC) Dyspnoea Scale categories. Accessibility to healthcare: five items (access to hospital day services; ease of access to primary care; in-person primary care; phone access to pulmonologist; in-person access to pulmonologist), each on a 0–5 Likert scale. Overall satisfaction with care: five categories from very satisfied to very dissatisfied. Instrument: PREM-C9 (9 items; six-point Likert 0–5; raw total 0–45). Conceptual grouping: 3 items on exacerbations, 3 on usual care, 4 on everyday life. Original UK version showed good Rasch fit and acceptable test–retest reliability and correlations with CAT and HADS. Cross-cultural adaptation: Followed ISPOR good practice. Forward and back translations for Catalan and Spanish by independent translators; cognitive debriefing interviews with 10 COPD patients (5 per language); face validity with 7 clinicians/nurses; harmonization between languages. Notable adaptations: replaced less natural terms (e.g., translated “frustrated” with “disappointed” or “worried”); reversed response scale numbering so higher scores indicate better experience (opposite to original UK version), to align with local conventions. Ethics: Approved by Hospital Clínic of Barcelona (HCB/2021/0540) and Hospital Mútua de Terrassa (CPMP/ICH/135/95). Written informed consent obtained. Conducted June–July 2022. Analysis: Excluded questionnaires with missing PREM-C9 items for known groups and regression analyses. Descriptive statistics (mean/SD or median/IQR; frequencies/percentages). Floor/ceiling effects defined as percentage at min/max (ceiling if ≥15%). Structural validity: confirmatory factor analysis (CFA) to test three-factor model; fit indices: RMSEA (good ≤0.06; acceptable <0.08; poor ≥0.10), CFI/TLI (good ≥0.95; acceptable 0.90–<0.95). Known groups' validity: non-parametric tests for three hypotheses—lower education worse scores; higher breathlessness worse scores; higher satisfaction better scores—with Wilcox’s Q for effect size between adjacent categories with significant differences. Multiple regression: outcome = total PREM-C9; predictors = age (<60, 61–70, 71–75, >75), education, satisfaction, breathlessness, hospital admissions/year, years since diagnosis, and accessibility (averaged across five items, treated as continuous). Similar models for subscales. Alpha set at p<0.05. Software: R 4.4. Reliability: Cronbach’s alpha assessed separately for Spanish and Catalan.

Sample: N=239 (male 68.9%, female 30.2%); 61.4% completed Catalan version, 38.7% Spanish; 50.6% aged >71 years. Breathlessness distribution: strenuous exercise 28.8%, hurrying on level ground 30.2%, walking slower than others 16.7%, stopping for breath on level ground 16.7%, too breathless to leave house 7.7%. Acceptability/data quality: Floor effect 0.8% (2/239), ceiling effect 7.5% (18/239), both acceptable (<15%). Reliability: High internal consistency—Spanish Cronbach’s alpha=0.802; Catalan alpha=0.875. Structural validity (CFA): Items generally loaded on intended dimensions; item 2 in “My everyday life with COPD” had lowest loadings (Spanish 0.48; Catalan 0.54); Spanish item 5 also loaded weaker than expected. Global fit close to acceptable for CFI/TLI but RMSEA indicated poor fit: Catalan CFI=0.85, TLI=0.78, RMSEA=0.14; Spanish CFI=0.88, TLI=0.83, RMSEA=0.14. Factor loadings and correlations statistically significant (p<0.05). Known groups' validity: Two of three hypotheses supported. Satisfaction with care: significant differences across categories (p<0.001); median (IQR) PREM-C9 scores—Somewhat/very dissatisfied 21.5 (14.75–27.75), Somewhat satisfied 26 (21–31), Quite satisfied 36 (29.25–40), Very satisfied 40 (33–43). Significant adjacent differences: Somewhat vs Quite satisfied (effect size 0.429), Quite vs Very satisfied (0.227). Breathlessness: overall p=0.023; significant difference between “Too breathless to leave the house” 27 (20–36) vs “Stopping for breath” 37 (30–40) with effect size 0.296. Education: no significant differences (p=0.34). Multiple regression: Poorer overall satisfaction and greater breathlessness associated with poorer PREM-C9 scores; accessibility positively associated (better access linked to better experience). Time since diagnosis ≤1 year associated with poorer experience. Age, number of chronic diseases, educational level, and number of hospital admissions were not associated with PREM-C9 scores. Model R2≈0.27 (adjusted ≈0.19).

The study confirms that the Catalan and Spanish PREM-C9 versions maintain strong internal consistency and demonstrate acceptable construct validity via known groups analysis, addressing the primary research aim to validate the instrument in new languages and contexts. CFA suggests a generally coherent three-factor structure aligning with the conceptual grouping of items, though some items (notably item 2 in everyday life and item 5 in the Spanish usual care subscale) loaded less strongly, indicating that while subscales are conceptually plausible, further evidence is needed before routine subscale scoring is recommended. Findings that satisfaction with care and dyspnoea severity relate to patient-reported experience, while educational level does not, add to the evolving evidence base on determinants of PREM scores. The regression results corroborate known-groups findings and indicate potential clinical utility: newly diagnosed patients, those with greater breathlessness, and those reporting lower satisfaction may be at risk for poorer experiences, highlighting targets for care improvement. Translation choices (e.g., response-scale reversal for cultural appropriateness) likely improved comprehension and usability, supporting the feasibility of cross-cultural deployment.

Catalan and Spanish versions of the PREM-C9 for COPD show good reliability and support for construct validity, making them suitable for measuring overall patient experience in these languages. While the instrument’s three conceptual domains are supported in part by factor analysis, current evidence favors reporting a single overall score rather than subscales. The tool can inform quality improvement by identifying associations between patient experience, satisfaction, dyspnoea burden, and access to care. Future research should: (1) further examine factor structure and the viability of subscales in larger samples and across languages (including the original English version); (2) assess test–retest reliability and responsiveness, particularly in interventions such as pulmonary rehabilitation; and (3) explore item-level screening utility to identify addressable negative experience elements (e.g., low support or inadequate information).

The study did not assess test–retest reliability or responsiveness due to logistical constraints. Known-groups analyses may have been underpowered for some comparisons. Results from Catalan and Spanish versions were pooled for known-groups and regression analyses due to unequal sample sizes and missingness in the Spanish version, limiting separate language-specific inferences. Some CFA fit indices (CFI/TLI below 0.90; RMSEA 0.14) indicate less-than-ideal model fit, and certain items loaded weakly on intended factors. The ≤1-year diagnosis subgroup was small.