Study design and participants

This observational study was conducted on two samples of COPD patients recruited in the Hospital Clínic of Barcelona and Hospital Mútua Terrassa based on the following criteria: age≥18 years; a confirmed diagnosis of COPD and able to consent and sign a consent form, able to follow written and verbal instructions in Catalan and Spanish. Only patients with a clinical diagnosis of COPD based on the patient’s medical history and the presence of airflow obstruction confirmed by spirometry were eligible for inclusion in the study. Participants who met any of the following criteria were excluded: asthma/pulmonary fibrosis, who are nearing end of life and had significant other co-morbidities such as severe heart failure. The patients who had been discharged from the hospital in the last 15 days were excluded. The study period was between March and June 2022.

The questionnaire was self-completed. The questionnaire was administered in paper and electronic form. The electronic survey was created using Lime Survey software.

Sample size

The sample size was based upon a criterion of ten participants per number of items in the questionnaire (i.e. 10 × 9 = 90)14 which was considered sufficient for the Confirmatory Factor Analysis (CFA) and multiple regression analysis.15

Study variables

Data was collected on the following variables in the order they are detailed below. All questionnaires were self-completed in paper or digital format. Participants were given the option of responding in Catalan or Spanish.

Sociodemographic data

Sociodemographic data collected included sex (male/female), age, educational level, and personal support. Educational level, (not completed; primary level; high school; Technical apprenticeship; university). Not completed means that the respondent did not finish obligatory primary education. Technical apprenticeship is a structured training program that combines on-the-job learning with classroom-based instruction. Personal support at home included the following response categories: living alone; living with company; living with one or more family relatives; living with non-family members; living in a nursing home).

Clinical data collected included

years since COPD diagnosis (>10 years; 5–10 years; <1 year); number of chronic diseases (0; 1–2; 3–4; >5); hospital COPD admissions per year (none; 1; ≥2).

The MRC Dyspnoea Scale

Reference16 was used to assess the degree of dyspnoea and the extent to which breathlessness affects mobility. The MRC Dyspnea Scale includes the following categories: breathlessness with strenuous exercise; breathlessness while; walking fast on level ground; walking slower than other people because of breathlessness; stopping for breath after a few minutes on level ground; being too breathless to leave the house).

Accessibility to healthcare services

accessibility to healthcare was assessed through five questions regarding various aspects of accessibility to healthcare services: access to hospital day services, ease of access to primary healthcare services, ease of in-person access to primary care services, ease of phone access to the pulmonologist, ease of in-person access to the pulmonologist. The response categories for each question were a five-point Likert scale ranging from 0 (nothing) to 5 (a lot).

The PREM-C9 Questionnaire

The PREM-C9 is a scale developed and validated within the English healthcare context to better measure and understand the experience of patients with COPD and the care they receive. It consists of 9 items distributed in 3 subscales that measure the impact of COPD in the daily lives of patients, their opinion of the health care they receive, and their expectations about the care they will receive in case of an exacerbation of the disease (Fig. 1). The PREM-C9 instrument describes the patient experience of COPD and items are scored on a six-point Likert scale from 0 (completely disagree) to 5 (completely agree). Summing scores from each item leads to a raw total score between 0 and 45. The 9-item version of the questionnaire demonstrated12 good fit to a Rasch model (xp = 0.33; PSI = 0.75) and good distribution of item scores. PREM-C9 scores correlated moderately with other measurement tools such as the COPD Assessment Test and the HADS Scale, and test-retest reliability was considered acceptable.12

Fig. 1

The figure contains the items of the COPD Patient Experience Healthcare Questionnaire.

Satisfaction with care

Participants were also asked to rate their overall satisfaction with healthcare services by choosing one of five response categories: very satisfied, quite satisfied, moderately satisfied, slightly dissatisfied, and very dissatisfied.

Cross-cultural adaptation

Before using an instrument such as the PREM-C9 in another cultural or linguistic setting, it is important to ensure a rigorous process of translation and cultural adaptation. In the present study, translation followed the International Society of Pharmacoeconomics and Outcome Research (ISPOR) Principles of Good Practice for the Translation and Cultural Adaptation Process for Patient-Reported Outcomes Measures17. As both Catalan and Spanish are Roman languages, which means they evolved from Latin and both are official languages in Catalonia being widely spoken, the PREM-C9 was translated into both languages. The adaptation process consisted of forward and back translation by pairs of translators working independently and in-depth cognitive-debriefing interviews in COPD patients with a range of demographic and clinical characteristics to assess comprehension and acceptability. The face validity testing was performed by a sample of COPD clinicians and nursing staff. Harmonisation between the Catalan and Spanish versions formed the last stage of the adaptation process. Some aspects of the cross-cultural adaptation have been published previously11.

Ethics

The study was approved by the Ethics Committee of the Hospital Clinic of Barcelona (HCB/2021/0540) and Hospital Mútua de Terrassa (CPMP/ICH/135/95) and was conducted between June and July 2022. Participants gave fully informed, written consent to participate.

Analysis

Questionnaires with missing values on any of the items of the PREM-C9 questionnaire were excluded from the analysis because complete data were considered a requirement for the known group validity and regression analyses.

For continuous data, descriptive statistics were expressed as mean and standard deviation (SD) for normally distributed and as median and interquartile range (IQR) for non-normally distributed variables. For categorical data, descriptive statistics were expressed as frequencies and percentages. The distribution of answers across response categories was assessed for each item, and potential ceiling and floor effects were estimated by calculating the percentage of subjects with the maximum and minimum possible scores for the PREM-C9, respectively. A ceiling effect was considered to be present when 15% or more of patients obtained the maximum score.18

Structural validity of the PREM-C9

Structural validity can be defined as the degree to which scores of a scale are an adequate indication of the dimensionality of the construct being measured19,20, Although the original authors only proposed one overall score for the PREM-C9, items are clearly grouped within the questionnaire itself into 3 categories (My everyday life with COPD, Usual care for COPD, and COPD exacerbation), suggesting that there may be an underlying dimensional structure to the items. We were interested in exploring whether this was in fact the case, as being able to score the questionnaire using sub-scales could facilitate interpretation. For example, patients could score well on one dimension but less so on another. To test this possible dimensional structure we used confirmatory factor analysis (CFA). The comparative fit index (CFI), Tucker-Lewis index (TLI), and root mean square error of approximation (RMSEA) were used to assess the adequacy of the model. Global model fit was evaluated by the following fit indices: root mean square error of approximation (RMSEA) - values close to 0.06 or below indicate good fit21,22. Values lower than 0.08 suggest adequate model fit22,23 and values ≥ 0,10 indicate poor model fit22,23. Comparative fit index (CFI) and Tucker-Lewis index (TLI) ≥ 0.95 suggest good model fit22,21 and values from 0.90 to < 0.95 indicate acceptable model fit24.

Known groups’ validity

This type of construct validity assesses an instrument’s ability to distinguish among distinct groups who are expected to show different scores on the construct or measure of interest based on their demographic and/or clinical characteristics. Known group validity was estimated using non-parametric tests. Three hypotheses were identified and tested:

1.

Participants with a lower level of education will tend to have poorer scores on the questionnaire.

2.

Participants with higher breathlessness scores will have poorer scores on the questionnaire overall, particularly because greater breathlessness will impact on their experience of the illness itself.

3.

Participants reporting higher levels of overall satisfaction with care will have better scores on the questionnaire overall.

Given the non-normal distribution of scores on the PREM C9, between group differences to test the above hypotheses were performed using non-parametric tests. Effect size between any pairs of adjacent categories for which statistically significant differences were found on the variables of interest was calculated using Wilcox’s Q.

Multiple regression analysis

A multiple regression model was carried out to determine which sociodemographic and clinical factors contributed most to differences in the PREM-C9 total score. The dependent variable was the overall PREM-C9 score. Independent variables included were: age ((categorized as <60years; 61-70 years; 71–75 years; >75 years to ensure a reasonably even spread of patient numbers in each age group)); Educational level; satisfaction; breathlessness; Hospital admissions per year (years since disease diagnosis and Accessibility to healthcare. Categories included for each variable within the model were as described in the section on study variables above, except for age which was categorised as <60 years, 61–70 years, 71–75 years, and>75 years to ensure a reasonably even spread of patient numbers in each age group, and accessibility which was transformed from a categorical to a continuous variable by calculating the average score of the five aspects measured. Similar models were constructed to test for association of the independent variables with PREM-C9 sub-scale scores.

In all statistical analyses, significance was set at p < 0.05. All analyses were conducted using R version 4.4.

Reliability analysis

Reliability (i.e. the extent to which the items comprising the scale measure the same construct) was evaluated using Cronbach’s alpha, with a coefficient of 0.70–0.90 considered to indicate satisfactory reliability. Reliability was assessed for the Spanish and Catalan versions separately.