Article Text
PDF
Abstract
We developed a chronic obstructive pulmonary disease (COPD) patient-reported experience measure (PREM-C9). 174 patients with COPD (86 [49%] with a confirmed diagnosis and 88 [51%] with a self-reported diagnosis of COPD) completed a 38-item list, COPD Assessment Test (CAT) and Hospital Anxiety and Depression Scale (HADS). Hierarchical and Rasch analysis produced a 9-item list (PREM-C9). It demonstrated fit to the Rasch model (χ² p=0.33) and correlated moderately with CAT (r=0.42), HAD-anxiety (r=0.30) and HAD-depression (r=0.41) (p<0.05). A substudy confirmed its ability to detect change prepulmonary and postpulmonary rehabilitation. The PREM-C9 is a simple, valid measure of experience of patients living with COPD, validated in this study population with mild to very severe disease; it may be a useful measure in research and clinical audits.
- respiratory measurement
- pulmonary rehabilitation
- COPD
- PREM
Statistics from Altmetric.com
Introduction
The drive to improve the quality of care for patients requires robust instruments to capture patients’ perceptions of the healthcare that they receive. The experience of patients living with chronic obstructive pulmonary disease (COPD) and their views on the quality of healthcare they receive is not captured in currently available Patient-Reported Outcome Measures (PROM).1 A new era of measures—Patient-Reported Experience Measures (PREM)—is emerging. PREMs assess patient experiences as opposed to outcomes (such as symptoms and quality of life) per se.2 PREMs are often used as a benchmark for care experience and the experience of living with a disease that complements information obtained with PROMs. Here, we present the development and validation of a COPD-specific PREM.
Methods
Study 1
A 38-item list, extracted from our previous qualitative work,3 the COPD Assessment Test (CAT)4 and Hospital Anxiety and Depression Scale (HADS)5 were administered to patients with COPD.6 Hierarchical item reduction and Rasch analysis were applied to our 38 items to identify those with the best measurement properties and potential for inclusion.7 This iterative process continued until a final item-set that met the Rasch unidimensional measurement model requirements was achieved.8 Test-retest reliability was assessed in patients who repeated the item-list 1 week later and reported that their general health was ‘about the same’. Concurrent validity was assessed using correlations between the final item-list (PREM-C9), CAT and HADS (see online supplementary file 1).
Supplemental material
Study 2
Further concurrent validity of PREM-C9 with the CAT and HAD, and its change scores prepulmonary and postpulmonary rehabilitation (PR) were tested in a different group of patients with COPD in a substudy (see online supplementary file 1).
Data analysis
Items were flagged for removal due to floor/ceiling effects (>40%), age bias (Pearson’s correlation), gender bias (independent t-test), item-item correlation (>0.7) and missing responses (>15%). Fit to the Rasch model was determined by a non-significant χ² statistic (p>0.05) and Person Separation Index (PSI) (>0.7). Test-retest reliability was assessed using the intraclass correlation coefficient (values>0.7). Construct validity was assessed in both studies using correlations (Pearson’s r) between PREM-C9, CAT and HAD. Statistical significance was set at p<0.05. In study 2, PREM scores pre-PR and post-PR were compared using Student’s paired t-test.
Results
Study 1: PREM development and preliminary validation
In total, 174 patients completed the questionnaire pack (mean age 71 years, SD 9; female 52%; mean FEV1 59%, SD 21.9) (online supplement table 2). Participants were recruited through British Lung Foundation Breathe Easy Groups (n=88, 51%),hospital and community pulmonary services (n=86, 49%).
Item reduction
Twenty-two of the 38 items were removed in hierarchical reduction: age bias (n=5), gender bias (n=1), missing data (n=6), floor effect (n=9), item-to-item correlation (n=1) and three items due to expert opinion (online supplement table 1). The remaining 13 items underwent Rasch analysis. Four items were removed due to poor fit to the Rasch unidimensional model. The final 9-item solution (PREM-C9) demonstrated good fit to Rasch (xp=0.33; PSI=0.75) and good distribution of item scores (logit range: −0.1 to +0.2) (figures 1 and 2). Each PREM-C9 item is scored 0 (good experience) to 5 (bad experience); total score ranges from 0 to 45 (figure 3). PREM-C9 scores moderately correlated with CAT (r=0.42), HAD-anxiety (r=0.30) and HAD-depression (r=0.41) (p<0.05).
Rasch modelling showing a well-fitting item ‘I have enough information about my condition (good)’.
Distribution of item thresholds based on a Rasch logit scale.
The patient-reported experience measure (PREM-C9) instrument.
Test-retest reliability was assessed in 88 (51%) participants and was acceptable (Intraclass Correlation Coefficients (ICCs)=0.7).
Study 2: PREM validation and change scores
Thirty-six patients with confirmed COPD (mean age 65 years, SD 10.97; male 61%; mean FEV1 53.4%, SD:19.4) completed questionnaires pre-PR and post-PR. PREM-C9 demonstrated very similar correlations to those seen in Study 1: CAT (r=0.48) HAD-anxiety (r=0.44) and HAD-depression (r=0.46). A significant difference in PREM-C9 scores pre-PR and post-PR was found (20.08±8.43 and 14.72±10.59, respectively; mean change 5.36±9.70, 95% CI 2.08 to 8.64, p=0.002), indicating improved experience.
Discussion
We report the development and validation of the first COPD PREM. The PREM-C9 offers a new approach by capturing patient experience and their interactions with healthcare systems and clinicians in three main areas of COPD: 'my everyday life with COPD' (items 1-4); ‘usual care in COPD’ (items 5–7); and ‘self-management and exacerbations’ (items 8–9). These sections draw on important patient-centred aspects of COPD.
PREM-C9 demonstrated good fit to the Rasch model confirming that its items relate to the same underlying structure—patient experience; enabling simple item summation to obtain an overall experience score, from good to bad. The PREM-C9 correlated moderately with PROM’s suggesting that the PREM captures a related but somewhat different patient perspective.
The focus on ensuring that patients remain at the heart of healthcare and reporting patient experience is becoming an essential part of quality improvement. Evaluation of healthcare experience is continuously evolving with the patient perspective increasingly sought.9 This approach differs from the measurement of symptoms and quality of life. Current satisfaction surveys are also limited as they are normally generic and loosely constructed.1 Assessment using PROMS is important but they do not capture the experiential views of patients.1
This short instrument should complement the range of instruments currently used within the COPD population when used alongside clinical audit and quality improvement strategies.10 Measuring and benchmarking patients experience in a systematic way pre-healthcare and post-healthcare intervention may be a powerful way to demonstrate quality improvement and outcomes for healthcare professionals and patients.
Our study does have limitations. Patients recruited indicated they attended Breathe Easy specifically for their COPD condition as advised by their medical/nurse practitioner and had spirometry at time of recruitment but we did not confirm this through accessing medical records. We conducted questionnaires in English only and results may not be applicable linguistically across a diverse patient group. Patients who did not have a good knowledge of English declined to participate as subjects from which the PREM was derived.
Conclusion
We have summarised the development and preliminary validation of the first published PREM in COPD (PREM-C9) (figure 3). The instrument was designed to present what patients consider is important to them and in relation to their care. We suggest this instrument should be used in routine practice to aid clinicians to understand the patient perspective and to form patient prioritised goals in codesigned management programmes.
Acknowledgments
We would like to acknowledge Dr Susan Walker for her contribution in stage 1 of the PREM development. We would also like to extend our gratitude and thanks to the patients and their families for participating in this study.
Footnotes
Contributors All authors (MH, CMR, SA, LG, PWJ and JY) contributed to the study design, data interpretation and contributed to the manuscript. MH and JY were involved further in data collection, statistical analysis and drafting the initial manuscript.
Funding This work was commissioned as part of the North East London, North Central London and Essex, Health Innovation and Education Cluster (NECLES HIEC) & UCL Partners, in partnership with Anglia Ruskin and Manchester Universities.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval Ethical approval for the study was provided by the Bloomsbury National Research Ethics Committee (ref: 12/LO/2022).
Provenance and peer review Not commissioned; externally peer reviewed.