The prevalence of comorbidities in COPD patients, and their impact on health status and COPD symptoms in primary care patients: a protocol for an UNLOCK study from the IPCRG

Funding The IPCRG provided funding for this research project as an UNLOCK Group study for which the funding was obtained through an unrestricted grant by Novartis AG, Basel, Switzerland. Novartis has no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. This study will include data from the Optimum Patient Care Research Database, and is undertaken in collaboration with Optimum Patient Care and the Respiratory Effectiveness Group, which will provide the data for this initiative without charge to the UNLOCK Group.


AIMS
To determine the prevalence of comorbid conditions in COPD patients, and to assess their association on the health status and on COPD symptoms in primary care patients in a range of settings across Europe. This study will address the following research questions: 1. What is the prevalence of comorbid conditions in COPD patients in real-life primary care? 2. Is there an association of one or more comorbid conditions with health status and with COPD symptoms in different primary care cohorts?

Design
The study involves a cross-sectional analysis of primary care data sets from different countries across Europe. At least eight data sets will be included in the analysis. Data may be collected in various ways and will be clearly reported and accounted for in the analyses, if appropriate.

Primary outcome
The primary outcome will be COPD disease-specific health status assessed using the CCQ and/or CAT.

Data analysis
The cohorts are derived from heterogeneous settings and countries, and therefore the data sets will not be combined, but the results will be analysed and presented separately. Data will initially be presented using descriptive statistics with graphical display. Missing values will be identified and excluded from the analysis with pairwise deletion. Differences between groups (based on the number of comorbid conditions) in health status measures (CCQ and/or CAT) and in COPD symptoms measure (MRC score) will be tested with factorial analysis of covariance (ANCOVA) models adjusting for important potential confounders (e.g., age, sex, smoking status, and disease severity). Estimated marginal means will be compared across data sets.
The χ 2 -test will be used to analyse differences in the prevalence of comorbidities. Predictive associations between variables will be studied with binary logistic regression models, and odds ratios and respective confidence intervals (e.g., 95%) will be reported. Statistical significance will be considered for P values o 0.05. The uncertainty of estimates, whenever possible, will be quantified in terms of confidence interval (e.g., 95%), and effect size values (e.g., r 2 , Partial Eta Squared) will also be analysed, interpreted, and reported. The analysis will be performed by the UNLOCK researcher. Analyses will be performed using SPSS, Version 20.0 (SPSS Inc, Chicago, IL, USA).
A sample size calculation is not applicable here, given that the study will be conducted by comparison of data from different samples that have already been collected. The known sample sizes will provide sufficient statistical power for even the very small effect sizes. For example, the values for a very small effect size such as d = 0.15 (Cohen's effect size for means difference) will require a sample of 1,500 patients for 90% statistical power: For d = 0.35, a sample size of 250 will be required for the same statistical power.
Ethical approval All included data sets must be approved by the ethics committees according to national law.
Participants identified in the data sets will remain anonymous, and patient confidentiality will be assured in the collection and merging of the data sets.

SHORT DISCUSSION
The diagnosis and management of comorbid conditions in COPD has important effects on exacerbations, health status, and mortality. [2][3][4][5] However, there is insufficient knowledge about the prevalence and impact of more than one comorbid condition in COPD patients in primary care settings. 8 Knowledge of comorbidities in COPD and a better understanding of their impact on health status may help to inform clinical practice and healthcare service delivery. 9,10 Although recent GOLD updates recognise the need to assess health status and include it together with FEV1 and exacerbations in the management algorithm, suggestions about comorbidities are limited to the recommendation that guidelines should be followed for each disease separately. 11 However, it is well known that health status is only weakly associated with spirometric values 20 such as the FEV1, and that some comorbidities as depression and anxiety are the factors that more strongly correlate with health status. 21 It is expected that this study will show to what extent other diseases or clusters of diseases also influence health status and will consequently inform our approach to COPD management.
In addition, this study of comorbidities will provide new insight into the development of preventive interventions, the reduction of burden of disease, and the adjustment of healthcare services to meet patients' needs in primary care. Expected results will provide knowledge about the prevalence of comorbidities in COPD, and their impact on health status between genders, age groups, and countries. We also expect that the result from this study, and future research on how and which multiple conditions affect patients' health status could be used to achieve a truly patientcentred care.

FUNDING
The IPCRG provided funding for this research project as an UNLOCK Group study for which the funding was obtained through an unrestricted grant by Novartis AG, Basel, Switzerland. Novartis has no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. This study will include data from the Optimum Patient Care Research Database, and is undertaken in collaboration with Optimum Patient Care and the Respiratory Effectiveness Group, which will provide the data for this initiative without charge to the UNLOCK Group. enrolment or completion of research: Almirall, Chiesi, Teva, and Zentiva; peer reviewer for grant committees: Medical Research Council (2014), Efficacy and Mechanism Evaluation programme (2012), HTA (2014); and received unrestricted funding for investigator-initiated studies from Aerocrine, AKL Ltd, Almirall, Boehringer Ingelheim, Chiesi, Meda, Mundipharma, Napp, Novartis, Orion, Takeda, Teva, and Zentiva. MR has received honoraria for educational activities and lectures or advisory boards from AstraZeneca, Boehringer Ingelheim, Chiesi, GlaxoSmithKline, Menarini, Mundipharma, Novartis, Rovi, and Teva. The remaining authors declare no conflict of interest.