Overview (S-000096)
Cloaked under the seemingly benign term of ‘Chronic Obstructive Pulmonary Disease (COPD)’, or simply put ‘long-term lung problem’, COPD is a rather complex progressive respiratory disorder with a diverse underlying pathophysiology contributing to the heterogeneity in presentation and progression. It is the leading cause of hospitalizations in Canada, the third leading cause of death globally and is emerging as the leading cause of mortality globally over the next decade. Currently, COPD management revolves around prevention and management of exacerbation or ‘lung attack’ since these aggressively impact lung function deterioration even among those at a milder disease severity stage. Thus, being able to predict exacerbation would enable clinicians to individualize treatment plans. The Acute COPD Exacerbation Prediction Tool (ACCEPT) prediction model is an important and timely undertaking in this direction and may hold the key to mitigate the burden of COPD on patients and the healthcare system. Though efforts are underway to assess and enhance its generalizability, ACCEPT remains to be validated in a population-based COPD cohort. We propose to assess the model performance of ACCEPT and ACCEPT 2.0 in the population-based longitudinal cohort of the Canadian Cohort Obstructive Lung Disease (CanCOLD; NCT00920348). As extension, we also propose to examine additional host characteristics and environmental exposure factors. We hypothesize that the currently proposed ACCEPT models are likely to underperform compared to exacerbation history alone in the milder disease population of the CanCOLD cohort. We propose to evaluate model performance (discrimination) by calculating the area under the curve (AUC) of the receiver operating characteristic (ROC) curves. Time-dependent (at 12 months) ROC curves will be used to account for the variability in follow-up time across patients and models will be compared using the DeLong test. Variables needed for this validation study are available in CanCOLD. These comprise of the predictors included in the ACCEPT model namely, number of exacerbations (event-based definitions of the Global Initiative for Chronic Obstructive Lung Disease) over the previous year, baseline age, sex, body-mass index, smoking status, domiciliary oxygen therapy, lung function, St George’s Respiratory Questionnaire (SGRQ) score, and current medication use. Where only CAT score is available, the model’s prescribed conversion to corresponding SGRQ score will be applied. The outcome variables of interest are moderate/severe exacerbations and severe exacerbations over one year. It is to be noted that ACCEPT model does not include host characteristics such as dysanapsis (airway to lung ratio on computed tomography) and biomarkers, nor characteristic of host’s environment - exposure to ambient pollution. The proposed study will inform potential model performance refinement in this population with the addition of predictors based on emerging evidence as a follow-up study in the CanCOLD cohort which will be explored as secondary objectives. Aligned with CanCOLD’s research objectives, the proposed study will support identification of COPD patients at milder disease stages likely to experience exacerbations and subsequent rapid deterioration towards a preventative early intervention approach among COPD patients from a primary care setting perspective.
Calendar
- Application Date
- 2023-03-13
- Approval Date
- 2023-03-28
Contact Details
- Name
- Dr. Jean Bourbeau
- Institution
-
Research Institute of the McGill University Health Centre (RI-MUHC) Note: DAR#907840