Using Clinical Prediction Models to Improve Treatment for Patients With Chronic Obstructive Pulmonary Disease (COPD)

Overview

Chronic Obstructive Pulmonary Disease (COPD) is a chronic disease of the lungs that affects more than 2.5 million Canadians. Patients with COPD experience episodes of lung attacks (or exacerbations). During these attacks, patients experience an intense increase in symptoms, such as breathlessness and cough. It is challenging to decide which patients should be put on treatments that would reduce the risk of such lung attacks. The digitization of health records in many clinics and hospitals means complex risk prediction algorithms can be used to predict the risk of lung attacks to enable personalized care. In this study, our team will implement a risk prediction tool (called ACCEPT) into the electronic health records in two teaching hospitals in Vancouver, British Columbia (BC), Canada. A clinical study will be conducted to evaluate if the use of this tool results in patients with COPD receiving better care with better outcomes, and if they are more satisfied with the care they are receiving.

COPD is a heterogenous and progressive disease of the airways that affects millions of people worldwide. However, current treatment guidelines fail to provide personalised, patient-centered disease management. In contrast, precision medicine emphasizes the tailoring of disease management to patient characteristics and values to optimize patient care and outcomes. Clinical prediction models (CPMs) are major enablers of precision medicine, and facilitate targeted therapies to patients who will benefit the most from them. The investigators developed a CPM called ACCEPT that improves risk stratification for COPD patients by predicting the risk of exacerbation at an individual level and thereby enabling personalized, preventive disease management. Using a stepped wedged cluster randomized controlled trial (RCT), the investigators aim to evaluate the impact of integrating ACCEPT into routine COPD care at two outpatient respiratory clinics in Vancouver, British Columbia, Canada. The 'stepped wedged' RCT has a cross-over design, with treatment assignment done in a uni-directional, staggered format that will provide opportunities to control for time trend. The total duration of the study is 30 months. There will be a one-month phase in period with patient recruitment and data collection starting on month two. The last physician assignment will occur in month 18, and patient recruitment will continue until month 24. Follow-up data will be collected until month 30 to ensure six months of follow-up data for all patients. Primary and secondary outcomes will be analysed using generalized estimating equations to account for possible clustering of endpoints (multiple visits for each physician). Further, following the intention to treat principle, clusters (physicians) will be analyzed according to their randomized crossover time irrespective of whether crossover was achieved at the desired time.