COPD Predictive Model

Chronic obstructive pulmonary disease (COPD) is a prevalent respiratory disease in the UK, affecting an estimated 1.16 million individuals. COPD is responsible for a significant number of emergency admissions to hospitals, with over 108,000 admissions annually.

The effective management of COPD exacerbations is crucial in reducing the burden of the disease on the healthcare system and improving patient outcomes. Arden & GEM’s Advanced Analytics Unit has developed a responsive and reliable predictive solution to identify COPD patients at risk of exacerbations and guide clinical decision-making.

The solution uses machine learning techniques to predict the likelihood of emergency hospital admissions due to COPD exacerbations.

• Over 120 variables were extracted and engineered from fully linked primary and secondary care data sources spanning three years, including geo-demographic factors, comorbidities, clinical investigations, polypharmacy, care utilisation and lifestyle factors.
• Both traditional and advanced machine learning algorithms were trained, evaluated and compared to select the best model.

The final model was integrated into an interactive dashboard that reports daily predictions to facilitate proactive identification and clinical decision-making.

The model is available to health and care systems through our collaborative intelligence platform, Athena, offering the following benefits:

• Supports decision-making for primary care physicians, respiratory nurses, respiratory consultants, physiotherapists and ICS programme managers
• Enables proactive intervention, at an ICS level, to avoid emergency admissions and prevent adverse patient outcomes
• Best practice interventions such as reviews and education can be implemented to level up the gaps within COPD cohorts and reduce inequalities
• Demonstrates the power of linked data in predicting adverse outcomes and providing a framework that can be easily adapted for other cohorts, disease areas and outcomes.