Developing predictive models for personalised prompt intervention in patients with asthma or COPD

Developing predictive models for personalised prompt intervention in patients with asthma or COPD

5 May, 2020

Asthma is the most common long-term respiratory condition affecting over 30 million people in Europe. Asthma together with chronic obstructive pulmonary disease (COPD) are major drivers to hospitalisation. Only in the United Kingdom both conditions are the biggest burden during the winter term with the highest patient hospital and emergency loads.

These diseases, along with the co-morbid presentation of other conditions like hypertension, heart disease, depression and diabetes, influence the effectiveness of treatment. However, with appropriate monitoring and interventions, sufferers can manage their conditions and live for many years. For this reason, early recognition of symptoms and deterioration is paramount with prompt personalised intervention to ensure specific needs are met.

In this regard, one of the pilots within the BigMedilytics project, led by the University of Southampton in the UK, seeks to develop predictive models from real-time patient data using mobile and web enabled platforms in order to provide alerts both to patients with asthma or COPD (to encourage them to take action) and to the clinical team (whether or not to intervene).

These models will support the move to a proactive approach to care and enable the targeting of limited resources to patients who need them in a timely manner. Further, while intervening early with treatment, this avoids hospitalisation and use of emergency care and improves clinical outcomes, encouraging patient engagement and autonomy.

The pilot explores heterogenous data sources such as patient self-reports, weather, atmospheric pollution, patient primary and secondary care, and contextual information on patient lifestyles. The application of big data analytics will help discover the most important features contributing to the development of powerful machine learning algorithms in order to detect and understand the principal factors that trigger COPD or asthma exacerbation events in individual patients.

The COPD/Asthma pilot is part of the theme Population Health and Chronic Disease Management tackled by the BigMedilytics project. The University of Southampton, AstraZeneca, and my mhealth participate in this pilot.