Obstructive airway disease phenotype discovery using cluster analysis techniques in CPRD

Date of Approval: 
2016-08-02 00:00:00
Lay Summary: 
Asthma and COPD are very common respiratory diseases, in which patients suffer from various problems that make breathing painful and challenging (these are sometimes called exacerbations). There are an increasing number of studies that suggest that asthma and COPD are not a single disease but a group of diseases that have different causes despite the fact that they share similar symptoms making their diagnoses and treatment challenging. This study will investigate if it is possible to discover these disease subtypes in COPD and asthma using an approach called cluster analysis. This method groups patients (in this instance patients with COPD or asthma) into groups based on individual patient clinical characteristics (such as other diseases and socioeconomic status or demographic information). Patients in the same group (called a cluster) are more similar between them than with patients in other groups. Cluster analysis has been used before for asthma and COPD, but on a much smaller scale, with smaller patient datasets and not covering the spectrum of both diseases in the same analysis. Furthermore, the accuracy of different clustering techniques (e.g. partitional and hierarchical) has not been investigated comprehensively and our research will aim to systematically compare and assess them.
Technical Summary: 
Asthma and COPD are both obstructive airways diseases, and are chronic. Asthma is characterised by recurrent episodes of wheeze, breathlessness, chest tightness and cough. COPD, is characterised by airflow obstruction that is not fully reversible, and patients often have symptoms of cough, breathlessness, recurrent infections and wheeze. The expiratory airflow limitation in asthma is reversible, in contrast to chronic obstructive pulmonary disorder (COPD). Asthma and COPD are very heterogeneous diseases, with groups of clinical, pathophysiological and demographic characteristics called phenotypes. Additionally, both disease can overlap (asthma COPD overlap). Currently, there is not much known about the relationship between pathological features, clinical patterns of disease and response to treatment. In this study, we will use an approach named clustering analysis to try to identify asthma and COPD phenotypes in CPRD. Clustering is an exploratory approach in which can be used to identify subtypes in diseases. Cluster analysis to identify asthma and COPD phenotypes has been used in smaller studies. However, cluster analysis of a larger patient database could lead to the improved identification and characterisation of asthma and COPD phenotypes which would otherwise be grouped into broader disease aggregates.
Health Outcomes to be Measured: 
Clinical phenotypes of asthma and COPD Death Hospital admissions
Application Number: 

Spiros Denaxas - Chief Investigator - University College London ( UCL )
Francis Nissen - Collaborator - Roche
Jennifer Quint - Collaborator - Imperial College London
Liam Smeeth - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Maria Pikoula - Collaborator - University College London ( UCL )

HES Admitted Patient Care;HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation