Many people live with two or more, or multiple, long-term health conditions (MLTC), such as cancer, heart disease, inflamed joints and mental health problems. People living with MLTC may progress to poor health and have a shorter life expectancy. Treating multiple health conditions require a balancing act. Often, they are prescribed many different medicines together (known as ‘polypharmacy’). Sometimes, these medicines can interact in unexpected ways, which can lead to adverse effects and cause further health problems. Our study aims to improve our understanding of the inter-relationships between MLTC, polypharmacy, personal and social factors to help optimise treatment for individual patients. We apply new developments in computer technology called ‘artificial intelligence’ (AI) to develop methods to analyse healthcare data collected from GP practices and hospitals from different regions across the UK to find out how long-term health conditions and polypharmacy change and interact over time, and how these patterns relate to personal and social as well as future health outcomes. These data are anonymised, large, and complex, but AI is very good at spotting patterns in these kinds of data. We will apply AI to look for patterns and trends on the interrelationships among long-term health conditions, prescribed medicines, personal/social factors and level of deprivation. In the long term, our research will lead to strategies for improved management of MLTC including targeted review of medicines.
Multiple long-term conditions (multimorbidity)(MLTC-M) is associated with premature mortality, significant treatment burden, and increased health care and socioeconomic costs. Deprivation contributes to its early onset while inequalities exacerbate it. Polypharmacy is associated with MLTC-M but the interrelationship is complex and its impact on health outcomes is poorly understood. Moreover, analysing these associations will require methods that can handle the size and complexity particularly of data extracted from large-scale electronic health records (EHR). We aim characterise the relationship between MLTC-M clusters and polypharmacy (MLTC-M-PP) and describe how MLTC-M-PP relates to inequalities and health outcomes using cutting-edge methodologies in data analytics. In this proposed study we aim to:
1. Develop novel and refine existing advanced statistical methods and artificial intelligence (AI) techniques to characterise MLTC-M and polypharmacy clusters and trajectories identified from large-scale EHR
2. Describe the associations of MLTC-M and polypharmacy, its impact on health outcomes, and how these associations vary by demographic factors and deprivation level
3. Describe how the interrelationship between MLTC-M and polypharmacy could inform risk stratification across demographic factors and deprivation level
We will apply AI techniques and explore using other advanced methodologies, including bursty dynamics theory from complex systems research, to characterise healthcare event structure, association rules mining to illuminate event sequence, topological data analysis, and market-basket analysis with topic modelling to define latent associated conditions. Quantitative methods will be optimised using CPRD resource, with particular focus on intersectional contrasts, such as by age groups, sex, ethnicity and deprivation levels.
• For identifying data-driven patterns (clustering and trajectories) of MLTC-M-PP: diagnoses, prescriptions, demographic factors and deprivation level (these are not 'outcomes' but variables to be used for discovering clustering patterns and trajectories).
• For determining impact of the derived MLTC-M-PP clustering and trajectories on health outcomes: all-cause and cause-specific mortality; disease clusters (e.g., subsequent MLTC-M arising from 'baseline' MLTC-M); health service utilisation (e.g., frequency of general practice clinic visits, frequency and duration of unplanned hospitalisation and/or readmission)
Nick Reynolds - Chief Investigator - Newcastle University
Dexter Canoy - Corresponding Applicant - Newcastle University
Barbara Hanratty - Collaborator - Newcastle University
Ellen Moss - Collaborator - Newcastle University
John Casement - Collaborator - Newcastle University
Liyuan Zhu - Collaborator - Newcastle University
Michael Barnes - Collaborator - Barts and the London Queen Mary's School of Medicine and Dentistry
Paolo Missier - Collaborator - Newcastle University
Rebeen Hamad - Collaborator - Newcastle University
Shaun Hiu - Collaborator - Newcastle University
Wasim Iqbal - Collaborator - Newcastle University
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation