Complex Multiple long-term conditions (MLTC) Phenotypes, Trends, and Endpoints (CoMPuTE)

Background Over 25% of adults in England have Multiple Long-Term Conditions (MLTC). Complex-MLTC (C-MLTC: 4 or more conditions) is more disabling, and is projected to increase from 10% to 17% by 2035. Current management of MLTC focuses on treating individual conditions, causing unnecessary polypharmacy, increased burden for patients/carers, and inefficient referrals to specialists, who may not recognise the impact of MLTC. Research has focused primarily on describing the problem, identifying common patterns of MLTC, and strong determinants like socioeconomic status. However, comprehensive knowledge is lacking regarding which conditions are found together, how the burden of C-MLTC relates to organ health, ageing and lived experience, and what risk factors and interventions may influence development of C-MLTC. Novel AI-based methods, utilising large-scale healthcare data, coupled with active ethical engagement of patients and stakeholders in developing these methods, provides a powerful opportunity to improve care pathways. Aims To harness the power of longitudinal EHR to develop AI-enhanced models and tools and improve the management (prevention and treatment) of mid-life and early old age MLTC and C-MLTC.(T1) To characterise the epidemiology, inequalities and costs of C-MLTC by clusters of disease trajectories identified in mid-life and early old age.(T2) To ensure decision-making tools and other outputs are fit-for-purpose and account for actual lived care needs and therefore, serve the needs and expectations of target audiences.(T3) Methods Each Aim is delivered through one of three Themes. The AI and NHS Capability Theme (T1) applies Bayesian latent trajectory modelling and deep-learning AI-clustering methods to longitudinal EHR from primary care, linked hospital and death data. External validation will be performed using alternative datasources. GPs and patients will participate in co-design to ensure relevance of outputs. The Epidemiology, Inequalities and Health Economics Theme (T2) will use a series of scoping reviews to identify risk factors and interventions. Causal relevance of risk factors with trajectories and clusters of C-MLTC and all-cause mortality risk will be ascertained. The association between deprivation, lifetime costs and C-MLTC after adjustment for key confounders will be quantified. The Ethics, Patients and the Public Theme (T3) was public generated. Led by a public co-applicant, it will examine ethical and social issues associated with AI utilisation in healthcare and iteratively feed into the development of the models and the epidemiological analyses. This theme will use literature reviews, workshops with often excluded groups, deliberative forums, and focus groups. Anticipated impact and dissemination This programme will address knowledge gaps in our understanding of C-MLTC by using innovative AI techniques, causal inference, qualitative methods, and public leadership. T1 will provide guidance to clinicians managing C-MLTC. T2 will inform risk-factor management and quantify impact of C-MLTC on health resources. T3 will address ethical and practical considerations of representation, inclusion and equality, and will consolidate T1 and T2. Patients and the public will lead on co-developing material for: (i)researchers and wider stakeholders and (ii)patients and public themselves.