Inflammation, nutrition, and the evolution of multiple long term conditions – an AI-based analysis of intersectionality in longitudinal health data (the InflAIM project)

Research question Can interventions be identified that limit chronic inflammation and improve nutrition to reduce the onset and rate of progression of multiple long-term conditions (MLTC)? Can these interventions be delivered in a way that modifies the risks of MTLC for individuals?

Background About 25% of the UK population have two or more long term conditions, associated with a reduction in quality of life, increased use of health services and reduced life expectancy.

Rational targeted prevention is essential, and policymakers have consistently called for better quality data on risk prediction and action, to inform decision making. Emerging evidence suggests common patterns among multimorbid conditions that develop over time.

However, understanding of the biological, social, behavioural and environmental factors act and interact to cause this clustering is limited by the complexity of the analytical challenge.

Of the mechanisms proposed, chronic inflammation is of particular interest: it is an established driver of disease risk for a range of chronic diseases individually and is amenable to intervention. It provides a possible link between nutrition and long-term disease. Variation in nutrition and malnutrition could provide an explanation for the social gradient in MLTC.

The role of nutrition in MLTC has received little attention to date and has significant potential for intervention at population scale.

Aims and Objectives By applying novel analytical methods to informative population datasets: (i) to identify the role of chronic inflammation, nutrition and malnutrition in the development of MTLC over the lifecourse and (ii) to quantify individual-level variation in risk according to social and ethnic background.

To develop and deliver interventions and policies targeting individuals and populations at risk to prevent the onset and progression of MLTC.

Methods Phase 1 will apply statistical and artificial intelligence-based methods to large-scale population datasets, to identify trajectories of MLTC throughout the lifecourse investigating how these are driven by chronic inflammation, nutrition and malnutrition.

We will explore how multiple aspects of disadvantage interact to determine MLTC, such as age, sex, ethnicity, socio-economic status, health inequalities, inclusion and diversity.

Phase 2 will explore the most acceptable and feasible intervention to prevent MLTC with people at risk of and healthcare practitioners.

Phase 3 will a deliver a fully powered randomised trial envisaged as a primary care-based intervention aimed at preventing the onset and progression of MLTC.

Timelines for delivery Prediction tools developed and validated: years 1-4; feasibility study and intervention protocol development: years 3-4; intervention trial: years 4-8 with long term follow up.

Impact and dissemination This programme aims to produce the tools, data, and analytics and targeted interventions needed to improve the lives of people at risk of MLTC.

To disseminate our research findings, in addition to research publications, we will produce policy briefings, reports, open-source software and data, and platforms for healthcare providers and policymakers to access data and models to inform their decisions.

We will centre our approach on engagement with stakeholders with MLTC; and extend to engaging with clinicians, local and national policymakers, to provide pathways to translate research findings into patient benefit.