Personalisation can mean two things: stratification, where a population is divided into sub-groups to facilitate better matching of interventions to individuals; or individualisation, where an intervention is specifically tailored to the individual. 

At CNELM, our principal approach is ‘individualised’.  However, we recognise the utility of both stratification and more traditional non-stratified evidence-based nutrition, and are interested in the advantages afforded by a blended approaches.

Individualised Personalised Nutrition interventions are as Personalised Synergistic Combination Nutritional Interventions (PSCNIs).  This technical definition represents a very familiar concept. 

Firstly, Nutritional Interventions (NIs) are simply the diets, dietary modifications, food supplements, meal timing guidance and related lifestyle changes suggested by clinical nutrition practitioners to their clients. Multiple interventions are given to the client at the same time – i.e. they are in Combination.

Moreover, we choose these interventions to work together – i.e. Synergistically – to produce an outcome.  Finally, the choice of which interventions to give will vary from individual to individual based on an assessment of their needs and health status: i.e. they will be Personalised.  Essentially, a PSCNI is a number of dietary interventions, designed to work together, chosen in a way that is relevant to a given individual.

Currently, the majority of PSNCIs are designed using Personalised Patho-Physiological Reasoning (PPR): this means clinically reasoning about the mechanisms of pathology or disease present in the individual to design a positive intervention.  However, whilst based in the scientific literature it is not fully evidence based: this is because PSCNIs present Evidence-Based Medicine (EBM) with genuine challenges. 

Combination Interventions, where two or more interventions are given together, are not the problem: there are many clinical trials on combination interventions. Synergy between the interventions is not the problem. Synergy simply means that the interventions ‘work together’.  Whilst this does mean that we can’t predict the combined effect from trials looking at the individual effects, a trial on the combined intervention will measure the combined effect, so this is not a problem.

The problem comes when we choose synergistic combinations of interventions in a personalised way.  This is because there are too many ways that a synergistic combination intervention can be tailored to the individual.  This makes it infeasible to expect there is a clinical trial on a precise PSCNI in individuals appropriately similar to a given client.

The solution is to use design Machine Learning algorithms to help support clinical reasoning, and to then trial these using a suitable clinical study.  At CNELM, our approach is to use Causal Models such as Bayesian Networks and Structural Causal Models to do this as they allow us to explicitly model and incorporate our pathophysiological reasoning approaches into machine learning algorithms. 

We use systematic reviews of the relevant scientific literature to support this process, particularly as these can also be used to support current Personalised Patho-Physiological Reasoning approaches to clinical practice.