Use of Machine Learning to Identify Patient Disease Patterns Requiring Dietician Interventions from Electronic Health Records for Optimal Care in General Practice

The increasing demand for primary care services, along with a decline in general practitioners, poses significant challenges to the National Health Service. This study aims to evaluate effective dietetic factors for optimised care in general practice and understand the role of dieticians in primary care. Applying statistical learning models and electronic health records (EHRs) from the SAIL Databank, we will analyse patient data to make data-driven decisions on dietetic care. The objectives include examining the frequency of attendance for patients with diet-related diseases, assessing clinical outcomes, and identifying optimal dietetic interventions from the diet-related features in EHRs. By reorganising and analysing comprehensive datasets and conducting literature reviews, this study seeks to address knowledge gaps and enhance the performance of dietetic interventions in primary care. The promising result is to demonstrate the efficient role of dietitians in improving healthcare outcomes for patients with diet-related conditions and the efficient way of staffing.

This study aims to apply machine learning techniques to large-scale electronic health records (EHRs) to identify patient conditions that may benefit from dietitian interventions in general practice. The objectives are to uncover diagnosis patterns linked to dietitian referrals, develop predictive models to support patient care pathways, explore multimorbidity patterns across referral groups, and build explainable models to guide future referral decisions.