Development and validation of a machine learning model for predicting nonattendance in NHS ophthalmology clinics
Brief plain language background
Eye clinics are the busiest outpatient service in the NHS, caring for people with conditions like glaucoma and diabetes that can lead to permanent sight loss. These patients need regular appointments, but many miss them, sometimes without notice. This leads to delays in care, longer waiting times, and avoidable vision loss. Missed appointments also waste valuable clinic time and cost the NHS hundreds of millions of pounds each year.
According to The Health Foundation, individuals from more deprived communities are more likely to miss appointments, meaning those already at higher risk of poor health fall further behind. Right now, hospitals have no precise way to predict who will miss their appointment. Better tools are needed to spot and support those at risk, so care can be given to the people who need it most, when they need it.
What problem/knowledge gap does it help address
Hospitals currently send generic reminders to reduce missed eye appointments, but these are not tailored to individual needs and often fail to reach those most at risk of non-attendance, particularly patients from disadvantaged backgrounds. Currently, there is no robust method for clinics to predict who is likely to miss their appointment. This project aims to address that gap by using anonymised data from the EPIC electronic health record system to develop a predictive model. EPIC allows access to large-scale data across multiple care settings, not just ophthalmology, enabling insights into broader patterns such as co-morbidity or missed appointments in other specialties. By identifying the underlying factors influencing non-attendance, such as illness, geography or socioeconomic status, we can better understand why patients miss care. This approach offers the potential to design personalised care pathways to reduce missed appointments and support efficient NHS resource use.
Aim of the project
To develop and validate a predictive tool using interconnected electronic health records to accurately identify patients at risk of missing their ophthalmology appointments. This model will inform future interventions aimed at reducing non-attendance and improving the delivery of timely eye care.
Potential impact on people with sight loss
This project will help identify patients most at risk of missing eye appointments by analysing patterns via routinely collected medical records. Beyond simply predicting non-attendance, it will enhance our understanding of the underlying reasons patients disengage from care such as poor vision, co-morbidities, distance to clinic or lack of transport. These insights will inform how future services can adapt to patient needs more effectively. By identifying vulnerability early, the research will inform future and larger projects to design targeted interventions aimed at helping prevent avoidable sight loss. With validation, meaningful patient benefit could be achieved within three to four years.