Improving equity in unscheduled care pathways through precision medicine

Background

Unscheduled care (USC) accounts for a large proportion of hospital activity, with over 3.2 million emergency attendances and 700,000 emergency admissions across Scotland in 2024/25.[1] Patients reach these services through a range of routes, including primary care, self-referral, and ambulance conveyance, and decisions made early in this pathway shape subsequent care and outcomes. Pressure downstream carries a clinical cost, with delays of more than five hours to admission from ED being associated with higher 30-day mortality.[2]

However, research and improvement efforts have tended to concentrate on performance within the ED itself, such as waiting time targets, rather than on the earlier decisions and routes that shape who reaches ED, and when. Comparatively little is known about risk and decision-making upstream, including within primary care and ambulance services, in part due to a lack of granular, linked data spanning these settings alongside ED and hospital records.

Vulnerable populations, including those living in poverty or with frailty and multimorbidity, face greater and earlier need for unscheduled care. In Scotland, multimorbidity arises 10-15 years earlier in the most deprived areas,[3] and emergency admission rates for 20-year-olds in the most deprived areas resemble those of 70-year-olds in the least deprived.[4] These patients often present with complex, overlapping needs poorly served by pathways designed around single conditions, and may face additional barriers to navigating appropriate care, from restricted access to primary care through to difficulty coordinating input from multiple services, potentially widening existing health inequalities.

However, how vulnerability shapes routes through care, and how system pressure interacts with patient characteristics to determine who is seen where and when, remains poorly understood. This limits our ability to identify which patients, and at which point in the pathway, interventions could most effectively reduce avoidable harm and improve outcomes.

Scotland's DataLoch resource links primary care, hospital, and ICU data, offering a unique opportunity to study unscheduled care as a whole-system phenomenon, building on prior work of the group to improve ED performance with predictive informatics. An approach combining data-driven care pathway analysis, identification of high-risk patients, and modelling of system-level dynamics is well-positioned to generate evidence for more effective and equitable unscheduled care.

Aims

The project aims to identify those at highest risk of adverse outcomes, evaluate the impact of patient vulnerability on care pathways, and model the system-level impact of scenarios affecting outcomes for patients in pre-admission care settings, to inform more effective and equitable unscheduled care.

Objective 1: To develop prediction models to risk stratify patients at highest risk of adverse outcomes (such as reattendance, readmission, or mortality) following contact with unscheduled care.

Objective 2: To characterise care pathways for patients across primary care, ambulance services and emergency departments, and evaluate how patient characteristics linked to health inequalities (including poverty, age, frailty, multimorbidity) impact these pathways.

Objective 3: To develop simulation models of unscheduled care pathways to evaluate the system-level impact of different scenarios, including operational pressures, care decisions and interventions, on outcomes for patients in unscheduled care.

Training Outcomes

The student will benefit from a supportive supervisory team with expertise in epidemiology, AI and emergency medicine, alongside a tailored training programme building on their prior knowledge and skills, including epidemiology, causal inference, and AI methods. The student will benefit from the vibrant environment in the Usher Institute, synergistic learning from related unscheduled care research undertaken by the supervisory team, and access to the interdisciplinary environment within the Fleuriot group in the School of Informatics. The studentship will provide a range of skills and, through the supervisory team's roles in national unscheduled care networks, potentially support translation of findings into policy and practice.

References: 

1. Public Health Scotland. Accident and Emergency: Overview. Available from: https://publichealthscotland.scot/healthcare-system/urgent-and-unscheduled-care/accident-and-emergency/overview/
2. Jones S, Moulton C, Swift S, et al. Association between delays to patient admission from the emergency department and all-cause 30-day mortality. Emerg Med J. 2022;39(3):168-173. doi:10.1136/emermed-2021-211572
3. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet. 2012;380(9836):37-43.
4. Liley J, Bohner G, Emerson SR, et al. Development and assessment of a machine learning tool for predicting emergency admission in Scotland. npj Digit Med. 2024;7:277. doi:10.1038/s41746-024-01250-1