Machine Learning Predictions Pinpoint Scotland's High-Risk Hospital Admissions for Patients
In a pioneering development, the SPARRA project - a machine-learning based tool - is set to predict individualised risk of emergency hospital admission for the majority of Scotland's population. This innovative project, initially developed by NHS Scotland's Information Services Division, is now being advanced with the collaboration of The Alan Turing Institute, using the latest machine learning and artificial intelligence methods.
The SPARRA project's score could potentially be utilised when informing primary care interventions, offering a proactive approach to healthcare management. The team behind SPARRA includes statisticians from the Mathematical Sciences Department, providing academic leadership and statistical analysis to improve the predictive performance of the next generation of SPARRA.
The new model of SPARRA employs a super learner of a collection of machine learning models, resulting in improvements to both precision and calibration. This score will be delivered automatically to GP surgeries, allowing for timely and informed decision-making.
To ensure reproducibility and maintain high standards, a reproducible data science environment is being produced as part of the SPARRA project's development. The Department of Mathematical Sciences, ranked 4 in the UK in The Complete University Guide 2023, offers high-quality teaching and research across a wide range of disciplines, providing practical experience to support future careers and employment prospects.
Multiple funding sources, including from the AI for Science and Government programme (Turing), Health Data Research UK, The Health Foundation, and EPSRC, have supported this work. The Department of Mathematical Sciences, with its dedication to the learning experience of students, offers postgraduate and undergraduate programmes.
Key members of the SPARRA project team include Dr Louis Aslett and Dr James Liley. Once deployed, the SPARRA project will produce a score every month for approximately 3.6 million patients, assessing their risk of emergency admission to hospital in the following 12-month period. This could mark a significant step forward in proactive healthcare management and personalised risk assessment.
Read also:
- Postpartum Period and Gestational Diabetes: Does it Persist?
- Controlled spree of Legionnaires' disease among Harlem residents ceased, city health authorities confirm; however, locals push for increased openness and information disclosure
- Transform City for the Better
- Prostate Cancer Examination Guidelines, Outcomes, and Financial Aspects
