Donate Today Call Us

Harnessing the power of machine learning to save lives

University Hospital Southampton is leading a pioneering research project about how machine learning - a form of artificial intelligence provides systems the ability to learn and improve from data and prevent deaths, not just in the UK, but globally.

Hospital statistics

In the UK, nearly 80% of patients who experience cardiac arrest in hospital don’t survive. Experts estimate that around 50% of those patients could have been saved with more accurate prognosis and earlier warning.

Dr Daniels, consultant in respiratory at University Hospital Southampton who is leading the project says: “Hospitals run on machines using monitoring devises that read patients’ vital signs like blood pressure, heart rate, breathing rate, and countless other factors. Currently, nurses monitor patients, taking observations throughout the day and night and the higher the score, the sicker the patient is. Doctors are called for deteriorating patients, and this is called an early warning score system.

Since its introduction, this approach has improved the quality of care; however it can sometimes identify a deteriorating patient too late. 

When assessing a patient, doctors need to consider huge amounts of information from many different sources in a short space of time.”

Machine learning can use every piece of digitally available data to give a highly personalised, accurate risk score. It has the power to translate this data into actionable information to aid human judgement to help staff make correct and timely decisions and to decrease the chance of errors. 

It eliminates reliance on ‘one size fits all’ clinical practice, giving recommendations for diagnosis, prognosis and treatment that is personalised to the individual patient.

Donate today



The impact 

Statistics a doctor must analyseMachine learning can use every piece of digitally available data to give a highly personalised, accurate risk score. This could benefit patient care and outcomes by: 

  • Identifying deteriorating patients earlier so they could be treated sooner and have the best possible chance of survival. This work could save hours, hence lives.
  • Accurately identify patients who are at greater risk so they can be monitored more actively and often.
  • Reducing the number of unnecessarily escalated early warning alarms. 
  • Resulting in fewer false alarms so staff can spend time with the patients who need them most.
  • Improving long-term outcomes for patients through earlier diagnosis and treatment, particularly from complex diseases.


UHS is at the forefront of digital healthcare innovation in the NHS. As a recipient of Global Digital Exemplar status and with a digital repository of clinical observations data, we are perfectly placed to integrate a machine learning technique into a healthcare environment to harness its potential to improve clinical outcomes and organisational efficiency. 

This project is cutting edge innovation in medical care. It is a truly unique pilot project which has huge potential to save lives and transform medical practice on an international scale.

Donate today


The investment

Machine learning in action

Implementing the machine learning techniques at University Hospital Southampton will cost just £132,000 - a relatively modest investment for such significant potential to save lives. This includes:

  • £50,000 investment in the IT and business intelligence support staff required to integrate the algorithms into the hospital systems. 
  • £20,000 to fund clinical leadership time to work with the maths team at the Alan Turing Institute to coordinate and design the implementation, as well as to understand and interpret the results. 
  • £50,000 to part-fund a PhD student to work with Professor van der Schaar to help complete the huge amount of data analysis required. 
  • £12,000 fundraising appeal costs

Help us to harness the power of machine learning to save lives. 

Call Southampton Hospital Charity on 023 8120 8881 to support this project today. Donate today!


Back to the Appeals page


« View all articles