Dr N Peek, Prof A Brass, Dr Mahesan Nirmalan
No more applications being accepted
Competition Funded PhD Project (European/UK Students Only)
About the Project
A vast amount of data is recorded from patients who are critically ill on the intensive care unit (ICU). Clinicians use this data to inform treatment decisions, to plan discharge and to predict mortality. The quantity of information that is available means that it is difficult to synthesise all available data when reaching a clinical decision.
Myriad scoring systems have been developed to help doctors make sense of the clinical data [1]. Recently researchers have suggested artificial intelligence methods to optimise these scoring systems by selecting the best features from different models [2]. However, most clinical scoring systems still use static, cross-sectional patient data such as singular blood tests and vital signs at discrete time points, ignoring valuable information on trends, variation, and periodicity that can be derived from longitudinal data. Critical care is uniquely placed to explore longitudinal approaches because most ICU patients are monitored continuously and there is a growing trend towards monitoring more patient parameters more frequently.
Could doctors make more intelligent clinical decisions if the continuous nature of critical care patient data was harnessed? There is already evidence that heart rate variability is depressed in a variety of critical illnesses. Could similar real-time measurements of all patient data be used to improve predictions of which patients could be discharged safely from critical care, to better predict mortality or to better define the illness severity of all patients on the ICU at a given time?
Our aim is to develop personalised, predictive decision support tools for ICU clinicians, by deploying artificial intelligence methods for the real-time analysis of continuously measured data – such as ECG, invasive arterial pressure wave forms, pulse-oximentry and end-tidal CO2 signals, from ICU patients. Artificial intelligence and machine learning techniques are already in use in other industries to predict system and individual behaviours, to analyse trajectories and to provide early warning of events that deviate from the norm [3]. To date, their application in healthcare has been limited. The research will use integrated clinical data collected from patients on the adult critical care unit at Manchester Royal Infirmary. We will explore novel methods and techniques for: artefact removal; temporal analysis; outcome prediction [4]; and quantifying regularity (or irregularity) in cyclical physiological data [5].
https://www.research.manchester.ac.uk/portal/niels.peek.html
https://www.research.manchester.ac.uk/portal/mahesh.nirmalan.html
https://www.research.manchester.ac.uk/portal/en/researchers/andrew-brass(e142f83a-d5c6-4468-9cdc-e2eac86cd58d).html
Funding Notes
This project is funded under the MRC Doctoral Training Partnership. If you are interested, please contact the Principal Supervisor to arrange to discuss the project further asap. You MUST also submit an online application form - full details on how to apply can be found on the MRC DTP website www.manchester.ac.uk/mrcdtpstudentships. Interviews will be held w/c 2 July.
Applications are invited from UK/EU nationals only who have been resident in the UK for the last 3 years. Applicants must have obtained, or be about to obtain, at least an upper second class honours degree (or equivalent) in a relevant subject.
References
1. Vincent JL, Moreno R. Clinical review: scoring systems in the critically ill. Critical Care. 2010;14:207.
2. Pirracchio R, Petersen M, van der Laan MJ. Mortality prediction in the ICU: can we do better? Results from the Super ICU Learner Algorithm (SICULA) project, a population-based study. Lancet Respiratory Medicine. 2015;3(1):42-52.
3. Jordan MI, Mitchell TM. Machine learning: Trends, perspectives, and prospects. Science. 2015;349(6245):255-60.
4. Verburg IWM, Atashi A, Eslami S, Holman R, Abu-Hanna A, de Jonge E, Peek N, de Keizer NF. Which models can I use to predict adult ICU length of stay? A systematic review. Crit Care Med. 2017;45(2):e222-e231.
5. Monfredi O, Lyashkov AE, Johnsen AB, Inada S, Schneider H, Wang R, Nirmalan M, Wisloff U, Maltsev VA, Lakatta EG, Zhang H, Boyett MR. Biophysical characterization of the underappreciated and important relationship between heart rate variability and heart rate. Hypertension. 2014;64(6):1334-43.
Continue with Facebook
