Cardiopulmonary exercise testing, machine learning and major surgery
Research type
Research Study
Full title
Does machine learning improve preoperative risk prediction of patient centred outcomes after major surgery using cardiopulmonary exercise testing?
IRAS ID
316321
Contact name
Jason Mann
Contact email
Sponsor organisation
Clinical Research and Innovation Office, Sheffield Teaching Hospitals NHS Foundation Trust
Duration of Study in the UK
2 years, 0 months, 0 days
Research summary
Major surgery accounts for one third of all operations and is associated with significant postoperative complications, mortality and decreased quality of life. Assessing functional capacity, using cardiopulmonary exercise testing (CPET), represents one of the most consistent tools for preoperative risk stratification. However, this focuses predominately on post-operative mortality and morbidity, whilst a significant proportion of those suffering reduced quality of life following surgery are not identified using current risk assessment strategies. Improving risk prediction prior to major surgery allows optimal resource allocation and informs shared decision making.
Machine learning (ML) algorithms provide an option for increasingly complex models with higher degrees of accuracy. CPET produces 7,000-10,000 data points, yet only one to three data points are commonly used to assign a patient as high-risk. Combining all the raw data produced from CPET with ML may increase the accuracy of postoperative patient-centred outcomes, as well as mortality prediction.
An epidemiological study using ML to develop a predictive model of days alive out of hospital at 30-days after major surgery using preoperative CPET, patient demographics and operative data. ML derived models will be compared to traditional statistical modelling to assess for increased predictive power.REC name
East Midlands - Nottingham 1 Research Ethics Committee
REC reference
22/EM/0279
Date of REC Opinion
2 Dec 2022
REC opinion
Favourable Opinion