Making hospital medical decisions safer

By Juliana Chan

SMU Office of Research & Tech Transfer – In an hospital intensive care unit (ICU), patients may require intubation, a process where an artificial ventilation tube is inserted into the airway and connected to a ventilator to assist with breathing during anaesthesia or sedation.

A common problem faced by patients after intubation is extubation failure, which is the inability to sustain spontaneous breathing after the ventilation tube is removed. In some cases, reintubation may be needed within 72 hours of a failed extubation.

Patients who are elderly, critically ill or suffering from pre-existing chronic respiratory and cardiovascular disorders are at increased risk of extubation failure. To prevent adverse patient outcomes, it would be helpful for intensivists to be able to identify patients at high risk of extubation failure so that they can be prepared to restitute ventilation early.

Supported by a 2/2019 Ministry of Education Academic Research Fund (AcRF) Tier 2 grant, titled “Optimal Personalised Medical Intervention Using Predictive Information in ICUs”, Associate Professor Daniel Zheng of Singapore Management University’s (SMU) Lee Kong Chian School of Business and colleagues are developing models to predict the risk of extubation failure in hospital ICUs.

“On the AcRF Tier 2 grant-funded project, we will integrate new predictive algorithms (predictive analytics) into decision-making models (prescriptive analytics) to provide more personalised decision support for physicians,” said Professor Zhang, an expert in healthcare operations management and analytics.

Together with Co-Principal Investigator SMU Assistant Professor Wang Hai and Collaborators Assistant Professor Ooi Oon Cheong and Dr Luo Haidong of the National University Hospital, the researchers will focus on extubation decisions by extending existing predictive models that only predict the current risk of extubation failure to also predict future risk.

Next, the team will develop a new decision-making framework that provides comprehensive decision support for extubation, utilising the predictive information to generate better early extubation recommendations for individual patients. Finally, the team will investigate the other ICU applications, such as those for rapid-response teams and ICU discharge, to test and validate the generalisability of the methodologies developed.

The AcRF Tier 2 grant taps on findings from Professor Zheng’s previous AcRF Tier 1-funded project. “In our earlier Tier 1 project, we successfully developed several models to measure and predict patient outcomes in the ICU, such as mortality and length of stay. The results include better risk stratification methods for real-time patient monitoring and a more rigorous estimation of treatment effects."

“The experience and results from our Tier 1 project provide a very strong foundation for the research team to explore more advanced predictive analytics combined with prescriptive analytics,” Professor Zheng said.

For more information, please contact:

Goh Lijie (Ms)
Office of Research & Tech Transfer
DID: 6828 9698
Email: ljgoh [at] smu.edu.sg

Back to Research@SMU Nov 2019 Issue