Adversarial Preventative Strategies for Infection-Propagation Control in hospitals

Centre for Smart Systems, AI and Cybersecurity (SSAICS)

We will leverage Machine Learning and Deep Learning capabilities to achieve robust detective and preventative strategies in order to understand infections spread and outbreaks in healthcare environments and minimise their impact

With the recent and increasing strain on public health care facilities, Infection Prevention and Control (IPC) has proven to be a challenging task. Despite the thorough infection-prevention protocols currently put forth, outbreaks are still critically probable, largely due to improper implementation of the procedures (i.e. human error). For instance, according to the NHS' public manual for Standard Infection Control Precautions (SPICs) and Transmission-Based Precautions (TBPs), all decontamination and cleaning/disinfection procedures are assessed solely by the healthcare worker carrying them out, leaving tremendous room for ineffective application.

Hospitals tend to have a large historical data related to infection control, but is not being exploited for preventative measures. This project will work on this historical data by leveraging Machine Learning and Deep Learning capabilities to achieve robust detective and preventative strategies in order to understand infections spread and outbreaks in healthcare environments and minimise their impact

Currently, control of infection spread in hospitals is a reactive activity, however, what we aim at is to be more preventative and predict potential outbreaks before they happen, in addition to identifying the main factors and behaviours which contribute to the spread of an infection. This is a challenging problem for any AI model as it involves multidimensional data.