A Belfast team took first place in Women Who Code's recent "Mission Predictable" global data science hackathon for Covid-19.
Women Who Code has been an invaluable lifeline for many during the Covid-19 lockdown period, running events and workshops for women in software engineering and the tech sector. The group's recent virtual hackathon tasked coders from around the world with using AI to solve problems caused by the Covid-19 outbreak.
A team consisting of Claire Burn, Mary-Jane McBride, Chloe Thompson, Claire Houston, and Lauren Burns representing Belfast's local tech scene won the event with its innovative machine learning prototype that could help hospitals predict serious complications in Covid-19 and ascertain which patients will likely need to be intubated.
The AI model used medical information such as whether the patient has asthma or diabetes or even whether they were pregnant to model the chance of complications requiring a ventillator. The model could help hospitals predict the numbers of ICU beds and supplies they're going to need during a spike in cases, and could be expanded to explore the effect of other conditions on the disease.
The team used anonymised data from a Mexican source on patients with Covid-19 that listed different co-factors such as obesity, asthma, and pregnancy. They built a website that could use a checklist of questions on admission to predict whether a person is likely to require intubation, which could help hospitals plan to get the supplies they need in advance and manage ICU bed numbers.
Source: Twitter
