The application of artificial intelligence within the healthcare industry has been relatively common for decades, with many PhD students applying machine learning techniques to various medical diagnostic tasks, but these have just been studies and typically not utilized for general usage or publicised. Only recently has the use of AI for medical applications been widely reported, with companies like DeepMind and Babylon Health, to name a few, developing medical applications and sharing publicly their advances.

However, some of the publicity has been negative with concerns of individual privacy and medical data being used without the individuals’ permission. While these concerns are absolutely valid, using aggregated and anonymised data can be exceptionally valuable for certain applications.

During times of crisis we must maximise our intelligence to benefit all, which is the fundamental ethos of the AI for Good campaigns that have focused on many different causes and benefits that the application of AI can deliver. Using advanced analytics to create insight and predictive models that can help us be the most efficient we can be with limited resources. While several companies are claiming to have discovered a vaccine for COVID-19, it will take many months for them to be tested and approved for use. Until then we need to find ways we can help reduce the number of deaths from the virus. There is already significant work being doing to collate useful data and research papers [1] from a number of various sources which will help in different applications against the virus.

However, let us acknowledge that in times of crisis, we can leverage the powers of advanced data analytics to the benefit of the humankind. Using aggregated and anonymised data removes many of the concerns of individual privacy and gives the data science community some useful datasets that we can demonstrate what advanced predictive analytics can help with.

Last week the government officials met with executives from the top technology firms at downing street[2] to discuss how AI can help the fight against COVID-19, and topics such as data sharing, fake news, drug discovery and robot cleaners where discussed. However, there are many other areas that AI can help.

We are already seeing the government implement plans to help reduce the peak of cases during a short time interval. Simple simulations[3] can show how this can be achieved by putting people with symptoms into self-isolation. This will help prevent any catastrophic spikes in demand that would be completely unmanageable because it significantly outweighs the available supply of hospital beds.

Ultimately, AI can be applied to many parts of any organisation, helping deliver efficiencies from the supply chain, routing of demand across the network of hospitals, diagnosis augmentation, as well as drug discovery, suggestions of medical care. But other departments in the back-office, such as finance, logistics and HR could also benefit from the use of automation and artificial intelligence, but obviously this is not a priority right now.

As well as being able to accurately predict the spread of the virus and determine the number of serious cases over time, helping to manage the capacity of the hospitals is an important aspect of ensuring we can match demand and supply of care.

The data science platform company H20 have been working with Stanford university, Kaiser and some other US hospitals to create a model that can help predict the capacity and demand for Intensive Care Units (ICU), this can then be used to manage the logistics around routing emergency cases to the hospitals with most capacity based on the regional predictions of cases over time. This is exceptionally important to ensure as many people falling seriously ill with the virus are able to receive the intensive care they need as soon as they need it.

Working in collaboration with Combined Intelligence, H20 are now looking to bring this predictive modelling into UK hospitals as part of our joint AI for Good campaign. We are looking to connect with as many UK hospitals as possible, so please help us, as we believe this modelling of capacity and routing can save lives.

H20 have pulled together an information deck on their work in this space and are happy to support UK hospitals with using their analytics, so if you are associated with a UK hospital and would like to have a copy please contact me directly and I will share with you and put you in contact with the H20 team.

Please share this article with your network, we are looking to reach as many executives within UK hospitals as possible to help them with this data analytics capability.

Footnotes:

[1] https://www.technologyreview.com/s/615367/coronavirus-24000-research-papers-available-open-data

[2] https://www.bbc.co.uk/news/technology-51851292

[3] https://www.washingtonpost.com/graphics/2020/world/corona-simulator/

[4] https://informationisbeautiful.net/visualizations/covid-19-coronavirus-infographic-datapack/