Due to the current worldwide spreading of COVID-19, there is an imminent urgency to find strategies against this virus. Thousands of research papers are being written and have been published and even more econometric and risk models are developed about this novel coronavirus. Yet there is still very little known about COVID-19. In this blog we discuss how artificial intelligence can bring together the newly available data and help us win the global fight against the coronavirus outbreak.
“AI can have a major impact on the fight against COVID-19 in multiple ways; it can help limiting the spread of the disease, increasing disease prognosis, speeding up diagnosis and aid drug development.”
As we see many global initiatives collecting and sharing COVID-19 related data, the road is paved for data and computer scientists to develop new algorithms. AI can have a major impact on the fight against COVID-19 in multiple ways; it can help limiting the spread of the disease, increasing disease prognosis, speeding up diagnosis and aid drug development.
Infection and hospital occupation rate modelling
AI based risk analysis modelling have helped, and will continue to contribute, to forecast the spread of the coronavirus. Such modelling is needed to decide on national strategies like social distancing, lock down, or no intervention at all. It gives better insights on the impact of each measure on infection rate, hospital admission rates and provides a better understanding of the urgency of the situation. [i,ii]
Risk analysis and taking pro-active measures
Having a nationwide strategy reassuring that the public receives timely, accurate, and transparent information regarding the evolving epidemic is crucial for protecting citizens. A nation that had learned from earlier epidemics and uses big data to pro-actively and efficiently is Taiwan. Due to its proximity to and high number of flights from China, Taiwan was expected to have the second highest number of COVID-19 cases. However, being on constant alert and ready to act, Taiwan understood the required actions and implemented them quickly. To date, Taiwan has managed to prevent a large-scale epidemic. On top of being an early recognizer of the crisis, Taiwan introduced daily briefings to the public with simple health messages and other governmental measures. They utilized the government’s rigid household registration system and mobile phone data to build an algorithm that tracked individuals based on their recent travel history and grouped them as high risk.[iii]
Diseases prognosis
Risk modelling not only shows effective for populations but also individuals. Prognostic modelling, for example, can help choosing the right intervention for a patient and ultimately help reduce mortality of COVID-19. For instance, based on 2799 corona patients admitted to Tongji Hospital in China, researchers of the Huazhong University of Science and Technology built a prognostic prediction model based on XGBoost machine learning algorithm.[iv] The model identified three key clinical features (Lactic dehydrogenase (LDH), lymphocyte, and High-sensitivity C-reactive protein (hs-CRP)) from a pool of 300 features and predicted the chance of survival for patients with more than 90% accuracy. The model helps doctors with an early identification and intervention, thus potentially reducing mortality.
Speeding up diagnosis
Having a timely and highly accurate system of diagnosing coronavirus infections can help managing the spread of the disease. The ESoMII*, two Dutch hospitals (EZT, NKI-AVL) and two technology companies** started the developing a novel COVID-19 AI-based diagnostic tool. They launched a research project on training AI algorithm on CT images from corona patients, hoping to facilitate and speed up the diagnosis and understanding of the disease. ‘We hope this will allow us to make a correct assessment of the degree of lung impairment. Everything is now in the starting blocks. This will be an exciting European project, and we hope to achieve results within 1 month.‘ – comments Erik Ranschaerts MD PHD of the EZT on his Linkedin.
(*European Society of Medical Imaging Informatics, **Robovision and Quibim)
Drug development
AI can be used for the identification and validation of novel drug targets for the COVID-19 virus as this requires screening of large streams of medical data, such as ‘omics’ data and electronic medical records. This information is used for machine-driven iterative drug discovery processes involving data mining, hypothesis generation, lead compound identification and optimisation. Subsequently, by rapidly and accurately scanning millions of combinations, AI can predict the suitability of a molecule for drug development based on structural fingerprints and molecular descriptors. AI can be used to assess potency, selectivity and binding affinity towards specific targets.
Read our whitepaper
In our whitepaper and blog series we give you our vision on AI. We discuss the basic workflow of AI tools, we provide an overview of the opportunities and challenges of implementing AI in healthcare and give actionable insights to help the different stakeholders in overcoming barriers and developing smarter strategies that accelerate the introduction of AI innovations.
Download whitepaper AI
In this whitepaper ttopstart will give you insights on accelerating Artificial Intelligence for healthcare and life sciences. This paper will provide an overview of the opportunities and challenges of implementing AI in healthcare and give actionable insights to help the different stakeholders to overcome barriers and develop smarter strategies that accelerate the introduction of such AI innovations.