We know disease exists, where it comes from and how it evolves - but we don’t yet know ahead of time when something changes to mark the start of a new disease, a new variant or just an increase in transmission.
Often reports and reflection on major events will talk about key points in time where ‘intervention’ or specifically ‘earlier intervention’ could have happened to change outcomes. In respect of Covid-19, there are many, some debatable, points in time where as a collective, nations, governments, health professionals, scientists and data experts could have made earlier interventions to reduce the spread of the virus and ultimately save lives.
But that’s the challenge in itself. Without knowing when a disease is going to spread among the population, we are already on the back foot and playing catch-up to get ahead of what is often an incredibly complex and rapidly changing situation. This is where the world's health professionals and data scientists want to move from intervention to prevention -and why so many clinicians are now relying increasingly on digital health analytics and AI.
Covid is, of course, going to be the great example for how we spot and act to prevent future pandemics, epidemics and reduce the spread of disease in our population. We can learn so much from it.
Researchers at University of California San Diego School of Medicine, and at the University of Arizona and Illumina, estimate that the SARS-CoV-2 virus was likely circulating undetected for at most two months before the first human cases of Covid-19 were described in Wuhan, China in late-December 2019.
Chinese authorities cordoned off the region and implemented mitigation measures nationwide. By April 2020, local transmission of the virus was under control - but by then more than 100 countries were reporting cases.
That’s five months it took to get the virus under control within one area - but it was already too late to contain. This is the exact type of scenario where data is already being used to try and prevent similar delays from happening again.
Using data to monitor the behaviour of viruses is nothing new. During a Cholera outbreak in London in 1850, physician John Snow discovered, through data analysis, that the areas that were being served by a particular water pump were more affected than others. Shutting down that pump helped to control the pandemic.
Also before the days of computers, epidemiologists used marbles to model the spread of an infectious disease. A white marble represented a susceptible individual, a red marble an infected individual and a black one a recovered immune survivor. What the epidemiologists did was place a red marble randomly into a group of white marbles. Where the red touched the whites, those marbles became infected, and so on, while adding recovering marbles over time.
Although very basic, this is actually a good way to visualise what machine learning models do. The difference now is that there are powerful AI and analytical tools available to healthcare professionals, governments and scientists to not only report but predict the future spread of disease.
AI uses algorithms to analyse huge datasets - both structured and unstructured -to make accurate predictions in a matter of seconds or minutes.
During the pandemic we have seen the UK government consistently use data to reflect on how the virus is spreading and then scientists have used that data to understand the geographical spread and peak of the spread. And of course contact tracing has been used to reduce and slow transmission.
But this is using data on the go to track the behaviour of the virus retrospectively, rather than being able to predict events ahead of time. Of course, without a modern day pandemic to compare against, it was incredibly difficult to predict how the virus would spread and mutate.
Now, organisations like WHO (World Health Organisation) and Center For Disease Control have masses of public data that, alongside government and private medical businesses data, and even mobile phone network operators’ data, can all be powered by machine learning to give us real insights into how to prevent and limit future outbreaks of similar viruses.
SAS supports clinical organisations all around the world through its powerful AI and Analytics to help track and understand how infectious diseases spread, so the world can respond more effectively. As one example, we have partnered with PERSOWN. It is developing a novel point-of-care diagnostic platform that plans to provide instant test results using ultra low-cost disposable test strips and a handheld reader, fully integrated with a robust suite of electronic health records applications.
This platform will present a cost-effective option for testing in parts of the globe where access to sophisticated medical diagnosticsis scarce, potentially creating billions of integrated data points across various disease states and among previously undiagnosed populations.
Through the partnership and the use of visual analytic and data management software we hope that PERSOWN will be able to effectively mine diagnostic data to uncover infection trends and visualise disease hot spots to better monitor and predict outbreaks. These early insights will help governments and health organisations implement practices that reduce disease spread, mitigate impact and create early interventions.
Through AI and machine learning, data sources are connected mathematically to predict how viruses will behave. A model is the representation of these mathematical connections and more and more clinicians around the world will be using these machine learning models to effectively predict and prepare to prevent the spread of infectious disease.
Data scientists and health professionals are working tirelessly to finally answer the question, when does something change that marks the spread of a virus?