Leveraging data to combat the (next) pandemic

Written by
Willem Prisse
In the past few months we’ve all been affected by the COVID-19 pandemic in some way, shape or form. Some have suffered irreparable damage while others have had to adapt only slightly. The sad fact of the matter is that we have lost over half a million people to this pandemic and are still counting. Surely, we all agree that a similar scenario needs to be prevented or, at the very least, combatted as effectively as possible in the future.
In these times, where healthcare and government systems are buckling under the strain of COVID 19, every ounce gained in efficiency is one that is welcome.

Now, more than ever, the importance of sharing accurate and trustworthy information has become clear. Had the danger of COVID-19 been more clear in an earlier stage, different precautions could have been taken, possibly saving many their livelihood and some even their lives. The social systems (healthcare and government) implemented to catch us in these times have been under immense pressure. Mercifully, in some countries the virus has started to slow down, but in others it has only started gaining traction. We must learn our lessons from both situations, based on the data they produce, to better understand the pandemic. In these times, where healthcare and government systems are buckling under the strain of COVID 19, every ounce gained in efficiency is one that is welcome.

In this article, the importance of sharing and gathering data during this pandemic is highlighted. Using the right data sources, we can develop methods to combat the pandemic and perhaps even predict its next move. We’ll illustrate this importance through several examples in which data is leveraged, which could help fight this or the next pandemic.

Not knowing where infection rates are rising and how it is spreading can mean that important resources aren’t allocated properly and with the necessary urgency.

Leveraging data for…

Outbreak Tracking and Forecast: Accurately tracking the way the virus spreads through contagion means we can better fight it and prevent unnecessary contagion. Recently, India’s infection and death rate shot up in 24 hours due to a back-log of information on infections and death rates. Not knowing where infection rates are rising and how it is spreading can mean that resources aren’t allocated properly and with the necessary urgency. Besides the official data streams i.e. reports from the WHO and hospitals, there are alternative data streams that can be tapped into that could help track the spread. A Canadian startup BlueDot strives to provide such a service. They tapped into public social media, news reports and government documents, and aggregate those data sources. By doing so, they were able to send out warnings as early as December about the then little known virus brewing in Wuhan. Sadly, those warnings were not heeded and the result is the situation we find ourselves in right now.

Diagnostic assistance: Data driven technology has already made its entrance into the diagnostic division of healthcare. As more and more data was being collected on COVID-19 patients, data driven diagnostics also started gaining traction. Infervision is among the leading companies in the world pushing this method of diagnostics. Their application can help radiology departments in hospitals significantly shorten their diagnostic process of patients. Flooded ER’s and overworked healthcare workers need every extra second they can get their hands on. These initiatives were able to help front-line healthcare workers diagnose and monitor the disease per patient more efficiently. Besides that, patients are now able to get their diagnosis in a matter of minutes instead of hours. Not only that, but the data supported diagnosis system was able to do so at a 95% accuracy rate. This saves all parties involved valuable time and streamlines an important part of a hospital visit.

Vaccine Development: Producing a vaccine is a massively time consuming process. Vaccines work in such a way that ‘parts’ of a disease are presented to the immune system of a human. This is done so that the immune system can more easily recognise and combat that disease in the future. Think of it as a friendly match to get to know the tactics of your opponent.

The two main parts that consume the most time are molecule sequencing and clinical trials. Generally, the development of vaccines takes anywhere from 5 to 10 years. The mumps vaccine for instance -considered the fastest ever approved — took four years to go from collecting viral samples to licensing a drug in 1967. However, with the right data and a good amount of computing power, this process can be significantly shortened.

The Argonne National Laboratory managed to narrow down the sequence of molecules of a possible vaccine, from a couple of billion to several thousands, in a matter of weeks. After that, they ran simulations on how molecules interacted with virtual human proteins. This use of data and virtual trialling does not mean that animal and/or human testing will be thrown out all together. However, it does mean that this phase of vaccine development can be greatly sped up and will need far less ‘human guinea pigs’.

A more ‘behind the scenes’ application of AI can also help researchers. Due to the enormous amount of literature that is being written every hour, it can be a never-ending story for researches to determine what has already been discovered. The Allen Institute for AI has developed an extensive tool that enables researches to sift through the tens of thousands of articles on COVID-19.

Information spread by chatbots: Although we live in en era where we have never been as connected as we are now, it still seems to be a struggle to spread accurate information on the COVID-19 pandemic. For instance, in the Netherlands there was a lot of confusion about what was or was not allowed under new regulations due to COVID-19. Data leverage might again be able to help here. Conversational technology, also known as chatbots, is something most of us are already familiar with. Tools like Siri or Alexa are part of many’s daily life. Chatbots have several advantages: First of all, they are available 24 hours a day, 365 days a year. Second, the information they provide can be tailored to the need of those who ask for it. Third, the information they provide is adaptable. This is useful because often, each country, state or city has their own laws, rules or guidelines for the pandemic. However, there are several challenges to overcome before chatbots can adequately support in dissemination of accurate information. In order for this solution to work properly there needs to be a support base. To create that support base, there needs to be trust.

Besides spreading information, chatbots could be of assistance in gathering information as well. Triage is an important part of the journey that a (potential) patient has to undergo. Healthcare systems have been under immense pressure due to the rise of patients seeking medical help. By using chatbots to assist in the initial triage phase, doctors, nurses and other medical personnel resources can be diverted elsewhere.

Targeted Loans: After addressing implementations that are directly linked to fight against COVID-19, we must not forget that there are many indirect consequences as well. The economy has been impacted by the pandemic and as a consequence of that unemployment has skyrocketed. The United States hit an all time high of nearly 15% percent in April of this year.

Since then, a 2 trillion (no, that’s not billion) dollar relief act has been signed in. However, once that bill was signed the hard work really began. Who gets that money? How much do they get? About a month ago, Google announced the launch of their PPP Lending Solution Data technology. This technology would help the government to determine which companies wouldn’t survive without a loan. Automating this process will allow banks and the State to shorten approval times and get loans out more efficiently and effectively. Subsequently, the technology would prevent malevolent applicants from receiving loans.

If we look at the relief programs put into play by the US through the Small Business Administration, their intent was to loan small businesses a short term low interest loan to survive this period of economic downturn. However, through loopholes many companies with well over 100M in market cap were able to poach those loans as well.

Leveraging the correct data could very well support those struggling systems in their functioning.

In conclusion

Summarised, the above method to leverage data to support during this pandemic can be of value in two domains: Efficiency and Accuracy. Simply put, there is a lot going on and it has put a lot of extra strain on already strained systems like the healthcare system and the governmental system. Both are struggling to keep up and any tool that can enable them to do their jobs more efficiently should be considered for development. Accuracy has also become an issue due to the rise in strain on systems designed to catch society when it falls. Leveraging the correct data could very well support those struggling systems. As seen in the US, due to the sheer size of the problem, even initiatives that mean well can unintentionally benefit those who don’t qualify for them. It must be said that leveraging data definitely would not solve all the problems that we face at the moment. However, in these times any support that makes any difference cannot be ignored.

Finally, it is important to realise that all of the above cannot be accomplished at the the flick of a switch. Leveraging data, more importantly the correct data, takes time and effort. There are many challenges that will have to be overcome for effective data leverage.

If this article sparked your interest in data leverage and using it to tackle the pandemic, be sure to keep a look out for my next article. In that article, I will highlight the obstacles that data leverage on a global scale might face.

Lege plus

Continue reading...

newsletter

Want to stay up to date?

Sign up for our newsletter, and we’ll keep you posted on our research, podcast and other AI goodies.
* We don't share your data. See our Privacy Policy
Thank you! You've subscribed.
Oops! Something went wrong while submitting the form.