How social media can help predict disease outbreaks

European Pharmaceutical Manufacturer sits down with Derek Wang, CEO of data analytics company Stratifyd, to discuss the Covid-19 pandemic and how technology can be better used to predict and potentially curtail disease outbreaks.

A data analytics company, Stratifyd uses augmented intelligence to help businesses gain a better insight into things such as customer data, in the hopes that this helps them uncover hidden avenues which can lead to increased revenue streams.  The company aggregates all data sources into one platform that then uses augmented intelligence (basically the assistive use of artificial intelligence into enhancing human understanding) to uncover insights which can help businesses spend less time focusing on data and focus their efforts elsewhere.   

In Wang’s words, the platform has been designed to help “individuals and teams discover, interpret, and tell their data story.” Data are processed by a range of tools which includes a Natural Language Understanding engine and a Structured Data Learning engine to uncover any hidden signals within those data points. Following this, the Stratifyd platform applies an artificial intelligence (AI) model which automatically analyses the data stream. 

For the pharma industry, this means companies can use the system to monitor customer interactions, gain a better understanding into how people are taking their medicines, potentially helping improve adherence and also discover insights into clinical trials. Stratifyd has worked with a number of pharmaceutical companies and Wang’s experience is that they’re really focused on improving patient engagement. 

“They’re open to leveraging technology to improve their patient experience, meaning the patient’s usage of the drug and the impact for them,” Wang explains.

But how can a platform designed to help companies improve their data analytics and help generate more revenue be used to help the world prepare for pandemics such as Covid-19? Wang believes that by applying the Stratifyd platform to social media it can help gather data on symptom related posts which can point healthcare professionals and governments towards the populations which might be seeing the occurrence of an epidemic. 

“So we’re able to – leveraging our algorithms –start extracting terms like painful movement, shortness of breath, all those automatic key terms start surfacing up from twitter, facebook, news media and then being able to map those into geographic areas,” Wang says. 

Once that data is mapped, Stratifyd would be able to start predicting how those symptoms and patterns grow and what the geographical separations are for those people reporting symptoms. For Wang, it validates how effective augmented intelligence and machine learning can be in helping leverage automatically predictive models for viruses and diseases. 

Wang does iterate however that relying solely on social media data will result in bias in any predictive model, indicating that the use of Stratifyd is to explore untapped avenues, rather than provide a singular, sole solution for viral monitoring. 

So the use of Stratifyd in a pandemic situation is focused on preventative measures that can help curtail the emergence of a virus or disease. But while Wang doesn’t think there’ll ever be a way to fully prevent situations like Covid-19, there are certainly lessons to be learned. 

“Nature is fearsome. I highly doubt there will ever be a way to completely prevent outbreaks such as Covid-19 — but I do believe there should be ways to curtail the extent to which such viruses ‘spread’,” Wang says. 

“Our interconnected world has facilitated the speed and force of such diseases spreading globally. I think there are lessons to be learned in why and how we are still fighting this as yet uncontained global pandemic — one of the most widespread in recent human history.”

European Pharmaceutical Manufacturer first spoke to Wang just when Covid-19 was becoming known to the wider public and countries had started to implement the very first lockdowns. At that point Wang was paying attention to the situation in Italy, where cases had become high in number in mere days.  

At that point in time, Wang was seeing the need for a wider predictive signalling discovering technology that could help provide more consistent modelling around the world. It makes sense when we’ve seen countries respond to Covid-19 in vastly different ways, with obvious successes and failures. 

One form of concern comes from the way that data and information related to Covid-19 has been shared without any factual basis. 

It’s a point of contention for Wang who we followed up with in June via email. 

“In addition to information shared by news media, there has been a significant influx of information floating around social media in addition to news media,” he said. 

“But identifying factual data vs. myth about Covid-19 has been and remains very challenging. Consistency is also a challenge, depending on the channel broadcasting the information. Given the disparity of sources and varying levels of information provided across health organisations and government bodies, there isn’t a single, consistently reliable source of data.” 

With information on any topic only a click of a button away, separating fact and fiction is always going to be difficult, especially for those not affiliated in complicated areas such as medicine or healthcare. It’s a problematic area, no more so than the privacy concerns that come with any kind of monitoring technology. 

When asked about the privacy implications for Stratifyd, Wang argues that it is policymakers rather than those working on the technology, that need to assess how and when certain applications are used. 

“This is more of a policy and political challenge than a technological challenge. The technology is there but it is up to policy makers to decide the speed and manner in which it is deployed.” 

Wang does add that in the case of Covid-19, it’s “evident that countries with closer tracking mechanisms recovered more quickly than countries with less robust tracking.”

Recovery is a word that we’ll see more regularly as countries start to come back to some sense of normality. For the patients and families affected by Covid-19 recovery may unfortunately still be some way off. 

Though people will argue over the merits and deficiencies of countries’ responses to Covid-19, Wang offers a more placating tone. 

“First off, there is no perfect playbook in responding to such a pandemic. It has been an incredibly challenging time and the majority of governments have been working as hard as they can to respond and address the myriad issues resulting from the pandemic. I commend their efforts to date, which takes courage and decisiveness. The ability to plan for such deadly and rapidly spreading viruses in the future will require a consistent, centralised, and factual global network early on. This will ensure everyone around the world is working from the same data and can plan more effectively and respond in a consistent and measured way.”