How we're using big data to tackle global health issues

July 18, 2020

We’ve recently published the results of two major research studies in the high profile scientific journal Nature Medicine: one studying symptoms of COVID-19, the other focused on food.

What do we know about responses to COVID-19?

The first showed how data from millions of users of our COVID Symptom Study app can be used to predict whether someone is likely to have COVID-19, even without a formal test. 

When the coronavirus pandemic came along, we saw that some people became seriously ill with COVID-19, while others were only mildly ill or had no symptoms at all. 

We quickly used our expertise in personalized data science and machine learning to develop the COVID Symptom Study app to understand why the disease affects people so differently, and with over 4 million contributors, it's now the world’s largest study of COVID-19 symptoms. 

Thanks to our symptom study data, and other research from around the world, we now know a lot more about the risk factors for COVID-19 than we did a few months ago.

Obesity, type 2 diabetes, and metabolic health are the three significant risk factors for worse COVID-19 outcomes. Genetics, gut health, and inflammation are all key players in determining how each of us will respond to the disease. And we also know that social factors like deprivation and ethnicity also play a role.

What do we know about responses to food?

The second paper details the findings from PREDICT - the largest study of nutritional responses in the world. 

Based on detailed data from more than 1,000 participants, we showed that everyone responds to food in their own way and that even identical twins, who share 100 % of their genes, can have very different responses to exactly the same meals. 

The best time to eat for nutritional health also depends on the individual rather than fixed “perfect” mealtimes - for example, we found that some people clearly metabolized food better at breakfast while others saw no difference.

We’ve also found that changes in fat and glucose in our blood after eating a meal differ hugely between individuals. Repeated, excessive blood sugar peaks and dips on a regular basis - for example, by always eating meals that trigger unhealthy fat and sugar responses - could underpin metabolic conditions such as obesity, diabetes and heart disease. 

Because everyone is unique in the way their bodies respond to eating food, this means that there is no one “right” way to eat. For example, someone who is particularly sensitive to glucose might need to cut their carbs, while someone else can eat them freely.

The power of big data

Both of these studies show the power of large-scale data to understand how the body works on a highly personal level, whether in sickness or in health. And they also reveal that there are plenty of similarities between how people respond to food and how people respond to coronavirus:

  1. People respond to food and coronavirus in a highly personalized way.
  2. The risk factors for severe COVID-19 are very similar to the risk factors for unhealthy responses to food (obesity, diabetes, poor sleep, inflammation, gut health and poor diet).
  3. One-size-fits-all advice about COVID-19 risks doesn't work, and neither does one-size-fits-all advice about nutrition.
  4. We can use artificial intelligence (AI) and machine learning to understand individual risk and predict how you, as an individual, are likely to respond to COVID-19. Using a similar process, we can also predict how you will respond to different foods.
  5. We can put all the power of personal prediction into an app that helps you monitor your health, understand your body, and decide what is best for you. 

There’s another important factor here too, which is the role of the billions of bacteria in our guts, known collectively as the microbiome, which can be altered by what we eat and plays a significant role in determining how we respond to food.

Although we don’t yet have firm data on how the microbiome influences individual COVID-19 risk, we do know that it plays an important role in immunity and immune responses. So there’s likely to be a connection there too.

So, what does this all mean for our health?

It’s time to think differently about food

Right now, nearly two thirds of the UK adult population and and around 7 in 10 adults in the US are classed as overweight or obese. 

Not only does this increases the chances of becoming overweight or obese, as well as the risk of developing conditions like diabetes and heart disease, but it also increases the chances of ending up in intensive care and dying from COVID-19, including for younger people. 

The healthier you are, the better your chances of remaining that way over the months to come and the better it is for us as a nation. But - as our PREDICT results show - simple "one-size-fits-all" diet solutions aren’t going to work when it comes to helping each individual move towards their best metabolic health. 

Our nutrition research has shown that improving your responses to food and your overall metabolic health relies on an intimate understanding of your own body. 

Using the results of our research, we’re developing a personalized home test and app to help you understand how your unique body responds to food and what you can do to improve and optimize your health. 


  • Two thirds of UK and 7 in 10 US adults are overweight or obese, increasing the risk from many health problems, including COVID-19.
  • We‘ve all started paying more attention to our health during the pandemic, so let’s seize the moment and get back on track. 
  • We all respond differently to food, so ‘one-size-fits-all’ approaches to tackle obesity won’t work. 
  • Modern technology like apps and cutting-edge data science not only help to monitor diseases like COVID-19, but they can help us all to understand our bodies better and optimize our health.

Find out more:

Discover your unique biology

Understand how your body responds to food with ZOE

Take the first step