The Big Data the other coronavirus vaccine?

Must Read

Brian Adam
Professional Blogger, V logger, traveler and explorer of new horizons.

The ‘Big Data’, the other vaccine for the coronavirus?

The 'Big Data', the other vaccine for the coronavirus?

A Canadian startup predicted, before the WHO global alert, the global expansion that the already feared COVID-19 would have and also several areas to avoid

Can an illness be prevented before it is diagnosed? “With pandemic predictions, the problem is that you cannot normally make reliable estimates of the basic reproduction number of the epidemic until the disease has penetrated society and there has been enough time to examine the statistics,” says Antonio Romero Sebastiá, Director of the Master’s Degree in Social Communication of Scientific Research of the International University of Valencia.

Smartwatches and bracelets are the new allies of the research departments, but Twitter and Google were already common in the search and prevention of new diseases. Wearables monitor the seven days of the week and 24 hours a day. “They wear the patch that is equipped with various sensors for a week. And although it is discreet, it provides us with continuous information on heart rate, breathing, physical activity and much more,” explains Dr Frank Kramer, who works with biomarkers in the Bayer Cardiovascular Experimental Medicine group.

The pharmacist, in a study, has monitored various markers with this high-tech patch to track the health of patients. This analysis has been joined by mathematicians, engineers and computer scientists, who are already working with experts in epidemiology.

Epidemiology is the science that studies the spread of disease and has been using data for its predictions for decades. “The ‘Big Data’ contributes factors that are proving very important to improve the quality of the predictive models,” explains Pedro Antonio de Alarcón, one of the managers of LUCA, Telefónica’s data unit.

“The challenge with these methods is to distinguish between activity related to an individual’s illness and those related to the media or increased awareness and interest in the virus during the flu season,” says Jennifer Radin, a researcher from the Scripps Research Translational Institute, in an interview with CNN. He has written a study on leveraging portable device data to improve real-time surveillance for flu-like illness in the United States.

“Epidemiologists are forced to perform extrapolations based on a small number of the first data points obtained. Usually, it is not possible to have reliable estimates until the disease has started its course and there are relatively reliable data, “says Romero Sebastiá. De Alarcón also validates this argument: «the prediction problem continues to be very complex due to the number of variables that affect it; ‘Big Data and artificial intelligence (AI) have improved prediction accuracy, but have not resolved it. Furthermore, the fact that each epidemic is different from the previous one requires a great deal of work to be done in each new outbreak,” he adds.

Until a few years ago, experts used models that made pre-calculated estimates based on the population volume of cities, the distance between cities and other economic parameters, but not the actual dynamics of population mobility. “Our mobile phones are a source of information of tremendous value in this regard,” adds the person in charge of LUCA.

Estimates vs error

Artificial intelligence continues to take giant strides in the medical field, although skills and ethics are two workhorses yet to be fought. “Like all models of this type, the estimates are accompanied by an error, which is attempted to be minimized by improving the diversity and quality of the data and the volume of the samples,” adds De Alarcón.

Bluedot, a Canadian startup, alerted its clients on December 31 of the expansion of COVID-19, the Chinese coronavirus. This algorithm tracks news reports in foreign languages, animal and plant disease networks, and official proclamations. Thus he was able to warn to avoid danger zones such as Wuhan.

On January 9, the World Health Organization (WHO) notified the population of the coronavirus outbreak; the U.S. Centers for Disease Control and Prevention jumped to Day 6.

The Canadian tool “uses ‘Big Data’ analysis to track and anticipate the spread of the world’s most dangerous infectious diseases.” However, according to its CEO, Kamran Khan to Wired, “The algorithm does not use social media posts because that data is too messy.”

BlueDot predicted that the virus would jump from Wuhan to Bangkok, Seoul, Taipei, and Tokyo in the days after its initial appearance. To make the diagnosis, a team of 40 employees have designed the analytical disease surveillance program, which uses natural language processing techniques and machine learning to examine news reports in 65 languages, along with airline data and outbreak reports from animal diseases.

“One of the key aspects that determine the spread of an infectious disease is the mobility of people. Understanding the locations that will collect the most movements is key to predicting where the next focus of the disease will be,” says De Alarcón.

In 2016, the World Health Organization declared an international emergency due to the spread of the Zika virus. That year the Mosquito Alert’ app’ was one of the most downloaded in the Google Play Store. It captures more than 5,700 images of mosquitoes, and half are of the tiger species, a possible transmitter of Zika. It also has a web version, which collects on a map the points where citizens have sighted insects.

Telefónica, through LUCA, has also investigated the spread of this virus. «The ultimate goal will be to unite all these different data sources to obtain the most accurate and punctual image of public health. Predictions based on data can be correct, but they can also fail”, says Antonio Romero Sebastiá, a researcher at the International University of Valencia.