2022-02-19

 

Social media helps understand sudden death in epilepsy


In a paper in the journal Epilepsy & Behavior,  in a collaboration with Ian Wood, Rion Correia and Wendy Miller sponsored by the National Institutes of Health, National Library of Medicine, we demonstrate that social media, in this case Facebook, could be used to detect behaviors preceding Sudden Unexpected Death in Epilepsy (SUDEP), the leading cause of death in people with uncontrolled epileptic seizures. This  interdisciplinary work involving informatics/complex systems researchers, clinical/behavioral epilepsy scientists, and supported by the Epilepsy Foundation of America, brought to light findings that are highly relevant for people living with this chronic condition. The team is currently working on a personalized web service for epilepsy, myAura, which will include diverse clinical and non-clinical data, namely self-reported patient entries regarding seizures, medication adherence, and physician encounters. This easy-to-use web service will also include the option for users to donate their social media timelines, making this data more easily accessible for larger studies.

You can read the article following the links in reference:

I.B. Wood, R.B. Correia, W.R. Miller, and L.M. Rocha [2022]. "Small Cohort of Epilepsy Patients Showed Increased Activity on Facebook before Sudden Unexpected Death". Epilepsy & Behavior. 128: 108580. DOI:10.1016/j.yebeh.2022.108580. Preprint: arXiv:2201.07552.

Labels: , , , ,


2020-11-22

 

Privacy apocalypse or data science for the common good?

Longer, updated version of op-ed article published (in Portuguese): "Apocalípticos da privacidade vs. StayAway Covid e a ciência de dados para o bem comum." Observador. October 25, 2020. 
After almost a year of this pandemic, it is already clear that countries that best contain the spread of SARS-CoV-2 suffer less in terms of both deaths and GDP contraction−two quantities that are ultimately correlated. Various democracies in Asia have been particularly effective at containing epidemic spread−with corresponding great results in both fatalities and GDP contraction−via some combination of three capabilities: (1) border control, including internal movement among regions; (2) speed and accuracy in testing and tracking; and (3) firm quarantine enforcement.

Integration of these capabilities is even more effective against the pandemic when public health authorities use data from individual citizen behavior, which raises privacy issues. Truthfully, there is no effective strategy against an airborne epidemic that does not restrict some fundamental, individual right. The lockdown of an entire country is, after all, one of the greatest restrictions on all freedoms that can be imposed upon the citizenry. Yet, life itself is the most fundamental right, and death its most brutal restriction. Hence, any public health solution requires critical thinking to optimize temporary restrictions on fundamental rights to minimize death (and economic recession).

Contact tracing apps based on the Google/Apple Exposure Notification System, carry almost no privacy risk compared to popular social media, weather, GPS/maps, messaging, or shopping apps − or even to visiting news websites given their collection of cookie data. The big problem with contact tracing apps is that, just like masks, they are only effective if a large percentage of the population uses them. Hence, the pressure from governments to encourage their use. Since it is difficult to guarantee their correct use by a sufficient proportion of the population, however, many generalize arguing that, given privacy concerns, it is not worth using any form of individual data to control the pandemic.

Note, then, that while Europeans and Americans are facing or already experiencing new lockdowns, Asian countries that use individual data to implement the above strategy have their societies essentially open. This is the case in Taiwan which not only used the power of its crowd-sourced digital democracy to be among the first countries to detect and respond to the initial outbreak, but also firmly enforces quarantine and monitors health status (of all arrivals at the border and positive cases) via mobile phone surveillance. It similarly makes great use of digitally assisted contact tracing using data from the major telecommunication companies, police, health records, and other sources, while furthermore making excellent efforts to ensure confidentiality.

Another country making extensive use of individual data is South Korea, where the government uses mobile phone and GPS data, bank card transactions, GPS, and even video surveillance networks with facial recognition, to quickly and effectively trace contacts of positive cases and thus contain outbreaks very successfully, all in a manner ethically proportionate to the crisis in terms of privacy.  South Korea had its first outbreak two weeks before Italy, with the two countries having very similar case numbers in early March. However, South Korea very quickly managed to contain the contagion, but Italy (followed by all Western countries) did not. Since then, new outbreaks continue to be rapidly controlled in South Korea, keeping the country essentially open, but in the West there is no such capability — only full or aggressive lockdowns can bring transmission and death numbers down.

The Asian response was markedly better early in the pandemic, but at the time, many in the West suggested that cultural factors such as wearing a mask were likely the cause of the difference. As we can see now, even with widespread use of masks and social distance, Europe and the USA are experiencing a much larger increase in cases. It becomes clearer now that Korean success, like other Asian democracies, comes from legislation democratically enacted following the 2015 MERS outbreak that allows for tighter surveillance in the event of an epidemic.

Still, when one talks about the success of Asian countries in this pandemic, it is common to hear: “but theirs is a different culture, we Westerners would never accept such an invasion of privacy.” Yet, this belief does not survive critical analysis, and may even suggest prejudice. First, the same notion was often heard in March regarding the use of masks, and now we all use them — necessity may know no law, but we can certainly make a virtue out of it. More importantly, Western democracies have already developed systems that go further in tracking citizen behavior. One could speak of surveillance systems implemented to counter terrorism, but even more fundamental is the entire tax collection system that allows governments to monitor bank transactions, income, and much more. We accept this interference in our most private data because it is a price we are willing to pay for the common good.

Similarly, the transaction made by the South Korean society, in which it accepts a temporary increase in surveillance so that the State can better respond to pandemics, is in all respects equivalent. Why shouldn't Western democracies want to use the same social contract to save lives and the economy? Why not pursue such deliberation democratically? One problem is that it is now in fashion to reify privacy as if it were an absolute right, the most fundamental right of individual freedom.

But even in a western liberal democracy, it makes no sense to render the right to life of risk groups less important than the right to not reveal that an anonymous infected person was less than two meters away in an anonymized place. It makes no sense to martyr the right to gather in bars, clubs, and other places of freedom, for privacy concerns over quarantine enforcement. It makes no sense to sacrifice whole economies to the altar of free movement by failing to (temporarily) control borders and travel between internal regions.

As Asian countries did after previous pandemics, we must defend our society better in this pandemic by preparing for others, likely more deadly, that are sure to emerge. Paraphrasing John Stuart Mill, freedom requires an appropriate adjustment between individual independence and the common good. In the context of a pandemic, this temporary adjustment certainly needs a non-absolutist conception of individual privacy and freedom, since it is individuals who spread the virus to each other.

There are no silver bullets in epidemiology, but there are better and worse results. While much more interdisciplinary science is needed to study the effectiveness of every factor, the difference in outcomes is striking. When privacy advocates frighten us about the imminent authoritarian apocalypse that using citizen data to combat the pandemic will bring (including by using exposure notification apps),  what they don't say is that in countries where entries, movements, and quarantines were controlled with such technology, deaths are orders of magnitude lower than ours (see figure), society and its spaces of freedom are largely open, and the economy has suffered a much milder recession. It is said that Westerners do not accept government intrusion into their data —naiveté and denial for anyone using iOS, Android, or paying taxes. I will gladly share my proximity and location data to get to the 1 or 10 deaths per million in Taiwan and S. Korea, respectively, rather than over 1200 deaths per million in Belgium, and even worse in several US states. Not because I want more surveillance, but because I want greater freedom. Indeed, those who enjoy greater freedom at this time are the citizens of Japan, Taiwan, South Korea, New Zealand, and others in the Pacific.

Deaths by cases of COVID-19 per million inhabitants o November 22, 2020. Vertical axis (deaths per million) on a logarithmic scale; Taiwan with 0.3 deaths per million inhabitants, South Korea with 10, Italy with 815, UK with 803, Belgium with 1337, and New Jersey with 1900 deaths per million inhabitants. Highlighted line markers: 20 (green), 200 (yellow), 400 (orange), and 1000 (red) deaths per million inhabitants on vertical axis, and 10000 (red) cases per million inhabitants on horizontal axis.  Some US states shown separately with blue dots. Take a moment to grasp a death toll 100 to 1000 times worse in the West in comparison to Asia-Pacific countries. Data from:  https://www.worldometers.info/coronavirus/#countries

Labels: , , , ,


2017-12-22

 

The Sweetest Taboo is Christmas

The year ends with tons of media attention, a bit of media whoring, really: The Independent, Time, Publico, GizModo, The National Post, Men's Fitness, Mother Jones, ArsTechnica, The Sun, Drive with Yasmeen Khan (at 17:30) (audio of interview), etc.

What is important is that it refers to a paper I'm really proud of. First, because it was a great collaboration that started with Joana Sá at Gulbenkian, who had the idea to look at the birth peaks problem. Then it involved Johan Bollen, a long time partner in crime, and Ian Wood, one my students who has been a real pleasure to work with. I am very proud with our demonstration that Data Science, Computational Social Science, and Complexity research can be used to obtain surprising results and invalidate previous hypotheses. In other words, Data Science that does not simply validate our pre-existing biases and accepted wisdom, but demonstrates new explanations. In this case, we found that online interest in sex rises at Christmas and other cultural celebrations, with more births nine months later, in what is the first global analysis of human birth-rate cycles. Another bonus was that to reach our results, we developed a novel spectral methodology for sentiment analysis on social media. To make things even better, it is fun Science dealing with the Sweetest Taboo, as all the media attention shows and Sade knew all along: Everyday is Christmas, and every night is New Years eve.


A post shared by DJ Angst and E-Trash (@riotbootique) on

Labels: , , , ,


This page is powered by Blogger. Isn't yours?

archives