• Galvanize | What does data science say about the coronavirus?

    By Wendy Gittleson for Galvanize

    The world is in the midst of what is potentially the deadliest pandemic since the Spanish Flu. While much of the world is on lockdown to prevent the spread of the SARS-CoV-2, or COVID-19, it often seems as if those in charge are flying by the seats of their pants. At the front line, healthcare professionals are risking their lives to save ours and infectious disease experts are working night and day to help develop treatments and a vaccine. Behind the scenes, though, data scientists are emerging as the unsung heroes of the pandemic response.  

    Flattening the curve

    Hundreds of thousands of Americans either have or have had the virus, which has put an unprecedented burden on our healthcare system. Until we have a vaccine, or at least effective medical treatments, scientists and front line responders are working to “flatten the curve,” which means we need to slow the transmission enough to ease the burden on the healthcare community. Flattening the curve does not mean the pandemic is over. In fact, if we relax after flattening the curve, infections could spike. Currently, medical science’s only way to slow the transmission is through sheltering in place and social distancing. Infectious disease specialists use a couple of metrics to determine how and when the curve will flatten.

    R0 number

    Infectious disease specialists in Spain teamed with data scientists and app developers to create a Flatten the Curve app. The app shows the “infection curve,” which predicts when the disease will peak and when the peak will fall. 

    The app uses an R0 (pronounced “Rnought” or “R zero”) number, which tells users how many healthy people a sick person can infect. If, for example, the R0 number is six, one person could infect up to six people. An R0 of one will cause the infection to die out. The criteria to determine R0 includes:

    • the proportion of susceptible people at the start and the density of the population;
    • the infectiousness of the organism;
    • the rate of disappearance of cases by recovery or death, the first of which depends on the time for which an individual is infective;

    Scientists currently estimate that the R0 number for COVID-19 is 1.4 to 2.5. This sliding scale can show you how long it will take for the virus to die out based on various R0s and responses. The R0 number varies from country to country and even city to city. The R0 number increases as population density increases. Highly infectious diseases have a lower R0 number than diseases that are difficult to pass on.

    Re number

    One part of the R0 number is fixed. It assumes that no one is immune from a disease. While studies are still preliminary, the latest research indicates that once someone recovers from COVID-19, they have immunity. The R0 number won’t change even if there is immunity, despite the fact that the number of people who will be able to contract the virus will shrink. The Re number (sometimes called the Rt number), or the effective reproduction number. The Re number is the number that can be infected at any given time. So, as more people contract the virus, and if people have immunity once recovered, the Re number goes down. Once the Re number is below one, and has remained there for two weeks (the incubation period), governments should be able to lift restrictions.

    Tracking the spread, once a site for data junkies wanting to learn about population growth, cars produced in a given year, cell phones sold, etc., has blossomed under the pandemic. Just three months ago, Worldometer’s web traffic ranked as the 8,794th most popular website. As of this publication, it sits at #49 of all websites worldwide. Watching the spread of the disease has become a worldwide hobby, but for professionals trying to stop the spread, it’s a necessity.

    The Center for Systems Science and Engineering at Johns Hopkins University created a live online dashboard for tracking the virus. It pulls data from the World Health Organization and the Centers for Disease Control as the disease progresses throughout the world. Not surprisingly, newsrooms, health organizations, and major search engines are all compiling data to help track the spread of COVID-19.

    While our current methods of tracking data are critical in the fight against the coronavirus, data scientists mostly rely on information provided by hospitals and testing labs. Asymptomatic people and people not sick enough to receive tests are generally not counted. Most of our current data is reactive. It does little to help stop the spread, other than to paint a picture for policymaking. In China, data scientists have taken a proactive approach. With the help of data scientists, China is working to identify and stop unwitting carriers of the virus. 

    Artificial intelligence scientists in China have developed a thermal scanner that can measure body temperatures. They installed the scanners in train stations in major cities. If their temperature is elevated and doesn’t drop after a few minutes, they isolate feverish passengers and alert the health authorities. They’re in the process of adding facial recognition to the technology.

    Contact tracing

    Since COVID-19 is asymptomatic for many people, and since testing everyone is currently unrealistic, one of the best ways to track the virus is through tracing every single person a COVID-19 patient has had contact with. Unless a patient was quarantined at the moment of infection, it’s nearly impossible for them to name every person they may have exposed to the virus. That’s where data science comes in.

    In 2011, two Cambridge University scientists developed a cell phone app that at the time was designed to prevent the spread of the flu. The FluPhone, which is an optional download, uses wireless signals to track the people you’ve been in contact with. If one of those people gets sick the app will notify you via text. Currently, only about 1% of potential users have signed up, but the developers believe the FluPhone could save lives during the COVID-19 pandemic.

    “The creators of FluPhone, Jon Crowcroft and Eiko Yoneki, certainly believe an app like theirs could help fight the coronavirus.

    ‘The health protection agencies could use it to populate anonymized map data,’ which might help reduce transmission, Crowcroft says. He says an app would also help researchers learn ‘how long the virus survives on a surface, what fraction of the population are asymptomatic carriers, and where to target critical medical resources.’”

    [source: Wired]

    US scientists are working on similar apps to the FluPhone. At MIT’s Media Lab, Professor Ramesh Raskar and colleagues are working on an app that would let people log their movements. The app will then compare their movements to the movements of known COVID-19 patients. 

    Others are asking Apple and Google to add to their phone tracking software that would only be turned on with user permission.

    There are a few drawbacks to cell phone tracking. While those advocating for greater tracking technology are reassuring people that the information does not have to go into a government database, people have privacy concerns. 

    Technologically, we may not be quite there yet. A phone’s location tracking is only accurate to about 20 – 40 feet, when the virus is transmittable to about six feet. This could mean a lot of false contact data. On the other hand, since the virus can live on surfaces for up to 72 hours, a person could come in contact with it at a gas station or grocery store and the app wouldn’t inform the user. 

    Herd Immunity

    The only sure way to lower the infection rate is through herd immunity. Herd immunity can be achieved in two ways. If enough people contract a disease and develop immunity, the Ro number will decrease, to a point where the risk will be very low. The problem, though, is that a lot of people can die before reaching herd immunity, and we don’t know whether people who’ve contracted COVID-19 develop immunity. 

    The second, and safest way to achieve herd immunity is to vaccinate as large a portion of the population as possible. Data scientists and research scientists use R0 to predict the number of people who need to receive the vaccine before reaching herd immunity. According to the Centre for Evidence-Based Medicine:

    “For example, if R0 = 2, immunization needs to be achieved in 50% of the population. However, if R0 = 5 the proportion rises steeply, to 80%. Beyond that the rise is less steep; an increase in R0 to 10 increases the need for immunization to 90%. Measles has an R0 greater than 10, which is why immunization of a large proportion of the population is so important in preventing the disease.

    Thus, if R0 is 10, a child with measles will infect 10 others if they are susceptible. When other children become immune the infected child who encounters 10 children will not be able to infect them all; the number infected will depend on Re. When immunity is 90% or more the chances that the child will meet enough unimmunized children to pass on the disease falls to near zero, and the population is protected.”

    [source: CEBM]

    When can states reopen?

    Several states are experimenting with full and partial lifting of restrictions, although experts warn that it’s too soon. Models from the Institute for Health Metrics and Evaluation at the University of Washington have been tracking the virus through deaths and hospitalizations. They believe that social distancing can be relaxed once there are fewer infections than one infection per one million people. According to the Mercury News:

    “Its updates on April 21 include an entry that has a date at which ‘relaxing social distancing may be possible with containment strategies that include testing, contact tracing, isolation, and limiting gathering size.’

    The institute’s website goes on to say: ‘The beginning of this timeframe is determined by our estimate of when COVID-19 infections drop below 1 per 1 million people in a given location, and is also influenced by each location’s available public health funding to implement new containment strategies.'”

    [source: Mercury News]

    Some European countries are loosening their restrictions, and they’re currently at risk of a second peak. Experts are afraid that the same will happen in the United States. Dr. Aaron E. Carroll writes in the New York Times, “Moreover, there’s a direct relationship between the number of cases that show up each day and the resources it will take to conduct contact tracing and isolation to prevent further spread of the virus. Locations that exit earlier will need more resources, not fewer, than those that wait until the number of new cases each day has been minimized. Almost no locations are prepared in terms of tests and public health personnel.”

    “We’ve engaged in policies that have slowed the number of cases,” said Ashish Jha, a professor of global health and medicine at Harvard. “We could have chosen to smash the curve, not flatten it. That would get us to a much more manageable place to reopen.”

    He and many other experts would like to see R numbers at one or below. 

    It’s not just medical professionals who are warning against opening too soon. So are data scientists. Researchers compiled economic data from the Spanish Flu pandemic in 1918, and they found that the longer cities held out from opening back up, the faster their economies recovered. They argue that it’s the pandemic, not the preventative measures, that hurt the economy. 

    Can we partially reopen?

    While the lockdowns are saving countless lives, the economy is in a tailspin. Before the pandemic, the US unemployment rate was about 3.5 percentApril’s numbers could rise above 16%. Some economists argue the real unemployment rate is closer to 25 – 35%. People are desperate to get back to work. Some businesses are safer than others, and some may need to rethink their business models. Restaurants, for example, could force social distancing by limiting their capacity, as could airplanes. Stores could continue to limit the number of people at any given time. Those at the highest risk, such as the elderly and the immunocompromised, may be advised to stay home until the virus is gone.

    Whatever happens in the short term, we will likely see some permanent changes as the virus dissipates. Working remotely could become the norm. Some parents might choose remote classrooms for their school-aged children. Basic doctors’ appointments might be held online, instead of forcing patients into cramped waiting rooms. While some jobs may never come back, home delivery jobs are on the rise, and retail workers have earned a whole new level of respect. 

    Tech professionals, and especially data scientists, have never been more in demand. Galvanize’s data science bootcamp can help prepare you to help prepare the world for the next pandemic.