Categories
Blogs

The Fitbit Flu

By: Ross Stokes

In modern day society, mankind is faced with an infinite array of threats—threats such as: global warming, super volcanoes, and terrorist groups. There is one threat in particular that the general public tends to discount. That threat is none other than the influenza virus. Microsoft founder Bill Gates claims a deadly influenza pandemic is one of the biggest threats to humanity and it could kill 33 million people in six months (Rocketto). In 1918 the Spanish flu infected 500 million people and killed an estimated 40–50 million people. Without the innovations of modern-day air travel, the Spanish flu was still able to infect one third of the planet’s population. Since 1918, the world population has grown exponentially, which may increase the probability of a similar event.

The Evolving Threat

The influenza virus is relatively unique among other viruses. It is capable of antigenic drift and antigenic shift. Antigenic drift occurs when the surface proteins or “antigens” of the virus change, making it unrecognizable to the human immune system. This is why humans are able to contract the influenza virus on multiple occasions. Antigenic shift occurs when an influenza virus among an animal population gains the ability to infect humans, much like the swine flu outbreak in 2009.

When it comes to preventative measures, options are very limited. People can not simply just quarantine themselves within their homes during the influenza season. The next best option is in the form of vaccinations. Unfortunately, influenza vaccines are only capable of protecting against certain influenza strains and take a significant amount of time and resources to develop. By the time a vaccine for a new influenza strain is developed, a majority of the infections and casualties have already occurred. However, there may be a way to curb the spread of influenza infections using a commonly worn electronic device.

The Idea

Beginning as nothing more than some circuit boards inside a wooden box, Fitbit made its debut to the public in September 2009. Fitbit devices are worn on the wrist and are capable of measuring health metrics such as: heart rate, number of steps, sleep quality, blood oxygen, and more. It is possible that all user data can be aggregated and placed into a database where it can be easily accessed. The database can be used to build and deploy machine learning and AI models that use health metrics measured by Fitbit devices as independent variables to predict influenza infections, ideally before the user starts showing symptoms. This idea may seem unattainable, but strides towards predicting infections using machine learning have been made. According to healthanalytics.com, researchers at Carnegie Mellon University’s Heinz College are using a machine learning algorithm to more accurately predict sepsis—one of the most dangerous infections typically contracted in hospitals (Kent).

The Costs

Designing, building, and maintaining a database is a complex process that requires the skill set of a seasoned data scientist. Building and deploying machine learning models are no different. The costs of employing data scientists will be substantial unless the process can be automated. According to Indeed.com, the average annual salary of a data scientist in the U.S. is $123,716 (Indeed). There will also be infrastructural costs incurred over time. Storing big data in bulk requires expensive storage equipment (servers, hard drives, etc.) unless cloud storage options are utilized, but even they come at a cost.

The Benefits

Although the costs of implementing such an idea may seem unsurpassable, the benefits have the potential to change the world. Once a person is deemed infected or soon to be infected, certain precautions can be made to stop the spread of the illness. For example, imagine the machine learning detects an anomaly typically associated with the influenza virus in a Fitbit users’ health metrics. A notification can be sent directly to the user informing them that they may have contracted the influenza virus. According to WebMD, a group of researchers have been examining and analyzing Fitbit data, and have found that changes in resting heart rate and sleep patterns are sometimes indicative of the influenza virus. As more people choose to wear Fitbits, positive network externalities will surely follow. Machine learning and AI models will become more powerful with a larger sample size, which will be beneficial to current Fitbit users and may attract new users. Since all of the health metric data will be aggregated in a database, data scientists will be able to track the spread of influenza infections and stop them before they become deadly pandemics.

Feasibility

This idea will require a large amount of Fitbit users in order for the data scientists to build accurate predictive models. According to the LA Times, Fitbit has roughly 28 million users, which represents a fractional proportion of the world population (LA). This may be a sufficient amount to begin with, but more users will eventually be needed to maximize social gains.  Unfortunately, it may be easier to teach a monkey to recite Shakespeare while standing on a pedestal than get an adequate number of people to wear Fitbits. Fitbits are rather expensive and may not be affordable to people of lower income groups. This is an issue because poverty has been linked to poor flu outcomes. The Center for Disease Control claims that hospitalization rates for influenza are far greater for those living under the poverty line (Boyles). Ideally, the cost of Fitbits to those that can not afford them should be covered in the form of government subsidies or donations, but if government officials fail to see the potential benefits of the idea, funding is unlikely.

Building accurate and reliable predictive models will not be an easy task. It possible that the health metrics supplied by Fitbit may not be good predictors of influenza infections. Fitbit may need to add sensors that measure other health metrics to their devices, such as a sensor that monitors body temperature. This will require investment on behalf of Fitbit in the form of research and development.

Conclusion

A global influenza pandemic is a highly probable scenario that must be taken seriously. Simply getting whichever influenza vaccine is available during a given year is not enough. However, a universal influenza vaccine is currently in the research and development phase. CNBC mentioned in an online article that the group that has been making progress towards a universal influenza vaccine has received funding from the Department of Defense to begin human clinical trials (Woods). This will ideally be the ultimate solution with regards to the influenza virus. However, there’s still a large amount of uncertainty in developing a working universal influenza vaccine. There still needs to be another option in place to help prevent such a global event.

In theory the idea of aggregating Fitbit health metrics in a database for analysis seems like it may have potential. There will be substantial costs incurred, in the form of human capital and infrastructure, that will be difficult to overcome. Sources of government and private funding will likely be needed, especially when it comes to providing Fitbits for the poor. Although the costs may seem unsurpassable, the benefits have the potential to save lives and improve the quality of lives for people everywhere. As more people choose to wear Fitbits, more data will be collected, which will improve the overall functioning and continue to create positive network externalities. If the influenza virus can be accurately predicted, then it is possible that other illnesses like the coronavirus can be predicted also.

Works Cited

Rocketto, Leah. “Bill Gates Says a Deadly Flu Epidemic Is One of the Biggest Threats to Humanity. It Could Kill Nearly 33 Million People in 6 Months.” Insider, 29 Dec. 2018, http://www.insider.com/deadly-flu-epidemic-biggest-threat-bill-gates-2018-learnings-2018-12.

Kent, Jessica. “Machine Learning, EHR Big Data Analytics Predict Sepsis.” Health IT Analytics, 24 Mar. 2018, healthitanalytics.com/news/machine-learning-ehr-big-data-analytics-predict-sepsis.

“How Much Does a Data Scientist Make in the United States?” Indeed, http://www.indeed.com/career/data-scientist/salaries.

“Google Is Buying Fitbit for $2.1 Billion, Betting on Fitness Wearables.” LA Times, 1 Nov. 2019, www.latimes.com/business/story/2019-11-01/google-to-buy-fitness-wearables-giant-fitbit-for-about-2-1-billion.

Boyles, Salynn. “CDC: Poverty Linked to Poor Flu Outcomes.” Center for Disease Control, 11 Feb. 2016, www.medpagetoday.com/pulmonology/uristheflu/56157.

Woods, Bob. “Scientists Are Now Rushing to Develop a Universal Flu Vaccine.” CNBC, 26 Jan. 2016, www.cnbc.com/2018/01/26/scientists-are-now-rushing-to-develop-a-universal-flu-vaccine.html.

Goodman, Brenda. “Could Your Fitbit Help Detect the Flu?” WebMD, 17 Jan. 2020, http://www.webmd.com/cold-and-flu/news/20200117/could-your-fitbit-help-detect-the-flu.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s