As another flu season approaches, health tech will play an educational role

1 Comment

The media may be humming with news of the Ebola crisis, especially with the recent death of Liberian Thomas Duncan in a Dallas hospital, but more people in the U.S. will die of the flu this year by several magnitudes. In fact, from 1976 to 2006, between 3,000 and 49,000 people died every year due to complications from the flu, according to the CDC.

What’s more, last year’s H1N1 epidemic saw more young and middle-aged adults killed than in previous years – people ages 18 to 64 accounted for nearly two-thirds of all flu-related hospitalizations and deaths, again according to the CDC. People who don’t take the flu seriously do so at their own peril, as healthcare workers at the neuro/trauma ICU at Portland’s Legacy Emanuel Medical Center told me as I toured the unit last week in the quiet before the flu storm. They see the worst of the worst from throughout the Pacific Northwest; guessing whether a patient will survive the flu can seem like a coin toss.

No single effort to combat the flu has been as impactful as the development of flu vaccines, first developed shortly after the virus was first isolated in 1933. But in recent years, various tech approaches are aiming to fill the information gap, boosting efforts to better inform the public about the virus and its spread, as well as where to get immunized and how to get help if you’re sick. From personal apps to big data mining and real-time interactive maps of outbreaks, the effort to revolutionize the public’s understanding of influenza has been extremely tech-driven.

“I don’t think we have something of the magnitude of the vaccine, that’s a pretty fundamental change that suddenly your immune system can deal with the flu,” David Lazer, Northeastern professor of political and computer science who has studied Google Flu Trends extensively, told me. “I do think, however, that it’s clear that if we knew where and when the flu would be bad that we could intervene in ways that save lives, so I do think this kind of methodology offers an approach that could allow us to use our resources more efficiently to save more lives. That’s clearly the potential. And I don’t think we’ve realized that potential.”

Help Remedies has a Facebook feature to track who got you sick -- or who is most likely to. A Gigaom staffer tried this with her account and was matched with a "prime suspect" from out of state, her boyfriend (who has mysterious check-ins) and two other friends who live in the same city. The app combs your Facebook for key words in statuses or check-ins.

Help Remedies has a Facebook feature to track who got you sick — or who is most likely to. A Gigaom staffer tried this with her account and was matched with a “prime suspect” from out of state, her boyfriend (who has mysterious check-ins) and two other friends who live in the same city. The app combs your Facebook for key words in statuses or check-ins.

First, the data

Nothing will likely ever rival the importance of the annual flu shot in minimizing deaths, hospitalizations, and the big bills that come with them. Modern public health campaigns to vaccinate everyone older than 6 months of age against influenza have ensured that the 1918 pandemic hasn’t happened again. (That was when H1N1 first reared its ugly head, infecting more than 500 million people worldwide and killing somewhere between 50 and 100 million, thereby wiping out nearly 5 percent of the planet’s population. Not surprisingly, people used to be a lot more cautious about the flu than they are today.)

Still, in most states less than 50 percent of the population gets the shot, due largely to lack of information and scientifically unfounded views about vaccine safety, according to infectious disease expert Paul A. Offit. (“The riskiest thing about vaccines is driving to the office to get them,” he recently told Time.) So apps that help people find nearby clinics with the shots, or warn people about outbreaks, actually serve to improve outcomes by simply getting more people to protect themselves via vaccines, isolation, better hygiene, etc.

Google trends measures the flu in San Francisco.

Like the vaccine, these approaches are far from perfect, offering potential that has yet to realized, as Lazer put it. Google Flu Trends caused a stir when its estimation of flu cases based on search results far overshot actual cases not just last year but over the past few flu seasons. In one study, “The Parable of Google Flu: Traps in Big Data Analysis,” doctors concluded that using weekly CDC reports from doctors on influenza-like illness, even though they lag by two weeks, outperforms Google Flu Trends as a predictor.

But co-inventor Matt Mohebbi, who left Google in 2013 to start his own company, Iodine, a drug info aggregator, said it wasn’t about outdoing CDC estimates but rather complementing them. “We’re only at the beginning of what’s possible with this big-data-style analysis,” he told the New York Times earlier this year. “And big data and small data are both very important.”

We’re all in this together

While Lazer agrees that the potential of big data has yet to be realized, he countered that if Google were more transparent with its data, as Twitter is, scientists could develop algorithms that improve Google Flu Trends and provide “finer granularity in their data.” This means that instead of offering data on six or seven regions in the U.S., as the CDC does, Google could be allow us to measure and predict outbreaks far more narrowly, perhaps even by zip code.

And the issue with projects using Twitter data is two-fold: First, there are lots of bots (“Do bots get the flu?” Lazer asked) that surely manipulate the data, and second, people who are well enough to be tweeting about symptoms may not be suffering from the flu (“When I had the flu, I couldn’t even flip the channels on the remote, I just lied on the couch like a dead person,” Lazer said).

FluNearYou.org shows a map of user-contributed flu reporting. (Source: FluNearYou)

FluNearYou.org shows a map of user-contributed flu reporting. (Source: FluNearYou)

Flu Near You, a community-driven flu tracker that offers real-time data to the public based on weekly reports that users submit through an app or email, has the potential to provide hyper focused, even neighborhood-level data about hospitalizations, immunizations, and symptoms (including whether people feel well, which is the box one ideally checks nearly every week of the year). But Flu Near You requires active participation, not simply the tracking of one’s tweets or searches, which will always limit its scope.

Beyond public health measures, personal apps have the potential to reduce the number of unnecessary visits to emergency rooms – and the spread of the virus in places where some of the most immune-compromised people are gathered – by empowering people to check their symptoms first. WebMD Symptom Checker is among the more popular, while Flu Defender offers not just symptom identification but prevention tips, recent flu activity, and a vaccine finder.

The CDC also releases flu data on its own app, but without symptom checking. And then there are apps like Task Rabbit to help those who are already sick delegate what they need to get done, or features like Facebook’s Who made me sick? to delegate, well, the blame (and suggest the person be quarantined.)

1 Comment

Nita.

Google flu trends is woefully inadequate as a tracking measure. It cannot distinguish between terms such as “I have a fever” and “I’m watching Saturday Night Fever” and is incredibly inaccurate. What we need is real-time geo-located data that is medically accurate. Let’s not forget that Google is NOT a health company. It is a tech company trying to get into healthcare, the next big thing. Let’s give the health guys a chance for once.

Comments are closed.