Android gene test presages future of data privacy

By: Jonah Comstock | Mar 19, 2013        

Tags: | | | | | | | |  |
Photo by Steve Zylius / UC Irvine Communications

UC Irvine researcher Gene Tsudik (Photo by Steve Zylius)

A team led by University of California-Irvine researchers has developed what’s been called an Android-based smartphone paternity test. The app would allow two people, if they had access to their digitized genomes, to run a paternity test comparing their genomes, determine how closely related they are, and even scan their genome for predictors of different diseases. The team also includes researchers from the Xerox Palo Alto Research Center (PARC) and the New York Institute of Technology.

According to researchers, though, the real innovation of the platform, called GenoDroid, is that it allows people to run tests using their digitized genomic data, while maintaining data privacy — not sharing their genome with anyone, even the other subject of the test.

“Clearly if you reveal a genome, anyone can compare it to another genome,” researcher Gene Tsudik told MobiHealthNews. “But if you give away your genome, you’re giving away a treasure trove of very, very valuable information.” And, he added, privacy concerns around genomic data are even more complicated than that.

“The problem is, [if you make your genome public] you’re not really just hurting yourself, you’re also cursing your own children and giving away information about your parents and siblings,” he said. “We believe ethicists and legal experts will have to be involved in this discussion.”

One such legal expert, Daniel Vorhaus, a lawyer with Robinson, Bradshaw & Hinson and editor of the Genomics Law Report, disagrees with Tsudik on the danger involved, suggesting there’s little evidence that a personal genome is especially valuable to anyone but the holder in most cases. In fact, more than two thousand people have opted to make their genome public, including Vorhaus himself, through the Personal Genome Project. GenoDroid was even tested using genomes publicly donated to science.

The conversation about the need for genomic privacy is an ongoing one, but it’s worth noting that many leading thinkers don’t agree that genomic privacy is as essential as Tsudik, admittedly a computer scientist and not a genetic expert, suggests. The position of the GenoDroid research team in their paper is that privacy concerns have at least been raised — because genome data “contains information about one’s ethnic heritage, phenotypic traits, and predisposition to numerous diseases and conditions, including mental disorders” and can’t be truly de-identified or anonymized because it is “the ultimate identifier.”

Of course, Tsudik is quick to point out that even though GenoDroid is a functional, usable app, it amounts to little more than a proof of concept. This is because it’s currently very expensive to have your whole genome sequenced, and very few people have access to that information. Even people who use services like 23andMe are only paying for digitization of certain parts of their genome, not of the whole thing.

In 2009, we reported that Illumina was offering full-genome sequencing for $48,000. According to their website, they now offer the service for $10,000. But most people in genomics agree that technology will continue to get cheaper.

“I see a company [like 23andMe] in the next 5 to 10 years offering the same service, for the entire genome, for $99,” Tsudik said. He also said that a human genome is too large a data set to store on a smartphone right now, but not too large to store on a desktop or an external harddrive, which the smartphone could then access via the cloud.

To run a paternity test, GenoDroid only needs to compare a small part of the genome, so the app can maintain some privacy by only sharing that portion of the data with the other devices. But Tsudik said the app will later be upgraded to run more complicated tests — for instance, two people who want to have children together might run a genetic compatibility test, or two people might run an ancestry test to see how much of their genome is shared. These tests would require more of the genome to be shared. For that reason, the app also encrypts the data and the only information it returns to the user is the result of the specific test.

One other area where Tsudik said the tech could be helpful is in the burgeoning field of personalized medicine. Pharmaceutical companies are increasingly developing cancer drugs that only work on people with particular genetic markers. The companies sometimes consider certain information about these markers to be trade secrets, which could lead, Tsudik said, to a data privacy stand-off.

“About a decade ago there was a Supreme Court decision that said you cannot patent genetic material, which means a pharma company can’t patent a genetic marker. This brings us to this conundrum — the pharmaceutical company has markers that are secret, the patient has DNA which is secret,” he said. Tsudik didn’t get it quite right in terms of the legal realities — the validity of gene patents is far from decided. According to this WSJ report from the end of last year, a case is pending that could provide more clarity, but, Vorhaus says, even if the finding bars certain gene patents it won’t be so broad as to rule out gene patents entirely.

Nonetheless, GenoDroid or similar apps could bolster personalized medicine by sidestepping the need to share such sensitive information. Tsudik said there are no plans to commercialize GenoDroid at present; the paternity test app is being made available (for free download in the Google Play store) for “a few well-meaning geeks in bioinformatics” to “play with,” and subsequent versions will be made publically available as well.

“I’m a security and privacy researcher,” said Tsudik. “Our goal is to look into the future and see what spheres of life will become digitized, computerized, and what privacy concerns will come out as a result. The real umbrella goal here is to anticipate privacy concerns in this brave new world where everyone has access to their own genome. People will lose their genomes, hackers will break in and steal them, and when that day comes we want to be prepared for it.”

UPDATE: This story has been edited from its originally published version to provide more balance and clarity about the complex issues involved in genomic law, thanks to prodding from Vorhaus.

Advertisement

Liveblog: Congressional hearings on FDA regulating apps

By: Brian Dolan | Mar 19, 2013        

Tags: | | | | | | | |  |

We are live blogging old school today: Hit the refresh button every few minutes for updates! (Live blog is now concluded.)

12:24PM: Rep. Ellmers very quickly concludes the hearing. That’s a wrap! Next hearing is slated for tomorrow at 10AM ET.

12:22PM: Rep. Ellmers asks if medical device tax will hamper innovation. Thompson, Dagi, and Ford all said yes. Others said unclear. Rep. Ellmers says she is a nurse and understands the need for patient safety. She asks all panelists that if an app is regulated as a medical device by FDA should the medical device tax be applied? Thompson said his group opposes it being applied. Most said it was unclear. Ford said it was two totally different questions.

12:20PM: Rep. Ellmers asks Thompson how long it takes to secure 510(k) clearance. He says it depends but usually between 90 days and 1.5 years. (MobiHealthNews found average time for mobile medical apps clearance has historically been about 110 days).

Ellmers

12:18PM: Rep. Gardner says this is not clarity to him. Dagi says we do not have the necessary clarity. Jarrin says it depends. Gardner says if it depends then it is not clarity. Jarrin says there is insufficient clarity for low-risk medical devices. Spalter says clearly there is not sufficient clarity. This panel of expert witnesses is having trouble, imagine what the developers who are trying to dream up these innovations is going through. Let’s put down on paper what needs to be understood and once we do the hockey stick of medical apps will go into the hundreds of thousands of medical apps. Keep reading>>

Readers react to sleep company Zeo shutting down

By: Brian Dolan | Mar 19, 2013        

Tags: | | | | | | |  |

Zeo Personal Sleep CoachLast week MobiHealthNews broke the news that sleep tracking and coaching company Zeo had closed up shop and was looking for a buyer. With more than 600 shares on Twitter and dozens of comments, it’s already one of our most discussed articles of the year. There were many worthwhile comments on our story, below are a few of our favorites from MobiHealthNews readers. These were edited for typos and condensed in some cases:

UPDATE: Whitley Howlit: As a former employee, one of the last in fact, maybe I can clear a few things up. First, as far as validation of the data, a simple Google search for “Zeo Validation” shows: http://www.myzeo.com/sleep/faqs/general-questions/how-accurate-zeo Secondly, at the end, Zeo Bedside was sold at $149.00 and the Zeo Pro was at $99. Please notice your price of $400.00 is an Amazon price and in order to understand Amazon pricing, you need to understand the Amazon Algorithm that determines pricing based on supply and demand: Less Supply + More Demand = Higher Prices. If the factory is in fact not producing anymore, demand go through the roof. Third, I really wish I could share information on all of the Universities, science and medical research teams, insurance companies, and government agencies both foreign and domestic who relied on Zeo for their sleep data, but unfortunately due to NDAs, I cannot. Additionally, if it was classified as a medical product, it wouldn’t have been available to consumers direct. Doing so would have been cost prohibitive for both Zeo and the end consumer.

Ejstensjn: Awful news. What I learned from this discussion: Investors do what other investors do. This is what creates fads and bubbles. Zeo failed to explain their product. Fitbit and others do nothing of the sort. Zeo: Great idea, but an idea before its time and without a valid business model. Sincerely hope someone else picks up the torch.

Dion Madsen: My perspective, as an investor who diligenced the company several times was that it never passed the “so what” test. Meaning, you buy the device because you have trouble sleeping and then you wear it and you confirm that you have trouble sleeping. So now you just confirmed something that you already knew, so what do you do next? Another aspect that the company tried to play up is the billions of “sleep nights” data that they stored in their database that surely had value. Well, I haven’t seen health companies turn these kinds of unstructured patient data sets into information that other people are willing to pay for.

Anonymous: In medicine we are drilled to only accept solutions that have a study (and preferably a meta-analysis) that shows statistical significance. Then we drill into the methodology, who funded the research and where it was published before letting it pass mentally as clinically successful. Everything else gets stonewalled with a ‘there’s no evidence’.

Gwern: Zeo is massively cheaper on a per-night basis than a polysomnograph, so one can get orders of magnitude more nights’ data. Between the better validity and the greater quantity of data, even if the Zeo analysis was only medium-correlated with a polysomnograph, it could easily deliver much better answers on some questions. And saying ‘no clinical evidence base’ seems exaggerated; as the article mentioned, Zeos were used by sleep researchers.

Gwern: Maybe the headband was too uncomfortable. Maybe the hobbyists and researchers weren’t a big enough market to justify further investment. Maybe the replacement headbands cost too much and scared people off. Maybe the competitors posted fake Amazon reviews. Presumably Zeo Inc employees know, but they’re not talking.

Michael Kremliovsky: This is an important lesson for other solutions looking for a problem in healthcare domain. This is also a lesson for startups that neglect human factors.

D Zar: I think the fundamental issue is that they were trying to be in the diagnosis/healing space yet were not FDA certified to do either. “Traditional” medical device companies constantly hear the “show me the study” mantra that has been vacant in so much consumer health start-up/app work. I’m hopeful things are changing. I’m all for innovative products in this space, but they must be proven to be useful. Anything else is just a fad or, worse, dangerous.

Check out the full discussion and weigh-in with your own comments here.

Nike picks 10 startups for TechStars Nike+ Accelerator

By: Brian Dolan | Mar 18, 2013        

Tags: | | | | | | | | | | | | | | |  |

Nike+ AcceleratorThis week Nike announced the 10 startups that would enter the first Nike+ Accelerator, which the digital sport company is offering in partnership with accelerator network TechStars.

As MobiHealthNews reported in December, Nike sought out startups that can leverage Nike+ technology and data to help people be more active. While Nike+ specifically calls its users “athletes”, the company also optimistically declares that “if you have a body, you’re an athlete”. The program will take place in downtown Portland, Oregon beginning this week until its Demo Day on June 13, 2013. Nike world headquarters is 10 miles away in Beaverton. Applications were due by February 3rd and selected companies were notified shortly after — February 18th, 2013.

In a written statement Stefan Olander, Nike VP of Digital Sport, noted that the company “recently celebrated the first year of NikeFuel and the Accelerator program is a natural next step to broaden and enhance the Nike+ ecosystem – allowing Nike to offer richer experiences to athletes of all levels.”

Nike said that hundreds of companies applied to participate in the accelerator and already thousands have requested access to the Nike+ API. The 10 companies entering Nike’s accelerator program are: Keep reading>>

USC’s Saxon to demonstrate body sensors in CNN feature

By: Neil Versel | Mar 18, 2013        

Tags: | | | | | | |  |

Digital health innovators continue to make the national, mainstream media.

Dr. Leslie Saxon, director of the Center for Body Computing at the University of Southern California, will be profiled by Dr. Sanjay Gupta on Sunday’s episode of CNN’s “The Next List,” which airs at 2:30 p.m. EDT. The show, hosted by CNN chief medical correspondent Gupta, explores innovation in science and technology.

In a video preview CNN released Monday, Saxon, a cardiologist at USC’s Keck School of Medicine, is shown watching people working out in a gym, saying, “I’m continually fascinated by how much athletes, patients, everyone wants their own data because they want to learn and be better and have better outcomes.” At least one weightlifter appears to have a wireless sensor on the outside of his shirt, the kind of device Saxon has been using to gather data from elite athletes for several years.

YouTube Preview Image

Also making appearances in the short clip are AliveCor founder Dr. Dave Albert, creator of the (previously named) iPhone ECG, and Cleveland Indians first baseman Nick Swisher. CNN did not immediately respond to an inquiry about other devices and people mentioned in the half-hour show.

In a tweet Monday afternoon, Saxon said the CNN crew interviewed several Center for Body Computing members and followed around her team.

This feature on a national TV program follows Dr. Eric Topol, director of the Scripps Translational Research Institute in La Jolla, Calif., and another physician in the vanguard of the digital health revolution, being featured on NBC’s “Rock Center with Brian Williams.” Topol also is set to be a guest on “The Colbert Report” on Comedy Central next week.

The iPhone ECG was also shown and debated in a short segment on Fox News’ “Sunday Housecall” Jan. 6, though participant mentioned neither Albert nor the product by name.

As MobiHealthNews noted last week, Topol used the iPhone ECG to diagnose atrial fibrillation of a fellow passenger – and possibly save her life – on his flight home from delivering a keynote address at HIMSS13. It was the second time Topol pulled out that device to treat a heart patient on an airplane, the first coming on a 2011 flight from Washington to San Diego that made an emergency landing near Cincinnati so a passenger having a heart attack could get immediate treatment.

Slideshow: 6 apps just like Fujitsu’s pulse tracker

By: Jonah Comstock | Mar 18, 2013        

Tags: | | | | | | | | | | |  |

Fujitsu announced the development of a technology, to be presented soon at the 2013 General Conference of the Institute of Electronics, Information, and Communication Engineers in Gifu, Japan, that detects a person’s pulse using a smartphone or PC’s built-in camera. The announcement generated a lot of buzz in technology publications.

If this sounds familiar to MobiHealthNews readers, it’s because a host of other companies, mostly small app developers, but also Philips Healthcare, have been offering similar products for the last several years. The technology tracks natural color changes in the face, which are invisible to the naked eye, to determine heart rate. These changes occur as blood is pumped through the body.

Here are six mobile apps, in addition to Fujitsu’s concept that all take instant pulse readings just by pointing your face at the camera:

Fujitsu’s Pending Technology

Fujitsu

Fujitsu’s technology might not be as ground-breaking as the company’s press release suggests, but at least they give a more detailed explanation for how it works than many other companies do.

“One characteristic of hemoglobin in blood is that it absorbs green light,” the release says. “Based on this fact, Fujitsu Laboratories has developed a technology that detects a person’s pulse by measuring changes in the brightness of the person’s face as blood flows through it. The technology starts to work by shooting video of a subject and calculating average values for the color components (red/green/blue) in a certain area of the face for each frame. Next it removes irrelevant signal data that is present in all three color components and extracts the brightness waveform from the green component. The pulse rate is then computed based on the peaks in that brightness waveform. This technique can measure pulse in as little as five seconds.”

Fujitsu also suggested that the technology could have applications besides allowing people to check their pulse easily on their phones — by building it into security cameras, it could detect people who are unwell or acting suspiciously. Keep reading>>