Skip to main content

Clinical Trials

Verily Says “We Grew Up”. We’ll See.

Some not-very-flattering stories have run about how things have been going at Google’s Verily life-sciences startup. For some time, they’ve been making big pronouncements about all their transformative technologies (tricorders! glucose-sensing contact lenses!), but actually turning any of these things into working products has been a bit more difficult than giving interviews about them and issuing press releases.

Here’s an update from Bloomberg Businessweek: the company has started a very large scale project to collect a baseline data set in 10,000 patients. They’re raking in as much biometric data as possible, hoping to assemble more medical information about a set of healthy volunteers than anyone has ever had before and use it to guide its product research and look for opportunities. I have no problem with that – as long as their signal/noise is solid enough, they have a chance of finding some interesting and useful things. But this doesn’t sound like the company that everyone was writing about:

While that sounds ambitious, it’s much more modest than the missions Verily promoted when it was officially part of Google. Years ago, the biotech division promised projects such as glucose-monitoring contact lenses and all-in-one medical scanners; those remain in the lab. Former employees say the internal code name for the life sciences division was Panacea—cure-all. That’s over.

“We grew up,” says Verily Chief Executive Officer Andy Conrad. The middle-aged geneticist has adopted the Silicon Valley T-shirt-and-flip-flops wardrobe of eternal youth, but he’s given up on a lot of the jargon, including Google’s onetime favorite word. Like some other Alphabet holdings, Verily has stopped talking about everything in terms of industry-changing “moonshots.” What next-generation technology requires in practical terms is “setting the goal and then getting down to the day-to-day practical drudgery,” Conrad says. “If you examine the real moonshot closely, you’ll see a dude whose job is to rivet and a lady whose job is to do some wiring.”

About the whole “tricorder” nanoparticle diagnostic thing (which first showed up in 2014), Conrad says that “Mother Nature defeated us wildly“. Now that’s the biopharma world that I know and that I’ve been working in, and I’m experiencing no small amount of schadenfreude to see someone else getting the treatment, too. This is the same Andy Conrad who told people (see that link in the first paragraph) that “We’ve done a lot, to be quite humble about it. Enough to give us great confidence that this is all likely to work” and that there was no need to test this kind of thing in animals, since it was so solid. In fact, the people in their Baseline Study (if you believe some of Conrad’s pronouncements) were all supposed to be walking around wearing these things.

What they’re actually wearing is the Verily Study Watch, which from what I can see is mostly tracking movement and cardiovascular data. That is, it’s a larger, more powerful (and presumably more reliable) tracker of the sort that all sorts of other people are selling, both as standalones and as apps for smartphones. It’s not that useful data can’t be obtained from such a device, but it ain’t no tricorder, neither. The tricky part is that you’re going to have to go out and get that big heap of data before you can say if there’s anything actionable in it. Don’t get me wrong – I’m glad that Verily is doing this study, because not many people would, and there could be interesting results. But interesting results are not guaranteed. They might get a big ol’ baseline of things that don’t surprise anyone at all; that’s the risk.

It’s going to be interesting to see if Conrad himself (and those who think along the same lines) has actually been humbled by his Verily experience, or if he’s saving up for another round of big declarations once things start to work a little. This behavior is one of the things that irritates non-Silicon-Valley types in health care the most, the “Step aside and let the disruptive world-changers do their thing, Pops” stuff. It’s a long way from Conrad’s earlier vision-tastic interviews to tracking heart rates and sleep schedules for the next four years. But that Businessweek article sums it up with a quote from former FDA Commissioner Robert Califf, who’s saying the same thing I’ve said around here for years now: “At times, Silicon Valley people are very naive about the complexity of health care. . .It’s going to be a lot harder than they think.”

22 comments on “Verily Says “We Grew Up”. We’ll See.”

  1. Mark Thorson says:

    Current watches and fitness bands barely get any useful data because they only can monitor acceleration, pulse, a few other things of little value. From these, software makes assumptions and derives metabolic rate, sleep schedule, etc. with data quality limited by the sensors. You could be jogging along, a dog barks at you, and your pulse goes up — but nothing else has changed and the software makes wrong conclusions. What is needed is a new generation of good sensors — not superexotic technology, but rational incremental advances in known technology. I’m working on sensors for blood flow and endothelial function. There already are sensors for both which do a good job, but making them suitable for a watch or fitness band is a challenge. Main problems are a) sources of artifact (ambient light or EMF, acoustic, motion, etc.), b) power consumption, and c) user comfort. Completely unacceptable are any sensor technologies which have needles that penetrate the skin, which immediately rules out many existing technologies. My goal is fine-grain recording of endothelial function, so if I smoke a cigarette within a few minutes my watch is alerting me my endothelial function has crashed. Same thing if I drink a sugary soda. That will give value to the consumer by presenting useful information that may modify behavior. The same goal might be achieved in a more crude form with fine-grain recording of blood flow by itself. My fingertip pulse oximeter gives remarkably good data. I can see my own dicrotic notch, which impresses me. Measuring the dicrotic notch might be a workable substitute for endothelial function, but that remains to be shown. Endothelial function has a much larger body of studies which establish its value as a general indicator of health which reacts strongly to both immediate challenges (smoking, sugar) and long-term challenges (exercise, obesity).

  2. tlp says:

    Yeah, one could say from their presentation at the ACS Meeting that they’ve got nothing close to initial claims

  3. Grammar sarge says:

    Last paragraph, first parenthetical – “and those who think…” rather than “and those who thing…”.

  4. Mad Chemist says:

    The physical sciences are a lot harder than computing, my friends.

    1. cancer_man says:

      Which is why we have to quickly move our brains to The Cloud and away from this silly carbon substrate!

      Something like that.

    2. dearieme says:

      That’s about the size of it.

  5. Moses says:

    So it’s something like the UK BioBank project then? Except with fewer people and a broader age range?
    Regeneron has an agreement to access and use some of the data that BioBank has already captured over the last few years.

  6. Anon says:

    Seems like Facebook needs a reality-check slap now also:

  7. Pennpenn says:

    Not that there’s as much of a correllation as people would like between saying you grew up and actually maturing and acting in a more responsible fashion. Oh well, here’s hoping they get some good science out of this.

  8. Humulonimbus says:

    Naivete got them into this, but I hope they are determined to take the lessons and build them into something that’s actually useful. I don’t think anyone can argue the status quo is cutting it. Breakthrough thinking can be indistinguishable from crackpot thinking until the breakthrough actually works.

    I don’t begrudge you the schaudenfreude, though. Obviously, not as easy as they claimed it would be.

    1. Mark Thorson says:

      This outfit raised $120M for their networked, apped $400 juicing machine that turns out to be hardly better than squeezing its fruit cartridges by hand. Google parent Alphabet was an investor.

      Good work, if you can get it.

      1. Blabla says:

        Please don’t link to the Daily Fail, it’s an odious publication full of lies and xenophobia. Don’t want to give them any more hits than they already get.

        1. Isidore says:

          A perfect example of scientific open-mindedness: Ignoring the message because of the messenger.

          1. Pennpenn says:

            Is it particularly scientific to persistently indulge provably unreliable and often horribly biased sources?

          2. Blabla says:

            As above. A good scientists judges the validity of a source. The Fail is not a source for factual information.

        2. skeptic says:

          You might be right about the Daily Mail, but it appears to be correct about the “Juicero” (I can’t believe people were able to raise venture capital for such a thing):

  9. tnr says:

    Ten thousand subjects with the only entry criteria that they wear a watch? That is laughably small in the epidemiology world. I’m involved with the TEDDY epidemiology study of over 8500 at risk subjects for Type 1 diabetes. We’ve been recording data on all subjects since birth and trying to tease out signal from noise in this highly specific cohort has been difficult.

    1. Wile E. Coyote, Genius says:

      I don’t think it’s even that they wear a watch, but that they wear Google tech. It is the marketing of the Google tech that matters, not the study.

    2. Emjeff says:

      You’re absolutely correct – 10000 subjects is a joke. But, it is probably what their budget allows, and they probably have not bothered to talk to a stats expert – why would they do that, after all – they’re programmers and know everything….

      This will be a colossal waste of money.

    3. MTK says:

      That was what I was thinking.

      10,000 subjects of varying age, weight, etc. wouldn’t seem to be powered to tell you much of anything. Well, not anything new or insightful at least.

  10. mallam says:

    Saw your answer, Derek, to the question on the March for Science at the AMA. Quite disappointing that you did not give an type of encouragement, and were so negative. The concept behind the March came from the current trend in abandoning scientific data and resulting conclusions around the environment, vaccines, drugs, chemicals, CO2 and more. If for no other reasons than this, you should have given some type of support for the March as a spokesperson for science that some people seem to respect. Any respect from me has now vanished.

  11. Alex Baxter says:

    UK Biobank has 100,000 participants (out of 250,000 total) with physical activity data (in addition to the standard UKB baseline questionnaire, blood sample, and soon DNA, plus they have resample questionnaires and imaging substudies that could be linked to the physical activty substudy).

    “World’s largest objective physical activity dataset now available”

    – note UKB has the raw accelerometer data so researchers will be able to do some interesting data processing on it. And of course all of this data is available to registered researchers.

Comments are closed.