If the health geek tribe had its own “Most Viewed” ranking on the New York Times site today, “Big Data is Great. But So Is Intuition” by Steve Lohr would be near the top. Everyone I respect (who’s awake, reading, and sharing) is tweeting about it.
Lohr writes:
Big Data proponents point to the Internet for examples of triumphant data businesses, notably Google. But many of the Big Data techniques of math modeling, predictive algorithms and artificial intelligence software were first widely applied on Wall Street… The problem is that a math model, like a metaphor, is a simplification.
Read the whole column — the online version contains links to background material on all the people and articles mentioned. But also consider the specific implications for health and health care. And be assured that others have been thinking critically about Big Data, too.
For example, my post last June about IBM Watson’s foray into medicine generated a spirited debate in the comments about what source material Watson would be fed and what other models might emerge to take advantage of health care data. Strata Rx 2012 featured multiple speakers on data & analytics, as did the Wired Health Conference: Living By Numbers.
Let’s keep the conversation going — and not just by RT’ing Lohr’s column. What other examples of Big Data hype do you see? What examples of Big Data’s promise do you see, as well? Where can people learn more about Big Data’s role in health and health care?
> Everyone I respect (who’s awake, reading, and sharing) is tweeting about it.
Oh, OUCH. 🙂 Silly old me, NOT reading, just shoveling and updating my website and ruminating about the future. Haven’t even opened today’s Globe yet, much less the “foreign” papers.
Well, maybe it helps that this post illustrates a thought as I was shoveling and pondering my inbox flood: “If I had to cut back to only following one blog, it would be SFox, because she’s most likely to catch things I’d rue missing.” Q.E.D.
===
To the issue:
When Web 2.0 was new, awareness spread like crazy, and the circles repeating it degenerated into less-and-less-informed groups. You know how things take on credibility, and thus echo-worthiness, if we’ve heard them more, even if we *haven’t* heard the definitions. It was the Gartner Hype Cycle, on the hoof. So when I started a job in 2006 everyone was buzzing about needing to get into Web 2.0, but nobody was sure what they meant. We figured it out.
Same thing’s starting to happen with Big Data. There’s a real “there” there, but it’s not easy yet to articulate what the real value is, and what’s just airheaded echoing.
I can’t answer your question about hype but I’ll say that my instincts (intuition?) suspect that the real power isn’t in what big data enables in today’s paradigms – it’s in what will become possible when we no longer need today’s power structures to let us know that something significant is happening.
A parallel from the Times in February: How Companies Learn Your Secrets described how, for instance, Target stores can predict that a woman’s newly pregnant before she’s told her family, just by analyzing shifts in what she’s buying. And that’s just what she’s bought at Target, early in the pregnancy.
Imagine what might start to be possible in medical self-awareness if some Big Data genius divines patterns that let you know “Yo, there’s a 79% chance you’ll get the sniffles next week.”
Ha! Good thing I put in that caveat about people I respect who are awake, reading & sharing. Many more are even more pleasantly engaged than you are, I’m sure, with sand shovels not snow shovels.
I just saw a tweet go by that I want to capture here, in case anyone wants to discuss it:
Hilary Mason (@hmason) is the chief scientist at bit.ly and wrote:
I’m troubled by the increasing interpretation of “big data” to mean “data without the scientific method”. When did that happen?
https://twitter.com/hmason/status/285163907360899072
And I think that was before she saw Lohr’s column.
There is another column to be added in favor of Big Data skepticism… or rather over-simplification in Big Data promises: “Is There Big Money in Big Data?” http://www.technologyreview.com/news/427786/is-there-big-money-in-big-data/
What strikes me most in health currently is a cognitive dissonance between embodiment and virtualization. That’s to say between a story telling of the “you deserve a highly personalized medicine” kind opposed to an Evidence Based Medicine behavior where MDs’ quality is based on their ability to comply to rules such as “X% or your female patients over 50 must have been screened against breast cancer”.
We probably are living a pivotal moment, when Big Data (and e-health, etc) can either get applied in the “usual way” (to say it roughly – the University Hospitals way) or in a truly innovative direction.
The usual way means extending Evidence Based medical domain, gain a better control of diseases in men, keep on digging deeper and deeper in the genome… The genuine Person is never concerned, being simply a patient in a cohort or a set of genes to be compared with other sets of genes.
Innovation may come from discovering a clear semantic for “Health”. Health is too often a hype term for “Medical” and I guess that by “Big Data in Health”, many will talk about mining into hospital information or patients web sites. But Health is about education, employment, living conditions, etc as well as “fighting against diseases”. A smart use of Big Data can hardly precede the elaboration of a proper model of what health really is about.
I’ve been thinking about this comment for a full 24 hours — even went on a run and devoted a good couple of miles to meditating on it. So many good points. One that I can’t get around is the question, “Is there money…?” Of course there is. Financial institutions have been mining consumer data for years, as have retailers (like the hapless Target data scientist who revealed too much to Charles Duhigg). Is there money in Big Data for health care? Again, of course there is. But then look at Kathy Kastner’s comment and think about the principle of “nothing about us without us.” Does health care have a different set of responsibilities and restrictions (read: regulations)? Will the Big Data entrepreneurs turn away? Should they? Do we want them to? I don’t know the answers, but I’m interested in the questions being raised here. Thanks so much!
Susannah,
First let me wish you a Happy New Year and hope you keep on raising insightful issues – like these skinny jeans of yours 😉
Back to this topic. I am not so much concerned about Money or not Money (I mean there is Big Money in health and the question is rather to know if it will keep on being massively wasted).
The true question is to know if Big Data is the proper way to get lost in complexity before discovering back old evidences or, on the contrary, to open the door (at least contribute to opening the door) to “the world to come”.
Two directions here :
– Embodiment, that I may describe as the result of a Copernican Revolution where Mrs Smith truly is the center her own world.
– Virtualization, where Mrs Smith is treated as a another patient in a cohort.
Big Data might contribute in both domains (for example fed with Quantified Self information on one side or with multi-patients genomic information on the other side).
To tell the truth, I am much more involved in contributing to making the Copernican Revolution happen before I can guess what Big Data can become useful for.
Another noteworthy tweet, from @EdwardTufte himself:
“Steve Lohr escapes the big data hype, notes instances of failed amateur social science. http://nyti.ms/YYhHRL #bigdata”
https://twitter.com/EdwardTufte/status/285250987596333057
Re @HMason’s concern about not knowing what big data really is: @JaniceMcCallum led a really good #S4PM #bigdata tweetchat a couple of months ago – she hasn’t taken my tweetbait today to come here – she knows her stuff, and she’s no fool re hype. Hope she’ll join in.
The transcript of @JaniceMcCallum’s 8/29/12 tweetchat on #bigdata in health is here.
On the definition issue:
@JaniceMcCallum … Why don’t we start w/ your perceptions of meaning of #bigdata?Typical def’n includes volume variety & velocity.
The other V that’s crept in is variability – a Google search for big data volume variety variability velocity gives lots of hits, including this from Edd Dumbbill of O’Reilly.
@JaniceMcCallum: Any thoughts on level of hype about #bigdata? @VinceKuraitis has post that says we need to focus on “small data” first. #s4pm
It was a busy tweetchat – 159 tweets in 59 minutes – worth a scan.
It may be too much to hope for on a Sunday, but I would love for more people to join us. I completely get that it’s easier to tweet a reaction or a link and I’m grateful for those, too.
For example, Keith Eric Grant (@ramblemuse) sent me this Nature article about the work that David Agus and Danny Hillis are pursuing:
Megadata: The odd couple
http://www.nature.com/nature/journal/v491/n7425_supp/full/491S52a.html
If you’re in a listening mood rather than a reading mood, check out:
David Agus at TEDMED 2010
http://www.tedmed.com/talks/show?id=7055
Danny Hillis at TEDMED 2010
http://www.tedmed.com/talks/show?id=7029
Researchers–academic and otherwise–would love to get their hands on Big Data-sets and have already been thinking through critically about actual *questions* to ask of them. It’s a matter of making sure we teach others how to manage, use, and think through the implications of having such data and to think a a research-minded way. We won’t necessarily receive answers to everything we need to know through Big Data. We’ll still need theories and questions to ask that inform critical use of Big Data.
Also, access to Big Data in retail and Wall Street has never been more abundant, but healthcare is a different beast due to regulations. If only we could aggregate patient data from medical records without legal barriers and really make sense of what health issues people are experiencing and treating, whether or not they are e-patients or not.
I decided to follow up with a post on my own blog. Thanks for the original post, Susannah!
A high compliment! Thanks so much.
Everyone, here’s her post:
http://rupapatel.com/2012/12/30/the-critical-intersection-of-big-data-research-and-healthcare/
You might also be interested in this talk by Hilary Mason (yes, I apparently have a brain-crush since this is the second time I’ve cited her in this thread):
Devs Love Bacon: Everything you need to know about Machine Learning in 30 minutes or less
http://www.hilarymason.com/presentations-2/devs-love-bacon-everything-you-need-to-know-about-machine-learning-in-30-minutes-or-less/
She stole my heart at minute 7 when she put up a Venn diagram entitled “Data Scientists” showing an overlap between engineering, math, comp sci, and curiosity and then suggested that knowing what questions to ask can often trump skills in other areas. It’s well worth 32 minutes of your time if you, like me, are going to school on machine learning, Big Data, etc.
Great debate. Thanks, Susannah et al!
Different kinds of data have different value. We are often asked over at clearhealthcosts if we are excited about big data sets about payment records, the assumption being that Big Data will reveal secrets about why health care costs so much. But that Big Data is expensive and hard to get at and interpret.
Most people don’t know that costs/prices/reimbursements vary by a lot, because it’s fairly recent that individuals have begun to be exposed in big numbers to the consequences of price variations and market opacity.
So we have chosen to do our pricing surveys and reveal the results in a “small data” way — sort of like the Slow Food movement. It’s democratization of information: we find out startling stuff, and reveal it to everybody who’s got access to the web.
For example, here in the New York area, we’ve found provider charges for a garden-variety MRI from $350 to $3,500. We’ve found payer MRI reimbursements from $400 to $2,300. Small data that’s relevant to real people, we think. And if your provider is charging $2,300 for your MRI and your insurance company is paying $2,300, then seeing that price info in the context of other prices brings you a real shock. Or if you’re uninsured or otherwise responsible for all or part of the cost, then you now have actionable information.
So in a way we are saying what Steve Lohr, a brilliant former colleague, and the equally brilliant Hillary Mason are saying.
Big data alone? Well, O.K.
What’s better: Context, analysis, intuition, thoughtfulness – and visibility.
<> Thank you Jean, for your put-perfecty perspective and for your Small Data experience.
Having lurked on @janetmccallum-hosted #S4PM I’m still befuddled by perceptions/descriptions/mechanisms of Big Data.
What stood out for me was the tweet that echoed my growing uneasiness with what BigData seemed to be missing BIg TIme:
In the last 10, Janet tweeted: ‘thoughts on incorporating patient-generated data?’
Lots of agreement and enthusiasm.
It was Scott Strangely @Strangely_T1 – who pointed out what was missing, and making me uneasy: In the spirit of “nothing about us without us:”
Scott: ‘Speaking as a diabetic, the patient-generated data may be the most important data set. I bring onehelluvalot of info into an appointment ”
Thanks Susannah for encouraging the convo, and motivating me to get my thoughts in order. Kathy
Thanks to you and Jeanne for answering my call to the table – tweets disappear like Champagne bubbles, but comments persist (like the Slow Food cited above). Yummy ideas, even uncomfortable questions: keep them coming. I suspect we’ll be talking about this well into the new year.
Happy New Year and here’s to a happy, healthy 2013 for all.
And to you!
danah boyd & co. have a very good article that I like to use that is a good piece to ground big data critiques beyond just throwing the word hype around. I can dig up the link if anyone is interested and post when I’m back on a laptop.
Jody, is it “Six Provocations for Big Data,” for the Oxford Internet Institute, Sept. 2011?
The provocations:
1. Automating Research Changes the Definition of Knowledge
2. Claims to Objectivity and Accuracy are Misleading
3. Bigger Data are Not Always Better Data
4. Not All Data Are Equivalent
5. Just Because it is Accessible Doesn’t Make it Ethical
6. Limited Access to Big Data Creates New Digital Divides
Fascinating list! And I love these opening quotes:
“Technology is neither good nor bad; nor is it neutral…technology’s interaction with the social ecology is such that technical developments frequently have environmental, social, and human consequences that go far beyond the immediate purposes of the technical devices and practices themselves.”
Melvin Kranzberg (1986, p. 545)
“We need to open a discourse – where there is no effective discourse now – about the varying temporalities, spatialities and materialities that we might represent in our
databases, with a view to designing for maximum flexibility and allowing as possible for an emergent polyphony and polychrony. Raw data is both an oxymoron and a bad idea; to the contrary, data should be cooked with care.”
Geoffrey Bowker (2005, p. 183-184)
And this quote by the authors, from page 1: “There is little doubt that the quantities of data now available are indeed large, but that’s not the most relevant characteristic of this new data ecosystem. Big Data is notable not because of its size, but because of its relationality to other data.”
Anyone who by now is not intrigued to read the full 13 pages (+refs) will never be.:) Thanks!
Oops, the URL is http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1926431
Dave & Susannah,
I’m not sure how I missed latching on to the bait that Dave apparently dangled in front of me to join the conversation here. No matter, I finally got the hint.
I’m preparing a talk on big data to present to the scholarly publishing community at the AAP/PSP conference in early February in DC in which I’ll use examples from healthcare/medical publishing. Until I have the preso to share, let me just say that using my interpretation of the term “big data” (analytic methods that can operate on data inputs that have high volume, variety, and velocity–and throw in the 4th “V” for veracity), this new buzzword can be applied to the future of evidence-based medicine that I’ve been talking about for many years: a learning model that incorporates structured and unstructured data from traditional clinical trials, outcomes data gleaned from registry data aggregated from EHRs, patient reported data, genomic data, etc. In short, we can extend the very-limited randomized clinical trial data that have been the “gold standard” for EBM to include much richer models of evidence that can be personalized by a range of variables, including such obvious ones as male/female.
I don’t view big data methods as being unscientific. Rather, I view big data methods as allowing richer analysis of more complex models. And I think we desperately need more complex models in medical research.
To wrap up, I highly recommend the talk on machine learning by Hilary Mason that I RT’d this weekend: http://www.hilarymason.com/presentations-2/devs-love-bacon-everything-you-need-to-know-about-machine-learning-in-30-minutes-or-less/.
I look forward to keeping the conversation going.
Happy New Year! I dip my toe gingerly into this conversation about Big Data and appreciate Liza Bernstein bringing my attention to the original article and Susannah’s post. I also appreciate Susannah Fox suggesting that Liza and I bring our twitter dialogue into this comment stream. I only drop in gingerly because we are still on holiday (kids don’t start back to school until Tues, Jan 8), and I worry that I won’t be fully able to engage in the back and forth I hope our comments spark.
The quiet reality of all of this brouhaha about Big Data in the last several months is that the answer to these issues has nothing to do with data. It also has nothing to do with privacy; nothing to do with identity; and nothing to do with security. It actually has to do with all of them, and until we move beyond each of these technology silos and see each of these topics, not as separate, but as axes emanating from a common problem, we will not unlock the power of data or more importantly, the power of the people that are connected to it. By addressing this as a linked people problem, not just a linked data one, then we will be able to unlock the incredible talent arrayed in the healthcare industry, the passion and power of engaged patients; and the analytics and analysis of data scientists.
The other challenge with the various conversations about big data, patient-centered data and many other related topics is what I describe as ‘chartopomorphism’ or paper-centric thinking. Just like we moved from geocentric to heliocentric thinking and now to the universe and even multiverses, our discussion of the Internet needs to leave the boundaries of paper behind, too. Discussions of “big data” whether in health, finance or any other industry are always framed as if we were moving around digital manila folders (EHR, financial statements or education records) from one digital filing cabinet (database) to another.
Just like we don’t know which power plant on the electric grid delivers the specific electrons to our homes or businesses to power our lights or appliances, we need to evolve our thinking around data in the same way. Relational database technology focuses on the power plants; contextual databases™ (CLOUD term) focus on the grid. We need to wrap our databases around the data, rather than the data around the databases. When we do, the data will never be big or small but will always be in context, weaving together information from across the Internet, as appropriate (atomistically and privately) rather than just the data we’ve moved from one place to another.
CLOUD [Consortium for Local Ownership and Use of Data, filed for 501(c)(6) status] was born in response to my wife’s journey through cancer treatments and has resulted in our effort to build the multi-dimensional language that will allow us to “Reweave the Fabric of the Internet.” This new language will not only unleash contextual databases but address the vexing issues of privacy, security and identity, too. If we are successful, then the unbelievably talented people making comments in this thread and many others throughout healthcare and other industries will be able to build on this new language and unleash a powerful new renaissance. At the end of the day, the power of innovation is all about doing new things, not just old things in new ways.
There is too much to the CLOUD way of thinking to pack into one post, so if you are interested in digging deeper, I’d suggest a visit to our home page at http://www.cloudinc.org. Start with my talk at TEDxAustin (link just below my shiny bald head) then click through some of the overview videos below the story of CLOUD to get a sense of the general architecture. There are also a number of posts on CLOUD’s way of thinking in the realm of healthcare (and other industries) under the ecosystems tab at the top of the page.
Thanks for the invitation to this part of the tribe and the opportunity to share a few thoughts as we start 2013!
Great discussion, thanks Susannah! Two short additions:
1. I recently heard from one of my colleagues in Chicago while discussing if there is a Big Data bubble: “anyone that talks about Big Data, is part of the bubble!”
Yes, there is a lot of richness to be extracted from large sets of data for drug development, treatment protocols, marketing, etc… but this inherent richness is not embedded in the size of the data. The richness comes from smart analytics and connections that can be made with information. “Big Analytics” would be more meaningful than Big Data, but is not as sexy, especially when thinking of the latest marketing tag-line.
2. Healthcare is not ONE thing, there is no Healthcare System, it s NOT a system; it really is more like an emergent network in the traditional sense. (this is why there are no systemic healthcare solutions that work, because you can’t solve for a systems that is not a system). So, if you agree with me that healthcare is an emergent network made of up different industries and players (i.e., biopharma, medtech, payors/insurers, providers, caregivers, patients); then it makes no sense to talk about how Big Data can help Healthcare. Analyzing large data sets for drug development is completely different from leveraging large data sets for understanding cost models, or human behavior. So, if we all want to have more meaningful conversations around large data sets within the complex emergent network that healthcare is, we need to distinguish if we are taking about drug/treatment development, patient engagement, adherence, costing models, etc… (by the way, for the same reasons discussed above, there is no such thing as “Healthcare Innovation”).
3. Given my interactions with my physicians and nurses in the last 3 years, all by the way at Harvard Medical School affiliated hospitals, the last thing I want is for my PCP to have more data. I rather have a smarter use of small data than access to big data. Short example, http://designandbiology.blogspot.com/2011/06/insanity-of-our-healthcare-system.html
Finally, while large data sets can be very useful for medical reasons if analyzed properly, these are not very useful to design good experiences. And we all would like to have better experiences when dealing with any part of healthcare. And here is where good service design kicks in, but that is another topic.
Intuition is under attack in medicine, and it’s not by Big Data but rather by “Small Data” in the form of Evidence-Based Medicine (EBM). Rather than individualized recommendations, we live by blanket guidelines derived from studies performed on small datasets. The sample size of a “huge” medical study would not even register as a blip on the screen of a true Big Data set. The problem with Big Data in Medicine is not its threat to clinical intuition but rather our failure to do it at all.
Why aren’t we heavily using Big Data in Medicine? Big Data requires really big datasets, and we don’t have really big clinical datasets. Medical informaticists still insist upon enforcing the antiquated data curation processes required to support “Small Data” analytics. The resulting dataset is small because the human resource intensive curation process is a bottleneck, and much of the incoming data is rejected for lack of conformity. Google and Amazon do not succeed at Big Data despite a lack of curation, they succeed because of it. Until every doctor has perfect intuition all the time, computers should help us provide better care. Big Data can play a big role in supporting care, provided that we start creating truly Big Data and stop trying to apply Small Data thinking to the process.
I expanded more on this idea on my most recent post at Wired here: http://bit.ly/ZQsLMw. Check it out!
Thank you Susannah. There is so much to say on this topic, so I’ll just add two thoughts to the discussion.
First, “Big Data” is obviously a suitcase term into which people put anything related to data aggregation, web data, data mining, analysis, prediction, etc. to the point where it has become somewhat meaningless – Watson, Google flu map, Hadoop infrastructure, Facebook social graph, etc are highly diverse applications and tools, yet painted with the same brush. As Dave says, something in those various categories is going to be very important to health, but there will be many people trying to ride the hype curve in the meantime.
Probably the most boring, but likely application of Big Data tools to health will be augmenting traditional observational research and comparative effectiveness research. There is healthy discussion taking place in the medical journals on the role of EMR data at scale in evidence based medicine. For example, how can we aggregate individual cases to yield better evidence or hypotheses for RCTs? Can we use this data for early detection of mortality or side effects (e.g. Vioxx)?
But this very discussion reflects one of the central problems in healthcare: the industry is driven by thinking about biology when it should probably be conceptualized more in terms of sociology or psychology. A well-run RCT may identify a highly effective drug, but an overworked doctor and under-informed patient have to both know about and understand how to use the results in a decision process that leads to . Only 25% of our best biological insight is successfully delivered through outcomes, the missing 75% of ideal health delivery is about the other stuff.
The spread of information, support for decision making, behavior change interventions, and systems improvement processes that can support better health are independent of the biology yet critical for outcomes and exactly the kind of problem that the hodgepodge of technologies and processes labeled “Big Data” are focused on in other fields.
The most exciting implication for me is that the technology trends we’re seeing are lowering the barriers to incorporate data sources that are traditionally ignored by healthcare (our online activity) or didn’t previously exist (consumer sensors like Fitbit, Zeo) that capture a broader set of behavioral and psychological indicators. These sources can significantly enrich the doctor-patient dyad and we’re seeing great examples of this at the C3N (http://c3nproject.org). At the health systems level, this data in aggregate can enhance our ability to engage in the quality improvement model that we are so impressed with at the ImproveCareNow network (http://improvecarenow.org). My greatest hope (beware: Twitter-style sound bite coming) is that Big Data is the gateway drug to Learning Networks. Google for “Learning Network Institute of Medicine” to learn more.
On the other hand, my fear about the Big Data movement intersecting with the healthcare’s culture of biology is that it commercial efforts will be pigeonholed into doing bad observational research on disease-focused problems, and fail to see the true potential: a better understanding of the patient who has the disease rather than the disease the patient has (thanks to Hippocrates).
What a great discussion and solid resources for learning on Big Data! I had bookmarked Lohr’s article to read quietly after the New Year celebrations (in Greece they are very important family holidays), something I did yesterday. Today afternoon though I glimpsed Susannah’s tweet:
“I blog about stuff I want to understand better. Learn along with me about Big Data and health care: http://bit.ly/TUbdSS”
on my TL and saw she was commenting on the same article..I enjoyed reading it as well as the comments, but I wonder how all this can apply to many countries in Europe, Greece among them…
How can we talk about Big Data in sectors or countries, where there are no digital data? Reading Lohr’s article, Susannah’s post and the comments I remembered Atul Gawande’s article “The Hot Spotters” because our situation in Greece resembles more what Jeff Brenner and Rushika Fernandopulle (note: Dave’s and mine classmate from Salzbiurg Global Seminar) had met.
In Greece, we try to put a brake to unreasonable health care costs without having the possibility to first study big series of data to pinpoint where the costs stem from…Simply, because no digital data are kept! We are still in the era of handwritten patient records and handwritten prescriptions, although now, after the experimental introduction of e-prescribing in 2011, they represent something around 70% of total prescribing and referrals.
One culprit is supposed to be medicines expenditure which was so drastically curbed that medicines disappeared from the pharmacy shelves! It was drastically cut down not because prescribing was rationalized, based on Big Data or some other data, but only on the last five years annual medicines expenditure that showed an escalation. It seems that the authorities did not bother to know which medicine categories most contributed to the rising costs, nor why (more patients, more acute conditions, directed prescribing??)…
So, how can then BigData that are still hidden in handwritten registers, documents, prescriptions, social insurance refunds, hospital supplies contracts, etc. be used to draw a meaningful picture of what needs to be reformed? How patients and populations that do not even know what is Big Data and how it can be used in healthcare be involved?
I agree with Jonathan remark that EBM tends to oust intuition. In Greece and in the other European countries under economic crisis, health care economists dictate that physicians should only practice according to EBM.
Gary Thompson remark: “By addressing this as a linked people problem, not just a linked data one, then we will be able to unlock the incredible talent arrayed in the healthcare industry, the passion and power of engaged patients; and the analytics and analysis of data scientists” says it all…
I just posted on twitter & FB about your post on #doctors20 and #opnhealth and we already have a contribution from good friend Francoise Soros/France, a physiotherapist and healthcare advocate
http://lecercle.lesechos.fr/economie-societe/social/sante/221146768/big-data-et-sante-13-opportunites
I have informed readers/friends on FB to post any resources directly in the comments of your blog post
The question about the role of intuition in the context of Evidence Based Medicine inspired a somewhat philosophical comment about the role of formal methods in the practice of medical problem solving. It went long enough and strayed sufficently off topic that I’ve moved it to my own blog, but readers of these comments may find it sufficiently relevant so I’ve included a link here. http://ianeslick.com/big-data-healthcare-and-the-human-lens
Why the skepticism about Big Data? We’re already seeing results when a volume of data from various sources that’s always changing is thoughtfully analyzed. If the data has been gathered in a standard way and verified, what’s the risk? Analysis of big data diminishes the value of intuition only when the intuition leads to poor results.
I’m a big supporter of better data>better analysis>better decisions in healthcare. Here are some 2012 examples I learned about in the Chicago market. http://bit.ly/YUjRxj
Thank you, Susannah, for spurring a lively discussion. With each new technology comes the irrational exuberance that this new toy will be the answer. Current examples include games in health, and now, big data. Yet, those of us who have treated patients understand that gizmos and gadgets are simply new tools in an ever-expanding toolkit.
In 2011, I researched mobile gaming and online support across the healthcare continuum, from wellness to chronic disease. Although the hype gathered around gamification, online support networks proved to be the more powerful motivating component. (The importance of peer-to-peer healthcare was first brought to my thinking via your work – http://pewinternet.org/Reports/2011/P2PHealthcare.aspx )
With the skeptical eye of a provider and financial analyst who still believes in people over machines, last summer I set off on a similar project about Big Data, wondering if the buzz was hype or hope or something in-between. Surprisingly, after interviewing more than 30 companies that use Big Data, or other advanced analytics in healthcare, we found an emerging ecosystem of companies interested in using Big Data to improve healthcare in six ways:
1. Support Research: Genomics and Beyond
2. Transform Data to Information
3. Support Self-Care
4. Support Care Providers
5. Increase Awareness
6. Pool Data to Expand the Ecosystem
Interestingly, among this sample of companies, three data usage trends emerged:
1. Working with limited data sets
2. Combining a greater variety of data
3. Pooling data for better results
Although not a panacea for the problems of healthcare, big data and analytics seem already to be useful new tools, obviously in genomics and proteomics, where data is of the essence. Like mobile health games, social media (which is big data by definition), and providing support for self-care seem to be blossoming applications. Details of this research can be found in Big Data in Healthcare Hype & Hope http://bit.ly/RbX0KR