Google and Microsoft Look to Change Health Care is an interesting piece by Steve Lohr in today’s New York Times. (In the print edition the headline reads “Dr. Google and Dr. Microsoft”.) It begins,
In politics, every serious candidate for the White House has a health care plan. So too in business, where the two leading candidates for Web supremacy, Google and Microsoft, are working up their plans to improve the nation’s health care.
By combining better Internet search tools, the vast resources of the Web and online personal health records, both companies are betting they can enable people to make smarter choices about their health habits and medical care.
“What’s behind this is the mass consumerization of health information,” said Dr. David J. Brailer, the former health information technology coordinator in the Bush administration, who now heads a firm that invests in health ventures.
Naturally the piece frames health care as a fight between giants. Even the Larger Context is cast in terms of Big Interests:
It is too soon to know whether either Google or Microsoft will make real headway. Health care, experts note, is a field where policy, regulation and entrenched interests tend to slow the pace of change, and technology companies have a history of losing patience.
I suggest we need to lose patience even faster than the tech companies, and come up with solutions that are not framed in terms of big company (or big entrenched interest) sports, but rather in solving a single problem from a single angle.
That problem is patient records, and that angle is the patient.
The patient needs to be the point of integration for their health care data. That doesn’t mean that doctors and hospitals shouldn’t also have their data. It does mean that the patient should have access to all of it.
The way the system is set up now, important facts about our bodies are not ours. Nor are they easily accessible by us. Yet when we go to a clinic or an emergency room, we are handed a clipboard with a paper form that we are expected to fill out from memory, about our immediate condition and our relevant (and even irrelevant) health care background.
A couple months ago, when I developed a condition called posterior vitreous detachment in my left eyeball, I had to fill out a pile of forms at two different locations. In the course of that experience I found my name spelled three different ways, while I was also challenged to remember how to spell out Type II dyserythropoietic anemia, a rare and relatively minor blood disorder which in emergencies only comes into play when anesthesia is involved — which it might after I fill out forms like this. Who knows, right?
In any case, it would have been handy if I could have auto-filled the forms from my own database, or my own metadata: data about data that lives elsewhere.
While it might be true that the giant sticks in the mud (more like huge pilings holding up a rusty pier) aren’t going to be too cooperative, so what? We — individual patients — need to be able to use our own health care data, for our own good, and for the good of the systems that depend on it, and will be in the line of blame when things go wrong.
I’m not sure patients need to “own”, “control” or even “manage” their health care data. But they clearly need access to it, especially when emergencies come up. Where problems need to be solved, there is business to be made, and I think this is one big one.
The key, as with all VRM projects, is that the solution needs to be anchored on the customer side — in this case the patient side — of the relationship. The answer to the silo problem is not yet another silo. It’s a silo buster, or a silo integrator.
Any ideas? I can think of a few, but would rather see the rest of ya’ll go first.
This page has some info:
http://www.centerforhit.org/x201.xml
The ASTM CCR really looks like a step in the right direction: XML format for patient information. Now, if only we can get health care providers to publish that data to us…
This looks interesting:
http://www.recordsforliving.com/HealthFrame/
Will let you store info on their site and sync it with your iPod.
Just my thoughts on how to do this with the best reward/cost ratio:
If we could just get the blank forms online as PDFs, from the doctor’s web site…to be filled out OFFLINE, this would be a major step forward.
1> We could do it at our own pace, thus reducing stress, and increasing accuracy.
2> We can ask our spouses or other care givers to complete/check the information. (My wife is an ER RN, she’s got a much better command of my medical history than I do, for example.)
If we could actually fill them out online.. it gets even better.
3> The importing of this information to the Doctors system can be accomplished without having to waste a clerks time.
4> It should be possible to get an XML file with a dump of the form and the data, for auto-importing into the next run at it.
5> Uploading of a patient XML file with previous history could shorten the data entry process, and improve accuracy.
Someone could make a business out of this, assuming they have appropriate disclaimers and/or sufficiently paranoid programmers who know how to do very STRICT input validation.
What do you think?
–Mike–
Secure, Online Forms does this now. (soforms.com)
But they’re driving from the doctor angle. Docs have to buy into it and get the service, offer it to patients, etc.
What if they offered this (free) to individuals who could then drive doctors to set up accounts to get the data out? That would be a feasible model, in my opinion. That creates portability, although I’m not certain it doesn’t create another silo of data. Or multiple silos of data. Potentially in multiple formats.
I’m in the electronic medical records business, and am very happy to see this thread. It extends far beyond intake and history forms, unfortunately. I’ve been a part of an ongoing conversation about what I call the portable healthcare record. It’s patient-driven, although it’s not owned wholly by the patient.
Think of the pHCR as an account that is credited each time a transaction takes place. It’d be much like a credit card statement, in that the providers of healthcare own each node of transactional data, much like a restaurant does when you pay your bill, but that the healthcare data is reconciled for viewing/sharing for the patient.
There would need to be a standard in place for this transactional data, which would be fairly easy to define, but much harder to sell to technology companies.
I’ve designed the first generation of a pHCR schema and have implemented it into our products, so that when the time comes, we’re ready to play with any open system. We’re doing double-duty right now, in that the data is kept within a siloed database, but also syndicated into an open architecture.
We supply integrators with the schema and open up a tunnel to the data. Bam! They can play with us. At very little cost, I might add. The integration is stupid simple.
Unfortunately, most electronic medical records companies think this is an odd idea. If they adopted this schema, doctors could much more easily switch platforms. Or curcumvent their interfaces altogether via web services.
There’s also an issue with unique patient identifiers, but that’s another ball of wax…
What do you think? What’s the first step?
The reason I suggested PDF files is that the process can be done WITHOUT ANY modification to the existing software, processes, etc. We’ve got a wonderful copier here at work that takes paper in, and puts PDF files in your home folder… making PDFs from clean blank forms costs essentially NOTHING.
If the doctors office in question has a web site, the cost of adding the PDF files as a courtesy to their patiences is almost trivial. The benefit to patients would be almost immediate once word got out.
I know the inertia of the millions of bureacrats embedded can’t be overcome directly…. this is one small way AROUND THEM.
Doc,
Since we IMed on this topic, my colleague here at UNC-Chapel Hill Gary Marchionini pointed me to part of a paper he published on Personal Health Records and their management.
He writes (quoting his paper):
The marketplace often drives innovation and PHR system development has been
more active than usability research. In 2006, the Markle Foundation
identified more than 167 different PHR products.[1] We examined
those that were publicly available (More than 20 of the websites listed led
to dead links). They range from full-featured systems available from
specific health care provider or payer organizations (92 were from such
organizations) to standalone products from small companies. Ninety-two were
aimed at health care organizations and their members, 57 were aimed at
general consumers and another 16 to consumers or health care providers. Two
targeted consumers with specific diseases (cancer, diabetes). Less than
one-quarter of these PHRs link to MHRs, and three-quarters of them are
Internet based. An important contribution of this work by Markle is the
categorization of 32 PHR functions into five classes: medical management
functions such as medication tracking, medical history, etc. (11 functions),
administrative management such as emergency contacts and bill paying (7
functions), other functions such as reminders, physician messaging, and data
monitors (7 functions), security such as firewalls (4 functions), and
privacy such as passwords (3 functions). Each of these functions will carry
a usability burden and thus inform the usability guidelines framework.
The American Health Information Management Association (AHIMA) provides
webpage links http://www.myphr.com/resources/) to PHR tools in three formats
(Internet, software, paper) and two cost levels (free, for purchase). Of the
74 different products; 39 are strictly Internet-based for purchase and 11
are strictly Internet-based and free, 15 are standalone software for
purchase and two are standalone software for free, five are paper-based for
purchase and two are paper-based for free, six are available in two
different formats for purchase, and one is available in all three formats
for purchase. AHIMA also provides a paper-based health information form for
adults (15 pages) and one for children (5 pages). The adult form has more
than 600 fields to complete, some of which are repeating fields. In addition
to reviewing many of the free systems, a small sample of commercial products
was purchased for examination. All these systems must display enormous
amounts of information and require scores of fields that users complete.
Figure 1 shows a screen display for the ICER-2-Go main page user
interface[2] . In addition to the system administration buttons near
the bottom of the screen, each of the documentation tabs leads to a screen
with scores of data fields. In the figure, the family information screen is
shown and personal and contact data fields are provided. Including a photograph is a useful feature (especially if these can be saved over time)
but does require knowledge of how to manage and import image data?a quick
and easy step for experienced computer users but a significant challenge for
novices. Clearly, using a PHR requires three substantial kinds of
experience: computer experience (to understand basic interaction mechanisms
such as using menus and tabs, entering data in forms, importing external
files, and opening, saving, and backing up the application and data), health
information experience with medical health concepts and terminology; and
personal record keeping experience to recall or find the information needed
to create the PHR records.
————————————————————————
[1] Thanks to Josh Lemieux of the Markle Foundation for sharing
this information.
[2] This is a screen display from one commercial product,
ICER-2-Go. This is a copyrighted screen display from one of the systems we purchased for review. Including this particular screen display is not meant
to be an endorsement or a critique of this particular system, but rather to
illustrate the complexity of a typical PHR.
Great stuff from all you folks.
I just arrived at LAX, en route to SBA. Will chew on this more when I get done unpacking, getting in sync with the family, swimming in the pool and other necessities.
Meanwhile, I’m wondering how much of the PHR business A) exists primarily for the insurance companies, B) is comprised of windows-only or other forms of silos with major compatibility problems.
Mike-
Excellent points. That could be a very good first step.
Our machine essentially turns paper into PDFs and adds taxonomy to those files so that they make sense to humans later on.This also supports the extensibility for machines to share data in a standard format unassisted.
That’s what I don’t want to lose: the metadata that supports the PDF(s). Otherwise I think it creates another set of problems, in that the PDFs have to be handled (sorted, stored, etc.) by a human being, which again opens up the sort of error-prone processes that electronic medical records are supposed to alleviate.
I think you make an excellent point about going around these beaurocracies. That’s probably the only way to create change in our lifetime. Do you have any thoughts about patients adopting an open platform, therfore pushing the doctors and technology companies to adopt that platform? Also, would you, as a consumer of healthcare, have any interest in being proactive about maintaining the baseline data that would be required to make such a platform viable?
Doc-
Proprietary formats and platforms are a huge problem in our industry. I haven’t taken my concept any further than our product because it would probably take an act of congress (literally) for developers to budge from their silos.
There has to be a creative way to make being closed a disadvantage, and it could very well be led by patients, or patients and doctors. I don’t personally believe it can be led by technology folks, although there will need to be enough adoption by tech folks to create that tipping point where you’re “damned if you don’t.”
As a clinician, I can tell you that everyday a patient comes in and complains that they have to fill out basically the same information that they filled out at the office of the Doctor that referred them to me.
I love the ideas of the patient using some sort of PDF file or even strong the data on the internet somehow (most medical offices have internet access at their reception desks for insurance verification). It saves time, money and would make the patients happier (they wouldn’t have to sit in a waiting area reading People Mag with a bunch of sick folks)
The comment about the compatibility problem is a big one, but if it could be overcome it could save the system oodles of money that could be used more effectively,
Doc, if you are ever up in SLO, I would be happy to show you how archaic the whole thing is (patient intake, billing….) in a small town medical system.
Sam-
I realize there’s a lot of value to be had in tying the PDF to metadata. That requires integration… and thus hits the silo wall. Just making the darned forms available outside of the time window just before seeing the doctor is a fairly understood and convincing way to open up the system, just a tiny bit.
Once the practice becomes commonplace, then it makes sense to try to “improve” the practice with the rest of the reforms we all want, like filling this stuff out once, and having it maintained by the Doctors who name the conditions we’re suffering from, and who actually have the best CLUE as to what’s going on.
Its maddening to me that a patient record as simple as a freakin’ ASCII text file that gets appended and copied everywhere could probably save thousands of lives per year, and millions of dollars… just won’t happen anytime soon. We’ve got intertia and a ton of laywers who will chop on any little mistake to discourage innovation in the US… so it’ll have to happen elsewhere first, then arrive here when it’s so freakin’ obvious to everyone that we have to change it.
I don’t expect it to get better for a very long time, but hopefully it will before I get too old to remember all the things that have happened to me…
Small changes at the outside, with the big picture in mind… flanking maneuvers are key here… a frontal assault WILL FAIL.
–Mike–
Following up on the post of Marchionini’s article clip. The full research into Usability of Personal Health Records including a great bibliography and a link to the 130+ systems evaluated is at this site http://www.ils.unc.edu/phr/
Mike-
I agree, any step in the right direction is a good step.
Your comment that it’s “maddening that a patient record as simple as a freakin’ ASCII text file,” is where we have to go. And where we will go. If it starts out as silo integration based on open architectures, it’s a good first flanking maneuver.
Maybe another direction would be to offer a unique patient/user authentication key in every silo – so that each silo would have an open door to an authenticated individual’s own data. If that facility was provided securely and in a standard way then a patient could access specific data, search across silos, correct/update name and address info, or chronicle/record symptoms for either future treatment or for therapeutic progress purposes.
Its a bear of an idea but the incentive might lie with government and private industry to improve service quality, efficacy, and cost-effectiveness for treatment prescribed versus cost-management for treatment delivered. A unique user account could simply be a number-based account (similar to a Swiss bank account) within each silo (and different for every silo) that would ensure complete anonymity for patients, and at the same time permit anonymous aggregation of data and statistics by healthcare providers, insurers, government, and industry.
Come to think of it – why shouldn’t government, education, transportation, and financial services work the same way?
Leake: I think having access to all of MY data is a good goal, but to the silo owners its a security nightmare. The point of a silo is STABLE storage of data… if anyone could change it the value of it disappears and goes negative. Read-ONLY access would be a fair compromise, if it’s truely secure.
The other reason for silos is centralization… if data is spread around the internet, it’ll be slow at best, and likely unreliable as hell, leading to inconsistent data depending on which set of servers happen to be reachable.
I’d start with just a simple legislative REQUIREMENT that read-only access be mandatory for personal data stored by any corporation, with appropriate restrictions for law enforcement, privacy, security, etc.
–Mike–
How to approach a relationship manager for health care and what are the benefits will get by this approach?