The people of the United States are finally pulling together around the
goals of reducing health care costs (by far the highest per capita in
the world) and improving outcomes (we have the worst health of any
developed country). Everyone seems to recognize the critical
importance of data and communications in these efforts. So several of
us at O’Reilly Media, having been involved with information
technologies for some time, are tracking the issues that come up in
deploying computer technology in health care–not just to streamline
payments, not just to facilitate access by doctors to records, but
actually to create new ways to deliver and track health care.
I recently attended a forum on how my state, Massachusetts, is
facilitating the move to Electronic Health Records, a prerequisite for
many things doctors, patients, and insurance companies can do to
improve health. It’s notable that the chief sponsor of the event, the
Massachusetts Health Data Consortium,
derives a lot of its support from insurance companies. Lots of invective has be
thrown at these companies recently, but the questions of technology
can pull together the insurers, providers, and patients in a common
[AO: My original blog said that insurance companies set up MHDC,
but this was incorrect.]
My own understanding of the progress and frustrations in deploying
heath care technology was enhanced by the conversations I had that day
and the statistics bandied about.
Smallness is not nimble
Small doctors’ offices went along with other industries and services
in the 1970s and 1980s by computerizing–but certain parts stayed in
the 1960s. So now, the big stumbling block that doctors face–adapting
their workflows to computerization–is reminiscent of the problems
other industries had in the 1970s and 1980s.
A typical scenario in those other industries was to use the waterfall
development model (or something even less structured) and setting
aside a requirements phase during which managers confidently told
application designers, “This is our workflow.”
OK, requirements phase over and done with. The applications were
deployed, and it turned out that hundreds of little things the line
staff did every day were left out of the workflow. In other words, The
official workflow that the manager knew about was not rich and subtle
enough to encompass reality.
Doctors are more aware than those managers of how hard it is to
formalize their workflows, but the mismatch between current practice
and computer ideal is just as great. I can understand doctors’
reluctance to install electronic systems because in their work, more
than most, when one standardizes their data ands fit their
observations to the structure of the electronic record, lots can
easily get lost in translation. Not only is every patient and every
symptom a bit different, but every doctor is different in the manner
of observing patients and recording results.
It would seem, therefore, that the unique individuality of the
doctor–as well as the patient–would make digitization of records a
problem that was independent of the size of the provider’s
organization. But still, some aspects of conversion are easily when
it’s done on a large scale. Thus, speakers at the conference suggested
that the biggest barrier to adoption is the fragmentation of medical
According to the president of the Massachusetts eHealth Collaborative,
Micky Tripathi, 80% of US medical practices consist of just one or two
physicians, and small practices handle 90% of all outpatient visits.
Given the difficulties of electronic record conversion and the
reluctance of doctors to put in the effort, only 1% of these practices
currently use electronic health records.
I suppose that small practices may improve care the way small
restaurants cook more gourmet food, but the structure of medical
practices clearly reduces efficiency. But electronic records hold out
hope as well. Doctors can use computer technology to accommodate their
preference for small practices.
The division of the industry into tiny segments also increases the
risk of lost records (especially when they’re on paper, but even
electronically). Forum speakers said the patient must ultimately be
the steward of his or her own records.
It’s great when a new practitioner can start up an office tightly
oriented around electronic health records. Converting an existing
paper-based site is much harder. And I wonder whether EHR proponents
are willing face the cost (in time as much as money) of training.
Issues arise when large institutions adopt electronic records,
too. For instance, who gets control and the right to use the data, the
employer or the doctor who actually generates the data? As we’ll see,
resolving questions about who benefits are critical to adoption.
Always somebody else’s job
Tripathi made another impressive point: doctors don’t see the benefit
to them in digitizing. In fact, they see it as benefiting everybody
else but them. I think this is because they don’t know of any killer
app that would make the change not only desirable but indispensable.
This is the same reason Linux hasn’t taken over on the desktop, even
though its applications and interfaces are quite easy to use and
highly functional. Linux can do everything Windows does–but for the
average computer user, it doesn’t do anything more than
Windows does. After looking over the applications provided on a common
desktop distribution such as Ubuntu, one could well ask–why bother
In short, if a system just does the same thing as the one it’s
replacing–even if it does that thing more efficiently and with fewer
errors–it won’t generate enough excitement to drive adoption and make
adoption seem worth the pain.
The move to electronic records seems to have garnered support from the
leadership of all the important stakeholders: the medical profession
(as illustrated by the participation of the American College of
Physicians and Massachusetts Medical Society today), insurers,
governments, and patient advocates. But as I’ve mentioned, support in
the medical profession hasn’t penetrated to the grassroots.
I wonder whether doctors could be swayed by stories of promising
apps–such as the one used by the Army’s mCare Telehealth initiative
for Wounded Warriors suffering from Traumatic Brain Injury (TBI). As
described to me by Stuart Vaeth of Diversinet Corporation, which
created the mobile platform underlying the application, doctors can
use this little utility to remind soldiers to follow treatment plans.
Better health, fewer doctor visits, lower overhead–all in one small
Another expert I talked to claimed that wouldn’t suffice. What doctors
need is to be offered an array of useful applications that are
guaranteed to work with the platforms they choose to install. This
suggests that standardization is even more important than is usually
thought, because it will create a platform for new applications. That
leads to my next point.
Without speaking about current health care bills, I’ll focus for a
moment on the medical part of the stimulus package, which suggest ways
government intervention can help fix health care.
First comes standard-setting. The most obvious role here is to promote
standards that ensure the different systems adopted by different
providers fit together. As reported by Ray Campbell, executive
director of the Massachusetts Health Data Consortium, the
Medicare/Medicaid part of the bill requires conformance to
requirements that will force conversion. God is still in the (yet to
be worked out) details concerning meaningful use.
Second comes actual investment. When up-front expenditures are high
and it takes years to recoup them, and especially when most sites are
small, somebody has to provide incentives. Campbell said the best
stimuli are actually up-front. In the case of the the American Recovery
and Reinvestment Act (ARRA) stimulus, the up-front stimuli aren’t even
direct cash bonuses (which make up the bulk of the stimulus, but after
conversion is complete and judged to conform). Rather, the up-front
stimuli are research centers that provide expertise to sites that want
Although I am still holding back on commenting about the current
health care bills, I should reveal that I worked on grassroots efforts
to pass the 2006 Massachusetts health care bill that inspired
others to push for the current efforts by Obama, Pelosi, and Reid.
With all its problems, the Massachusetts bill put all of us on the
line in this state to work together to solve our health problems.
So instead of fearing a takeover of health care by government (it’s
strange how Americans tend to view “government” as some homogeneous
monstrosity, like the mysterious obelisk in the movie 2001),
the public can view legislation as a way to stand together as a
The one thing of which I am convinced is that we need to replace the
current adversarial model of allocating health money with a
cooperative one. Only if providers and patients get lots of data, and
are willing to use it creatively, can we lower costs.
It worked with the energy utilities. Until recently, they operated
under a simple model where the more energy the sold, the more money
they made. Regulators and utilities worked together to find new models
that rewarded them more for conservation than for consumption. Health
care is a more complex problem–more levers and pulleys to the system,
and subjects who face a diversity of diagnoses–but with all that is
at stake, we should find a way forward.