The Economist: Health Care & Information Technology

Fascinating and timely article from the Economist on health care & IT

The inability, and reluctance, of doctors and hospitals to use information technology more widely is killing thousands of people

IT is strategic – when used wisely.

Apr 28th 2005
The inability, and reluctance, of doctors and hospitals to use
information technology more widely is killing thousands of people
“WHETHER or not a treating doctor has Alex’s full medical record
available can literally mean life or death,” says Cynthia Solomon of
Sonoma, California. Her son Alex, now in his 20s, grew up with
hydrocephalus, a rare and life-threatening condition in which fluid
accumulates in the brain and needs to be drained through special
shunts. So Ms Solomon had no choice but to become a walking filing
cabinet of records on allergies, pituitary-gland problems, brain scans
and “every piece of paper a doctor ever wrote about Alex’s case.” She
worried constantly. There were close calls, such as the time that Alex
went on a trip and ended up, unconscious, in some distant hospital. Ms
Solomon could not get his paper records to the new doctor and had to
pray that Alex would not get the wrong antibiotics or be laid on his
back, which might have killed him.
To Ms Solomon the information problem with health care today is so
glaring that she eventually took matters into her own hands, as best
she could. She took out a second mortgage, hired software programmers
and developed a computer system, called FollowMe, for online medical
records that any doctor can, in theory, access anywhere and anytime.
FollowMe will not fix the world’s health-care industry–only about 400
families now use it–but Ms Solomon has correctly identified the
woeful, even scandalous, failure of the health-care industry worldwide
to adopt modern information technology (IT).
The solution seems obvious: to get all the information about patients
out of paper files and into electronic databases that–and this is the
crucial point–can connect to one another so that any doctor can access
all the information that he needs to help any given patient at any time
in any place. In other words, the solution is not merely to use
computers, but to link the systems of doctors, hospitals, laboratories,
pharmacies and insurers, thus making them, in the jargon,
This may be obvious, but today it is also a very distant goal.
According to David Bates, the head of general medicine at Boston’s
Brigham and Women’s Hospital and an expert on the use of IT in health
care, the industry invests only about 2% of its revenues in IT,
compared with 10% for other information-intensive industries.
Superficially, there are big differences between countries. In Britain,
98% of general practitioners have computers somewhere in their offices,
and 30% claim to be “paperless”, whereas in America 95% of small
practices use only pen and paper. But, says Mr Bates, this obscures the
larger point, which is that even the IT systems that do exist cannot
talk to those of other providers, and so are not all that useful.
It shows. People on the right side of the digital divide increasingly
take for granted that they can go online to track their FedEx package,
to trade shares, file taxes and renew drivers’ licences, and to do
almost anything else–unless, of course, it involves their own health.
That information, crumpled and yellowing, is spread among any number of
hanging folders at all the clinics they have ever visited, and probably
long since forgotten about. The most intimate information is, in
effect, locked away from its owners in a black box.
Many IT bosses find this baffling. John Chambers, the chief executive
of Cisco Systems, the world’s largest computer-networking company, says
that health care is down there with mining as the most technophobic
industry. Jeff Miller, a manager at Hewlett-Packard, a large
computer-maker, calls health care “one of the slowest-adopting
industries”, which is especially surreal because hospitals often
splurge on the latest CAT-scan or MRI equipment, but are stingy with
their back-office systems. It is, he says, like “Detroit putting out
futuristic hydrogen cars but using paper processing and manual labour
for the manufacturing.”
This has perverse consequences. According to the Institute of Medicine,
a non-governmental organisation in Washington, DC, preventable medical
errors–from unplanned drug interactions, say–kill between 44,000 and
98,000 people each year in America alone. This makes medical snafus the
eighth leading cause of death, ahead of car accidents, breast cancer
and AIDS. “It’s like crashing two 747s a day,” says Mark Blatt, who was
a family doctor for 20 years before he joined Intel, the world’s
largest semiconductor-maker, to manage its health-care strategy. There
should, he says, be more outrage.
Improving computer systems, of course, would not eliminate all medical
errors. But most researchers believe that they would reduce them
dramatically. One study in America estimates that IT could prevent 2m
adverse drug interactions and 190,000 hospitalisations a year. Another
study reckons that electronic ordering of drugs can reduce medication
errors by 86%. By contrast, research published in March in the JOURNAL
OF THE AMERICAN MEDICAL ASSOCIATION warns that IT, if the software is
badly designed, could actually increase errors. But almost everybody
agrees that well-designed IT is essential to improving quality in
health care.
The same goes for its cost, an increasing burden to ageing societies in
the rich world and even in poor countries such as China. HP’s Mr Miller
reckons that redundancy and inefficiency account for between 25% and
40% of the $3.3 trillion the world spends on health care every year,
and could be eliminated with proper IT. A study from a clinical
research centre at Dartmouth College in New Hampshire reaches a similar
conclusion, estimating that a third of America’s $1.6 trillion in
annual health-care spending (as of 2003) goes to procedures that
duplicate one another or are inappropriate.
Estimating how much IT could save, after taking account of the
considerable cost of applying it widely, is not easy. Writing in HEALTH
AFFAIRS, an American journal, in January, Jan Walker and five
colleagues (including Mr Bates) at the Centre for Information
Technology Leadership in Boston concluded that a fully interoperable
network of electronic health records would yield $77.8 billion a year
in net benefits, or 5% of America’s annual health-care spending. This
includes savings from faster referrals between doctors, fewer delays in
ordering tests and getting results, fewer errors in oral or
hand-written reporting, fewer redundant tests, and automatic ordering
and re-fills of drugs. It does not include, however, perhaps the
biggest potential benefit: better statistics that would allow faster
recognition of disease outbreaks (such as SARS or avian flu).
The key word in all such estimates is always “interoperable”, says Mr
Bates, pointing to the differences between two pilot programmes in
America. In one, the Californian city of Santa Barbara set up a
city-wide peer-to-peer network (in which the computers of different
practices and clinics can talk directly to one another). This allows
doctors, say, to pull up portable-document-format (PDF) files from one
another. But the information in them–text, with numbers buried in
it–is “unstructured” and so not very useful. It is the equivalent of
faster faxing, and not what people mean by interoperability.
The other American pilot, located in Indianapolis and managed by the
Regenstrief Institute, a non-profit medical-research organisation,
comes closer. It has created a city-wide network in which physicians
can, with the patient’s permission, log on to a complete medical
history that includes all previous care at the 11 participating
hospitals. Already, the database contains 3m patient records, 35m
radiology images, 1.5 gigabytes of diagnoses, 20m order-entries by
physicians, and so forth. The key difference is that, wherever
possible, the data is entered in a structured and formatted form. Test
results are in neat rows and columns and tagged in a way that every
other computer can recognise and compare against other appropriate
numbers. This is the sort of IT solution that not only cuts waste and
errors, but also helps physicians to make better decisions.
What, then, would the ideal IT architecture of health care in future
look like? It would start, says Intel’s Mr Blatt, with wireless data
entry by nurses and doctors. Practices and clinics would have secure
“Wi-Fi hotspots”–using a radio technology called 802.11–and staff
would walk around with small handheld devices that transmit all inputs
to the database in the back office. Another source of input might be
tiny radio-frequency identification (RFID) chips that are attached to
patients and send basic information when they come in range of a radio
field. Patients could also add inputs themselves. A firm called Health
Hero, for instance, makes a cute little device called a Health Buddy
that patients take home and plug into their telephone lines. A couple
of times a day, it asks them basic questions or takes their heart rate,
and sends the data to the doctor.
Behind the scenes, all this data would be formatted and stored
according to recognised standards. Contrary to widespread concerns,
this does not require a single central repository or any other
particular hardware architecture. Instead, it relies on common software
protocols and formats so that individual computer applications can find
and talk to one another across the internet. Most of these standards,
such as XML, SOAP and WSDL, already exist and are used by many
industries. Others, such as HL7, LOINC or NCPDP (spelling them out
makes them sound no less obscure) are unique to the health-care
industry and govern data interchange between hospitals, laboratories
and pharmacies. On top of these, there need to be hacker-proof layers
of authentication and password protection so that only the right people
get access.
There is still some work to do to refine these technologies. In
January, eight of the world’s largest IT companies–Microsoft, Oracle,
IBM, HP, Intel, Cisco, Accenture, and Computer Sciences–teamed up to
form an “interoperability consortium” for that very purpose. In
general, however, “the technology is very, very ready,” says Robert
Suh, the technology boss at Accenture, a consultancy that is helping
Britain’s National Health Service (NHS) and regional governments in
Australia and Spain to implement electronic health records.
In fact, Britain’s–or rather England’s–NHS is among the pioneers
worldwide. This year, it will begin rolling out a GBP6.2 billion ($12
billion) project in which five regions in England will form networked
IT”clusters” so that 18,000 NHS sites, including all family doctors and
acute-care hospitals, can share standardised information on patients.
These clusters will eventually be linked through a “spine” (called the
N3 and run by BT) with huge bandwidth to create, in effect, one
national network. Scheduled to be completed by 2010, the plan, like
most IT projects, has had some early hiccoughs and has been greeted
with cynicism by some doctors. But other countries will be looking to
it as a model.
Another pioneer is Denmark, which began rolling out a similar network
for the region around Copenhagen in 2001 and expects to complete it by
2007, before covering the rest of Denmark. Torben Stentoft, the boss of
Hvidovre Hospital in Copenhagen and the head of the city’s network,
says that his main concern is the nitty-gritty of dealing with all of
his legacy computers which need to be tweaked or replaced. But he feels
that he has his society’s full support. “Nobody is against this.
Everybody is asking for it,” he says. In particular, the Danes find
nothing terribly controversial in the idea of a national health
identification number, which they already have, and spend little time
worrying about how to fund the new systems, since their tax kroner are
doing that.
Mr Stentoft is in an enviable situation, especially if viewed from
America, which has the world’s largest and costliest health-care
system. America is as enthusiastic as any country about electronic
health records. President George Bush has embraced the idea, and he
spoke about it publicly some 50 times last year. He has even appointed
a “national co-ordinator for health information technology” to create a
fully interoperable, nationwide network within ten years. But America’s
health-care system is so different from others that it faces some
special complications.
The first big difference is that, whereas most other rich countries
have “single-payer” (ie, government-run) health-care systems, America
has a highly fragmented industry with many private providers and
insurers doing business alongside large government programmes (such as
Medicare, for old people). This means that in funding a new IT
infrastructure “the financial incentives are not exactly aligned,” says
Mr Bates. In single-payer systems, the expenditures come out of the
same pocket–the taxpayer’s–that the savings go into. But in America,
he estimates, the practices and hospitals that pay for the IT only get
11% of the cost savings, with the rest going to insurers and employers
(who buy the insurance). The resulting mismatched incentives, says Mr
Bates, could derail the entire project: “It’s a situation where America
could end up far behind.”
This calls for some combination of government subsidies and
private-sector financial incentives, argues the Markle Foundation, a
charity in New York that is dedicated to the proper use of IT in health
care and national security. Over half of all doctors in America work in
small practices. And, say Markle’s researchers, a typical practice
(defined as five doctors handling 4,000 patient-visits a year) would
make losses if it had to pay the estimated $15,000 a year for three
years that it costs to install an interoperable IT system and to learn
how to use it.
The practices, Markle concludes, therefore need incentives of $3 to $6
per patient-visit, or $12,000 to $24,000 a year, which comes to $7
billion-14 billion a year for three years, or between 1.2% and 2.4% of
total ambulatory-care revenues. The trickier question is how to
administer this largesse, whether it is provided by insurers and
employers or the government. The money could be disbursed directly and
specifically for the IT systems. Or it could be given indirectly in
some sort of pay-for-performance arrangement.
The other big difference between America and countries such as Denmark
is public perception of the robustness of privacy laws. The European
Union has stricter privacy laws than America, and Europeans have
relatively more confidence in them. For information sharing, “ours is a
much more porous environment,” says Alan Westin, a professor at
Columbia University who has written several books on privacy issues.
This is not primarily an IT issue, although the internet does seem to
raise the stakes. In February, one database broker, ChoicePoint, had to
inform some 140,000 people that it had accidentally sold sensitive
information about them. Also in February, a statistician of the health
department in Palm Beach County, Florida, inadvertently e-mailed a list
of more than 6,000 HIV carriers to all employees of the department.
This makes many Americans suspicious of plans that involve sharing
sensitive health information. Although opinion polls in Europe show
overwhelming support for interoperable medical databases as long as
these are properly regulated, a February poll by Harris Interactive
found that Americans are currently evenly split, with 48% saying that
the benefits outweigh the privacy risks, and 47% saying the opposite.
Some 70% of Americans in the poll worried that sensitive data (on
sexually transmitted diseases, say) might leak.
This is unfortunate, says Michael Callahan, a health-care lawyer at
Katten Muchin Zavis Rosenman, a law firm in Chicago, since a weighty
tome of legislation was passed in 1996 precisely to prevent such leaks.
Called HIPAA (short for “health insurance portability and
accountability act”), the law defines strict codes for sharing medical
data and takes effect in stages, with a large chunk of compliance
falling due this month. HIPAA creates a national “floor”, says Mr
Callahan, with some states following even stricter statutes, and
involves the federal government in enforcement and prosecution. HIPAA
is not quite as strong as equivalent laws in Europe, he thinks, but
strong enough.
Mr Westin disagrees. The HIPAA rules are “not at all adequate” for
shared medical records, he says. So the only way to sell such records
to the American public, he says, is to design the whole system with
privacy as a priority. This rules out any form of medical
identification card, to which Americans would be hostile (even though
they think little of giving their social-security numbers, a de facto
ID, when renting DVDs). It also means avoiding a central database that
could be hacked. The best approach, says Mr Westin, is to emulate the
“locators” used by American police. Cops in California who arrest a New
Yorker cannot access information about that person directly, but can
view a directory of such information and request it from the
authorities in New York. Finally, rather than allowing sceptics to opt
out of the new system, says Mr Westin, the system should from the start
require patients actively to opt in.
As the Markle Foundation puts it, the technology must be designed in
such a way that “decisions about linking and sharing are made at the
edges of the network” by patients in consultation with their doctors,
and never inside the network. This goes to the very heart of the
matter. For even though it is fine to start hoping for the day when
interoperable electronic health records create vast pools of medical
information that could be used to find new cures and battle epidemics
in real time, their ultimate purpose is to make one simple and
shockingly overdue change: to enable individuals, at last, to have
access to, and possession of, information about their own health.
See this article with graphics and related items at