cards: Making the grade
Will a click of the mouse replace word of mouth
when it comes to picking a hospital?
These days, hospitals — like students — have
to worry about getting a bad report card. Employer groups, health
insurers, government agencies, and even newspapers are amassing mountains
of data about them, crunching the numbers, and assigning scores and
rankings. Several private companies have turned the collecting and
assembling of hospital data for public consumption into a nice little
business. Patients and their families are increasingly expected to
act like savvy shoppers, making smart choices based on all this computerized
That’s the idea, anyway. Reality isn’t
quite so simple. There’s no easy way to summarize quality.
You get at it indirectly through what can be measured; but the individual
measurements have problems. Mortality rates must be adjusted to take
into account how sick patients were to begin with, and adjustment
formulas vary. Patient volume is often used to judge quality because
it’s easy to calculate. Studies have shown an association between
volume and better outcomes, but it’s not a perfect correlation
by any means. Any overall grade is based on assumptions about the
relative importance of individual measurements — assumptions
that are open to question.
There are practical problems, too. The information
used to create the report cards is often years out of date because
of lags in the release of data from Medicare and other government
agencies. Participation is often voluntary, and some hospitals have
balked at providing information.
A 1999 Institute of Medicine (IOM) report on medical
errors sparked this interest in grading hospitals. The sound bite
version of the report: At least 44,000 and as many as 98,000 Americans
die each year from preventable medical errors committed in hospitals.
Putting information in the hands of the public was supposed to be
part of the solution. It should be, but now we have competing report
cards and perhaps some information overload. Whether these report
cards will be used to restrict access to certain hospitals is another
We’ve thumbnailed six of the hospital report
card services below. But if you’re interested in these quality
measurements, you should get on the Web sites yourself. That’s
really the best way to get comfortable with this kind of information.
How credible are the report cards and rating systems?
They each fall short in some way. And they’re certainly not
the only thing to go by when you’re picking a hospital — or,
by extension, a doctor. Think of them as pieces in the puzzle. And
we’d still ask around to see if people think it’s a good
U.S. News & World
The news magazine started ranking hospitals in 1990,
and the winners love to brag that they’re at the top of the
standings. The top 50 hospitals in 17 specialties are listed. Searching
on the Web site is free. But the magazine’s rankings put great
weight on “reputational scores” calculated from surveys
of doctors. Is that anything more than conventional wisdom? Bigger
also tends to get rated better because hospitals are graded partly
on how much high-tech equipment and how many services they provide.
Large employers started Leapfrog after the IOM report
came out. The stated goal is to use their purchasing power — and
consumer choice — to get hospitals to reduce errors and improve
overall quality. The group first asked hospitals to supply information
on patient volume, computerized drug ordering, and several other
items. Now it is surveying them about 30 safety practices that will
be combined into an overall safety score. Access to the Web site
is free, but you have to click through several pages before you get
to information about a particular hospital. Participation is voluntary,
so you may not find anything about the hospital you’re interested
Consumer Checkbook is a consumer publisher and Web
site supported by the Center for the Study of Services, a Washington,
D.C., nonprofit organization. Its guide includes “desirability” ratings
for hospitals based on surveys of physicians, risk-adjusted mortality
figures, and adverse outcome rates for several surgical procedures.
The desirability scores are just a matter of physician opinion. The
mortality and outcome information is based on 1996–99 Medicare
data, so it’s out of date. The Web version of the guide can
be hard to follow. The guide costs $19.95, either in print or through
the Web site.
Health Grades, Inc.
Health Grades, a Lakewood, Colo., company, uses a
five-star grading system based largely on mortality data from Medicare.
It doesn’t give hospitals overall scores, but rates them by
individual procedures and conditions; a hospital could get five stars
for coronary bypass, but just one for hip replacement. Health Grades
has been criticized for keeping its formula for adjusting mortality
rates a secret. Yale researchers evaluated the company’s report
cards in a paper published in 2002. Over all, higher ratings by Health
Grades correlated well with better care, but researchers also found
marked differences between hospitals with the same rating. Access
to the Web site is free.
National Voluntary Hospital
This is Medicare’s foray into hospital quality
rating. So far, three conditions are covered: heart attack, heart
failure, and pneumonia. Hospitals have been asked to provide information
on certain aspects of care for those conditions — for example,
the percentage of heart attack patients that receive an aspirin upon
arrival. Each hospital’s performance is compared with results
for the top 10% and 50% of hospitals. For example, if 99% of heart
attack patients at the top 10% of hospitals get an aspirin upon arrival,
but only 70% of the patients at your hospital do, maybe heart attack
care at your hospital isn’t what it should be. This is just
a pilot project, so many hospitals aren’t listed. It would
be helpful to have some explanation for why these particular aspects
of care are important.
Select Quality Care
Select Quality Care is a Web site developed by HealthShare
Technology, an Acton, Mass., company. Unlike the other sites on this
list , it’s not accessible to the public — only to people
enrolled in health plans that HealthShare has contracts with. At
least one other company is selling a similar, members-only report
card Web site to health plans. One nice feature of the site is that
the hospital rankings depend on how you personally rate the importance
of patient volume, mortality rates, and several other factors. If
you revise your preferences you can see how it affects the rankings.
The site also has graphics that make the averages for your area clear,
so you can easily tell whether a hospital falls above or below the