The Role of Hospital Transparency

1
The Role of Hospital Transparency
2
The Problem
3
Premise
Transparency the public reporting of cost and
outcomes information will lead to improved
value in health care. The T word is now
political Mom and apple pie. Consumers of
healthcare, made price sensitive by appropriate
product design and informed by detailed outcomes
and price information, will shop for value and
bring market forces to bear on both cost and
quality.
4
The Providers dilemma
5
Four issues with transparency

• Data limitations -- relates to data sources and
  appropriate use of statistics
• Some tasks are harder to accomplish than others,
  some patients are sicker than others -- risk
  adjustment
• The realities of media/consumer understanding
• Unintended consequences of public reporting

6
Johns Hopkins, transparently
2005 Ratings
2005

• Did not earn clinical excellence award 4
  other Baltimore hospitals did
• Only award for specialty excellence was for
  stroke care
• One of the lowest rated hospitals in Baltimore

• Ranked 1 overall, 2 in the nation for
  neurology and neurosurgery, 3 in cancer, 4 in
  heart and heart surgery
• No other Maryland hospital made the top 50

Data limitations
7
Some tasks are harderRisk adjustment is an
attempt to level the playing field

• WHITE COAT NOTES NEWS FROM BOSTON'S MEDICAL AND
  SCIENTIFIC COMMUNITYA NEW WAY TO RANK HOSPITAL
  QUALITY
• Boston Globe, March 2, 2004
• Tops in Heart Attack Care
• Winchester Hospital
• Melrose-Wakefield Hospital
• South Shore Hospital
• Brockton Hospital
• Massachusetts General Hospital (5)
• Beth Israel Deaconess (23)
• New England Medical Center
• Brigham and Women's Hospital (3)
• Boston Medical Center
• Beverly Hospital

An example of what happens when data are reported
without risk adjustment
The hospitals at the top of the list usually
transfer their most serious heart attack patients
to the hospitals lower down.
Risk adjustment
8
How consumers view surgeon quality 43 non-MDs
and non-RNs were asked to select a surgeon. They
were given this chart, assured that there was no
right or wrong answer, and asked to list their
1st, 2nd and 3rd choice
Number of Cases Deaths Expected O/E
ratio Mortality Surgeon A 100 6
(6) 4 1.2 Surgeon B 90 7 (8)
9 0.9 Surgeon C 240 14 (5.8)
3 1.9 
Response Number of Total
C-A-B 23 53
B-A-C 9 21
A-C-B 6 14
C-B-A 4 9
B-C-A 1 2
The Result All three surgeons were ranked FIRST
and LAST by some respondents. Surgeon C was
first for 62. Surgeon B was last for 64.
Media/consumer understanding
9
Adjusted CABG Mortality in Northern New England
n37,599
8
7
6
5
4
Mortality Rate ()
3
2
1
0
1988
1992
1994
1996
1998
1990
2000
Public reporting
10
Uses of Quality Measurement
Improvement
MOTIVATION
Motivation
THE NNE AND NEW YORK STATE EXPERIENCE
11
Quality Measurement in Aortic Valvuloplasty
Creating conflicts

• To palliate congenital aortic stenosis, the valve
  is dilated with a balloon
• Therapeutic success is achieved by maximizing the
  amount of dilation/gradient relief -- use a
  bigger balloon
• Safety is achieved by avoiding rupture/damage to
  the valve -- use a smaller balloon
• Do not measure quality of aortic valvuloplasty
  purely by procedural morbidity/mortality, need a
  measure of efficacy and long term benefit as
  well, otherwise the incentive is purely to use a
  smaller balloon

Lee TH. Torchiana DF. Lock JE. Is zero the ideal
death rate?. New England Journal of Medicine.
357(2)111-3, 2007 Jul 12.
12
Mom and apple pie?
13
Mom and apple pie?
There are two groups of PCI (angioplasty) patients
14
There are six times as many PCI patients in shock
in MA. Why?
Mom and apple pie?
15
Which is preferable?

• A lower mortality rate for PCI?
• A lower mortality rate for MI?
• Patients who receive PCI for MI with shock are
  67 more likely to be alive after 6 years than
  those that dont.
• For 130 of 1000 patients with shock/MI, PCI is
  the difference between life and death.

Source JAMA, June 2006
16
Provider Behavior - Risk aversion
benefit
Risk benefit
too well
too sick
risk
therapeutic range
Severity of illness
17
A way forward
Keeping Score in the Transparency Era
18
Applying the model

• Patient safety
• never events accountability
• safe practices public reporting (and payment)
• Hospital Acquired Infections public reporting
• AHRQ Patient Safety Indicators - improvement

19
Responding with transparency
20
Summary
"there is nothing either good or bad but
thinking makes it so. Hamlet

• Quality measurement has proven value for
  accountability and care improvement
• The notion that transparency and consumerism will
  add value to healthcare is an attractive but
  unproven hypothesis.
• But patients will be harmed by poorly thought out
  reporting
• Avoid zero sum scenarios where one piece of
  public data improves at the expense of another
  hidden outcome
• Use statistics appropriately
• Be aware of the limitations of administrative
  data
• Be careful with reports at the physician level,
  these are the most likely to change behavior
• Some data can be worse than no data