Its Time to Create an Industrial Revolution in Health Care

1
Its Time to Create an Industrial Revolution in
Health Care

• George C. Halvorson
• Chairman and CEO, Kaiser Foundation Health Plan,
  Inc.
• and Kaiser Foundation Hospitals
• World Bank Presentation
• September 15, 2005

2
About Kaiser Permanente

• More than 8 million members
• 142,000 employees
• 32 hospitals
• 12,000 physicians
• 30,000,000,000 in annual revenue

3
About Kaiser Permanente

• We may be the largest non-governmental,
  vertically integrated care financing and delivery
  system in the world.
• We are the largest system of that type in the U.S.

4
Its time for an industrial revolution in health
care.
5
Its time for health care to stop functioning as
a highly localized, unacceptably idiosyncratic
cottage industry with the exam room functioning
as a medical cottage.
6
Its time to bring a significant degree of
systems support and systematic thinking to
health care.
7
Its time to give the heath care industry a
whole new set of tools.
8
Health care is state-of-the-art science,
significantly handicapped by antiquated,
cumbersome and often dysfunctional information
dissemination and data application
approaches.(The results are expensive and
sometimes dangerous).
9
Health care today is inconsistent, unmeasured,
sometimes dangerous and often wasteful.
10
American Health Care

• Most expensive in the world
• Superb high technology
• Superb techniques
• Mediocre outcomes
• Inconsistent quality

11
In a country with the highest per capita spending
on health care in the world, best care is not a
guarantee for everyone.
12

• More than one half of Americas diabetics receive
  inadequate care.

Source RAND Study, New England Journal of
Medicine, June 26 2003
13
Half of Americas heart attack patients do not
receive proper follow-up care.
Source Institute of Medicine, Committee on
Health Care in America. Crossing the quality
chasm a new health system for the 21st Century,
2001
14

• Nearly 55 of older women with a fracture did not
  receive osteoporosis medications after their
  diagnoses.

Source KP Center for Health Research, Journal of
Bone and Joint Surgery, 12/03
15
UCLA 2003 Heart Patient Study

• 31 of patients considered ideal candidates for
  ACE inhibitors were sent home without them.
• 72 of heart failure patients were released
  without recommended discharge instructions.
• 69 of smokers with heart failure were never told
  to quit.
• 18 of heart failure patients didnt have the
  pumping power of the left ventricles measured.

16
U.S. Inconsistency

• One hundred and thirty-five doctors diagnosed the
  same patient and recommended 82 different
  treatments.

Source Strong Medicine, 1999
17
Ugandan Inconsistency

• Three doctors
• diagnosed the same patient and recommended 3
  different treatments

Source Strong Medicine, 1999
18
Its true everywhere.
19

• Doctors cant keep up
• and
• doctors cant remember everything.
• There are more than 20,000 medical journals.
• Nearly 10,000 articles on clinical trials are
  published annually.
• About 1,000 new drugs were developed over the
  last decade.

Source Foundation for eHealth Initiative,
11/15/03, and the National Library of Medicine,
2001
20

• Care delivery is not set up to manage care
  outcomes -- the care delivery non-systems are all
  built around incidents of care -- maybe episodes
  of care -- but not optimal, patient-focused
  delivery of all appropriate care.

21
Cost Distribution of Care
1 of people
100
of Healthcare Costs
80
30 total cost
60
40
Premium level
20
10 total cost
?
of People
0 total cost
0
0
20
60
80
100
40
20 of people
70 of people
Source Milliman USA Health Cost Guidelines
2001 Claim Probability Distribution
22
Any rational macro system would define and
consistently deliver best care to the truly sick
and would intervene to prevent or delay illness
in and care needs for the other 95 of the
population.
23
That level of intervention would require

• Systematic thinking
• A system

24
The doctor in the exam room needs all of the
information about each patient, about each
patients disease, about best care for each
disease, and about the support needed and
available for the patient.
25
Wobbly Parts of U.S. Health Care

• Five are glaringly obvious

• The medical record
• Inconsistent access to current science
• Patient compliance
• Patient follow-ups
• Outcome tracking

26
Wobbly Part 1The Medical Record

• Incompatible
• Inconsistent
• Often illegible
• Generally isolated
• Too often inaccurate
• Not interactive
• Woefully inadequate

27
Wobbly Part 2Inconsistent Access to Current
Science

• Physicians have a very hard time staying current
  with medical best practices.
• The translation of medical research into practice
  is slow. It takes an average of
• 17 years for new knowledge generated by
  randomized control trials to be incorporated into
  practice.

Source Foundation for eHealth Initiative,
November 3, 2003
28
Every other industry has used systematic quality
improvement techniques to improve its products.
29
DVD Players

• 700 five years ago
• 70 today

(better functionality at one-tenth of the cost)
30
How does that happen?
31
By constantly re-engineering the product.
32
How does re-engineering actually work?
33
How did General Electric achieve Six Sigma
production standards?
34
By using basic re-engineering tools.
35
Four MAIC Steps

• Measure
• Analyze
• Improve
• Control

Source "Jack Welch and the GE Way", Robert
Slater, 1999
36
Measure

• Identify the key internal process that influences
  critical-to-quality characteristics (CTQs) and
  measure the defects generated relative to CTQs.
• Defects are defined as out-of-tolerance CTQs.
• Success occurs when you can measure the defects
  generated for a key process affecting the CTQ.

Source "Jack Welch and the GE Way", Robert
Slater, 1999
37
Analyze

• Why are defects generated?
• Brainstorm, use statistical and other relevant
  tools to identify key variables that cause the
  defects.
• The output of this phase is the explanation of
  the variables that are likely to drive process
  variation the most.

Source "Jack Welch and the GE Way", Robert
Slater, 1999
38
Improve

• Confirm the key variables and then quantify the
  effect of these variables on the CTQs.
• Identify the maximum acceptable ranges of the key
  variables.
• Make certain the measurement systems are capable
  of measuring the variation in the key variables.
• Modify the process to stay within the acceptable
  ranges.

Source "Jack Welch and the GE Way", Robert
Slater, 1999
39
Control

• Ensure that the modified process now enables the
  key variables to stay within the maximum
  acceptable ranges using tools such as statistical
  process control (SPC) or simple checklists.

Source "Jack Welch and the GE Way", Robert
Slater, 1999
40
The process starts with data.
41
Health care lacks data.
42
Where does health care get the data it uses now?
43
From paper medical records.(And not even one
complete paper record per patient).
44
One patient, four doctorsFour unrelated,
unconnected, non-communicative, non-intuitive,
non-interactive, often inaccessible and too often
illegible paper medical records.
45
Four doctors cannot always coordinate care.
They often repeat tests and prescribe
conflicting drugs.
46
So health care lacks both data and real care
coordination.
47
No outcomes data.No process data.No data
sharing.Minimal data and minimum data
availability.So what happens when no one has
data?
48
Bad and expensive care.
49
Hormone Replacement Therapy for women
50
Knee Surgery
51
The Knee Surgery Story

• More than 650,000 relatively expensive
  arthroscopic lavage or debridement procedures
  were performed each year to relieve the pain of
  osteoarthritis of the knee.
• Outcomes Seemed Satisfactory
• About half of the patients reported some relief
  as a result of surgery.

Source New England Journal of Medicine, July
11, 2002
52
Scars vs. Cuts

• A study was done. Some patients were just given
  scars instead of surgery fake surgery in
  effect.

Source New England Journal of Medicine, July
11, 2002
53
In a clinical trial with a matched control group,
surgery was no more effective than a placebo in
eliminating pain and improving function.
The Result?
Source New England Journal of Medicine, July
11, 2002
54
Or the Vioxx Situation?(More than 2 million
people worldwide were taking Vioxx in 2004).
Source The Washington Post, October 1, 2004
55
Based on Kaiser Permanente Data
Vioxx Estimate Up to 140,000 Got Heart Disease
40,000 Unnecessary Deaths
Source The Lancet, January 25, 2005
56
Or autologous bone marrow transplant treatment
for breast cancer?
57
So what should be done?Think about how costs
are actually distributed in health care.
58
Cost Distribution of Care
1 of people
100
of Healthcare Costs
80
30 total cost
60
40
Premium level
20
10 total cost
?
of People
0 total cost
0
0
20
60
80
100
40
20 of people
70 of people
Source Milliman USA Health Cost Guidelines
2001 Claim Probability Distribution
59
We need systematic approaches to determine what
best care looks like.
60
We need systematic approaches to support
physicians in delivering that care in the exam
room at the point of care.
61
We need real and useful data about what really
works.
62
Patients need to know which doctors detect breast
cancer well or fix knees well or cure skin
cancer most often.
63
Major differences in performance exist within the
care system.
64
Life Expectancy for Cystic Fibrosis Patients
Years
SOURCE Data from 1997 Cystic Fibrosis
Foundation Patient Registry, as reported in The
New Yorker, December 12, 2004
65
So what happens when care is systematically
supported?
66
Kaiser Permanente is investing over 3 billion
dollars over the next 5-plus years to implement a
total computerized physician tool kit including
an automated medical record. Why are we
spending that money?
67
Kaiser Permanente pilot studies show that
systematic support of consistent approaches to
care really do work.
68
Improved Outcomes Ohio Death Rates From Heart
Disease
2.18
(per 1,000)
1.36
Ohio
KP
0.87
0.32
0.29
0.13
Ischemic Heart Disease
Myocardial Infarction
Congestive Heart Failure
Ohio vs. KP in Ohio
69
Improved Outcomes Renal Disease in Southern
California

• Computer tracking of follow-up care resulted in a
  31 reduction in the death rate from end-stage
  renal disease.

• Source Department of Health and Human Services
  Medicare Demonstration Project

70
Improved Outcomes Anti-Coagulation in Colorado

• What happens when an AMR is attached to a
  pharmacy review program to monitor and track
  patients on anti-coagulation therapy?
• A 79 reduction in bleeding complications.

71
Improved Outcomes Reduced Drug Interactions

• A recent study of intensive care patients by
  Kaiser Permanente found that when physicians used
  a computerized physician order entry system
  (CPOE), incidents of allergic reactions and
  excessive drug dosages dropped by 75.

• Source Foundation for eHealth Initiative,
  November 15, 2003

72
Standard breast cancer detection scorekeeping
73
90 percent of cancers detected before reaching
Stage II
74
E-support

• -- e-visits
• -- e-scheduling
• -- e-reminders
• -- e-care

75
And e-efficiencies.
76
Nurse Shift Change (paper records)

• Cumbersome
• Error-prone
• Time-consuming
• Dangerous to patients
• 43-minute average

77
Electronically-Supported Shift Change

• Better information
• High level of accuracy
• Less time away from patients
• (43 minutes to 12 minutes)

78
Ordering lab tests for hospitalized patients
--switch from first in/first out (FIFO) to
prioritized scheduling based on urgency and
discharge date
79
Eliminated one day in the hospital for every
four patients.
80
Potential Savings 80,000 hospital days for
diabetes alone.(based on systematic follow-up
care)
81
How Can Technology and Data Help?

• CPOE in the outpatient setting would prevent more
  than 2 million adverse drug events and 190,000
  hospitalizations per year, saving 44 billion in
  costs for medication, radiology, laboratory work
  and hospitalization.
• (Its like refrigerating food in restaurants.
• It helps).

Source Center for Health Transformation, 2004
82
Final reason to create that AMR
toolkitResearch Real and timely research
requires digitalized data.
83
Its necessary to get the health care industry on
board.
84
How Can We Get Health Care Leaders on Board?

• All physicians want to do the best job for their
  patients.
• Use data examples and anecdotes to prove that the
  best job involves physician support tools.
• Make the system ergonomically useful not only
  for physicians but for the entire health care
  industry.

85
How Can We Get Health Care Leaders on Board?

• Use physician leadership and governance at all
  key points to avoid any sense that the
  patients best interest is not the primary focus.
• Make it interactive. Involve both the patient and
  the caregiver at the point of care to show that
  individualized care, competent care and
  culturally sensitive care are not mutually
  exclusive.

86
The opportunities are immense.
87
Those opportunities cannot and will not be
realized without
88

• An electronic medical record
• Computerized physician support tools

89
Without these computerized tools

• Care is inconsistent
• Prevention is ineffective
• Huge amounts of resources are wasted
• No one knows what actually happened (in truly
  useful ways)

90
Both are needed.Both are possible.Both are
overdue.
91
The result will be a true industrial revolution
in health care better care for less money.
92
Thats the number one opportunity for health care
today
93
Lets truly modernize medicine.Its overdue.