Transcript Trends in costs (2002 $)

Disease Management at
Kaiser Permanente in
Northern California:
Impact on quality and costs
AcademyHealth 2004 Annual Meeting
Bruce Fireman, Joan Bartlett, Joe Selby
Division of Research, Kaiser Permanente
Overview
•
Review the design and implementation of
four DM programs that are generally
consistent with the Chronic Care Model:
– coronary artery disease (CAD), heart failure
(HF), diabetes, and asthma
•
•
Summarize evidence of the programs’
impact on quality and costs
Discuss several “lessons” for KP
2
Kaiser Permanente in
Northern California (KP)
•
•
•
•
•
Integrated, group model, nonprofit HMO
3 million members, ethnically diverse,
similar to California in age distribution
30% of Northern California population
Comprehensive services delivered by
4,400 KP physicians at 17 KP hospitals
and 34 clinics
Longstanding partnership between health
plan and medical group
3
Why KP invested in
disease management
•
•
•
“It’s the right thing to do” to improve
quality
DM promises a “win-win” strategy to
reduce costs by improving quality, thereby
reducing exacerbations and complications
DM builds on KP’s historic strengths and
highlights them
4
Disease management has
evolved and grown over a decade
Disease
registries
Innovation
grants
Guidelines
Early
1990s
1996
Incented quality
measures
Self-care
classes
1998
2001
Expanded
care manager
programs
Plans for primary
care redesign &
First care DM expansion
manager
programs
Primary care
teams
2003
Continue major
IT investments
Expanded
case management
program
5
Programs are consistent with
the Chronic Care Model
•
Key components—clinical guidelines, disease
registries, patient self-management education,
reminders, proactive outreach, risk stratification,
multidisciplinary care teams, and performance
feedback to providers—are integrated in a
comprehensive effort to help:
– clinicians to plan and deliver evidence-based
care; and
– patients to play an active and informed role in
caring for themselves.
6
Trends in prevalence of the
four targeted diseases
Prevalence per 1,000 adults
80
70
60
50
40
30
20
10
0
1996 1997 1998 1999 2000 2001 2002
7
FTEs delivering care
management services
Program
May
1997
July
1999
July
2001
July
2003
CAD
Heart failure
15.2
9.4
19.9
19.8
30.3
34.9
30.9
44.5
Diabetes
Asthma
39.6
2.4
88.0
13.7
127.3
37.7
145.7
36.9
Cholesterol
management
21.9
31.9
54.8
61.7
Case management
23.4
23.2
50.0
75.1
112.0
196.5
335.0
394.7
Total
8
Three types of analyses
•
Examined trends in the four targeted chronic
diseases regarding:
–
–
–
–
•
•
use of recommended tests and medications
risk factors, morbidity, and mortality
use of hospital, emergency room, and clinic services
costs by type of service and overall
Examined whether facilities with more favorable
quality trends experienced more favorable cost
trends
Examined quality indicators for patients served
by facilities with more care manager services vs.
patients served by facilities with less care
manager services
9
Overview of quality and cost
trends from 1996 to 2002
•
•
Increase in use of recommended tests and
medications
Improvement in lipid and blood pressure
control
– appears to have been accomplished more by
medications than by weight management and
exercise
•
Trends in use largely similar to rest of KP
– increase in pharmacy, decrease in visits to
MDs, increase in visits to non-MDs, decrease
in ER visits, same or higher inpatient days
10
Overview of quality and cost
trends from 1996 to 2002
•
Increase in costs (2002 $) for each of the
four conditions
– Costs increased more in real dollars yet by
less or equal percentages, when compared to
patients of the same age and sex and medical
center without the disease.
11
Trends in quality indicators:
coronary artery disease
Quality indicator
LDL
Statin
Blood pressure
Antihypertensive
Beta-blocker*
Any test
Median LDL
Any rx
Median days
1996
1998 2000 2002
43.6% 60.5% 75.4% 86.0%
125.4
112.7
Median days
Any rx
Median days
98.8
27.1% 43.5% 60.0% 73.1%
277.4
306.5
BP < 140/90
Any rx
104.4
318.0
324.3
57.7% 68.4%
80.5% 83.3% 87.0% 89.7%
430.9
472.1
547.1
608.8
43.9% 54.2% 66.2% 75.5%
291.8
302.1
315.7
323.1
12
Trends in quality indicators:
heart failure
Quality indicator
ACE-I or ARB
Beta-blocker
Blood pressure
Antihypertensive
Any rx
Median days
Any rx
Median days
1996 1998
2000
2002
62.0% 68.3%
73.2%
75.8%
340.5
337.5
332.0
17.5% 27.3%
43.9%
59.1%
296.4
314.7
44.2%
50.2%
95.9%
96.3%
713.5
761.8
327.5
247.0
267.7
BP < 130/85
Any rx
Median days
93.0% 95.0%
618.1
665.6
13
Trends in quality indicators:
diabetes
Quality indicator
HbA1c test
LDL test
Statin
Blood pressure
Antihypertensive
ACE-I or ARB
1996
1998
Any test
68.6%
74.9% 82.6% 86.9%
Any test
28.2%
43.5% 65.3% 81.8%
Median LDL
Any rx
Median days
132.0
10.3%
266.9
123.1
Median days
Any rx
Median days
115.6
108.4
16.4% 27.1% 45.2%
276.3
BP < 130/80
Any rx
2000 2002
274.6
279.1
28.1% 35.0%
58.4%
391.6
32.1%
326.2
63.0% 69.0% 72.2%
417.1
449.4
493.7
40.8% 50.9% 56.6%
326.7
323.5
324.9
14
Trends in quality indicators:
asthma
Quality indicator
1996
1998 2000 2002
Inhaled Any rx
corticosteroid Median days
71.5% 80.2%
Long-acting Any rx
inhaled beta2- Median days
agonist
4.4% 11.4%
Short-acting Any rx
inhaled beta2- Median days
agonist
83.9
101.4
94.5
99.1
90.1% 90.5%
106.3
95.1
82.6% 85.0%
99.2
103.4
17.1% 22.7%
100.4
106.6
89.2% 88.6%
87.5
72.4
15
Underuse persists in 2002 but
is less common than in 1996
•
•
•
•
Among CAD patients with an LDL test (86%), 88%
have LDL < 100 or were given statins
Among CAD patients with BP e-recorded (92%),
92% have BP < 140/90 or were given 2+ classes of
antihypertensive medications
Among diabetes patients with an LDL test (82%),
68% have LDL < 100 or were given statins
Among diabetes patients with BP e-recorded
(88%), 68% have BP < 130/80 or were given 2+
classes of antihypertensive medications
16
Change in behavioral risks
factors in survey respondents
with the four conditions
Behavioral
risks
Obese
1996
32.2%
(BMI 30 or (29.7%-34.8%)
higher)
Physically
inactive
49.7%
(Exercise < 3 (46.9%-52.4%)
times per week)
Smoker
1999
34.9%
2002
38.9%
(32.4%-37.4%) (36.4%-41.3%)
46.0%
49.9%
(43.4%-48.7%) (47.5%-52.4%)
11.2%
11.5%
10.9%
(9.5%-13.0%)
(9.9%-13.3%)
(9.4%-12.5%)
17
% change since 1996
Trends in use and costs
for all adults
70%
60%
50%
40%
30%
20%
10%
0%
( 10%)
( 20%)
( 30%)
1996 1997 1998
rx costs
clinic visits
1999 2000 2001 2002
total costs
er visits
inpt days
18
Trends in use for patients with
the four diseases
•
•
Similar to comparison group
– Pharmacy costs soared
– Visits to doctors decreased but visits to other
clinicians increased (total visits rose except
for asthma)
– ER visits decreased
– Inpatient days stayed the same or increased
Different from comparison group
– ER visits decreased more for asthma
– Inpatient admissions decreased (diabetes and
asthma) or increased less (CAD and HF)
– Inpatient days increased less for CAD,
diabetes, and asthma
19
Trends in costs (2002 $):
coronary artery disease
$15,000
$12,000
$9,000
$6,000
$3,000
$0
1996
1997
1998
1999
2000
2001
2002
20
Trends in costs (2002 $):
heart failure
$20,000
$16,000
$12,000
$8,000
$4,000
$0
1996
1997
1998
1999
2000
2001
2002
21
Trends in costs (2002 $):
diabetes
$8,000
$6,000
$4,000
$2,000
$0
1996
1997
1998
1999
2000
2001
2002
22
Trends in costs (2002 $):
asthma
$5,000
$4,000
$3,000
$2,000
$1,000
$0
1996
1997
1998
1999
2000
2001
2002
23
Would costs have increased even
more without the DM programs?
•
•
By the methods of some DM vendors, we
would conclude that—without DM—costs for
CAD, HF, and diabetes patients would be 8% ,
5%, and 11% higher in 2002 (yielding savings
of about $200 million)
But there are plausible reasons apart from
DM to account for a lower percentage
increase in costs for chronic disease patients
– Decomposition of costs by type of service
– HF cost trends similar to COPD cost trends
– Earlier detection of diabetes
24
What are the causal pathways by
which DM can achieve savings?
•
Quality improvement: A diabetes DM
program can:
– Increase use of effective medications, and
– Improve diet, exercise, and self-monitoring.
– Then—further down the causal pathway—the
medications and self-care reduce BP, LDL, and
HbA1c, and
– Eventually—still further down the pathway—
reduce exacerbations and complications, and
– Thereby reduce costs.
25
What are the causal pathways by
which DM can achieve savings?
•
Utilization management: Patients are
given a supportive care manager who
coaches diet and exercise, and also:
– Steers patients to less costly providers,
medications, and hospitals,
– Encourages patients to forego services that
would be duplicative or unnecessary or even
potentially harmful, and
– Facilitates timely discharge from the hospital.
26
What are the causal pathways by
which DM can achieve savings?
•
Productivity improvement: The DM
program could re-engineer the delivery of
usual care in ways that boost productivity:
– Deliver care by telephone, the internet, and
group visits rather than office visits,
– Offload work from physicians to less costly
nurses and health educators, and
– Streamline patient flow, the delivery of
treatments, monitoring, and follow-up.
27
Quality pathway from DM to
savings has intuitive appeal—
what’s wrong with it?
•
DM won’t be cost-saving unless the
treatments it recommends are costsaving.
– most of the treatments recommended in
clinical practice guidelines are cost-effective
(but not cost-saving)
•
By remedying the underuse of costeffective treatments, DM can increase the
value of health care, but won’t reduce
costs
28
Quality pathway from DM to
savings has intuitive appeal—
what’s wrong with it?
•
•
•
Forces that drive up nationwide healthcare
spending are complex and little is known
about their relation to changes in health
This is overlooked when we assume that
because sick people are costly it follows
that the prevention of sickness would help
control rising healthcare costs
Cross-national studies of medical
expenditures do not suggest that a
populations’ health drives per capita
medical expenditures
29
“Lessons” for KP
1.
2.
3.
DM programs can contribute to rapid
improvement in quality when effective
medications are underused.
It was more difficult to reduce obesity or
increase exercise than to increase the
use of effective, well-tolerated
medications.
Opportunities for quality improvement
change over time as usual care
improves. Care manager programs need
to adapt as some of the shortcomings of
usual care are fixed.
30
“Lessons” for KP
4.
5.
Targeting is difficult. Interventions that
target well for underuse of effective
treatments will target poorly for overuse
of costly services.
The causal pathway from improved care
to reduced morbidity to cost savings has
not produced sufficient savings to offset
the rising costs of improved care.
31
Conclusions
•
•
•
•
KP implemented DM programs on a
large-scale
DM is a promising approach for improving
quality of care
DM has not been cost saving
DM should be championed when it
improves health at a reasonable cost
regardless of whether it saves money
32