Transcript Niek Klazinga (University of Amsterdam)

Health System Performance
Management
quality for better or for worse
• Niek Klazinga, April 27 2010
• London LSE/NHS Confederation
Dept Social Medicine
Reasons for international comparisons on
performance related to quality of care
• Accountability
• Strategic decision making
• Learning/improvement
Dept Social Medicine
Table 1.1 Conditions under which performance measurement
is possible and problematic
Performance measurement possible
Performance measurement problematic
 An organization has products
 An organization has obligations
 Products are simple
 An organization is product-oriented
 Autonomous production
 Products are isolated
 Causalities are known
 Quality definable in
performance indicators
 Uniform products
 Environment is stable
and is highly value-oriented
 Products are multiple
 An organization is process-oriented
 Co-production: products are
generated together with others
 Products are interwoven
 Causalities are unknown
 Quality not definable in
performance indicators
 Variety of products
 Environment is dynamic
Source: Managing performance in the public sector. De Bruijn H. (2002), p. 13
Dept Social Medicine
Measurement and Management
• A measure on quality of care does not exist independently
• validation is dependent on the use/purpose
• Validation is dependent on the boundaries of the universe it
is supposed to signal upon
• Measures need to be integrated in management/decision
making mechanisms of government, financiers, managers,
professionals and patients
• Apart from reliability and validity, relevance and usefullness
are important criteria for selecting quality measures
• As a consequence the users should be involved in the
development of the measures
Dept Social Medicine
Health systems performance management
• Health Systems (scope , components and boundaries)
• Performance (objectives on various dimensions such
as health results, efficiency and equity –
measurement challenges)
• Management (heterogeneous national governance
models, integration of performance indicators in
management mechanisms)
Dept Social Medicine
Related policies
•
•
•
•
•
•
•
Health system sustainability
Integrated care
Prevention
Patient Centered Care
Equity
Regulated market
Incentive structures
Dept Social Medicine
Conceptual Framework for OECD
Health Care Quality Indicator
(HCQI) Project.
(shaded area represents the current focus of
the HCQI Project)
Source: Arah OA, et al. A conceptual framework for the OECD
Health Care Quality Indicators Project. International Journal
Quality Health Care. 2006; Sep 18; Suppl.1:5-13.
7
7
Combining various rationalities
• Public Health
• Medicine
• Management sciences
• Economics
• Societal / individual values
Dept Social Medicine
Performance indicators and benchmarking
related to mortality data
- avoidable mortality (health system level)
- standardized mortality rates (hospital level)
- limitations of death statistics
Dept Social Medicine
Dutch hospital standardised mortality
ratios 2001-3(HSMRs) vs hospital
(standardised for age, sex, urgency/readmission, LOS within 50 CCS groups leading to 80% all deaths,
excluding small hospitals and those with poor data recording, using year 2000 standard)
140
HSMRs (95% CIs) 2001-2003
120
100
80
60
40
20
0
Hospital number (assigned by BJ)
Performance indicators and benchmarking
related to cancer care
• CONCORD study
• Eurocare
• Limitations of cancer registries and limited
possibilities for linking with other (administrative)
data-bases
Dept Social Medicine
5.7.1. Cervival cancer screening, percentage of women
screened aged 20-69, 2000 to 2006 (or nearest year)
5.7.2 Cervical cancer five-year relative survival rate, 19972002 and 2002-2007 (or nearest period)
83.5
United States 1
United Kingdom 1
79.4
Sweden 1
78.6
75.6
Norway 1
Canada 2
France 2
76.5
74.1
Korea
Canada
71.9
61.9
Iceland
71.0
72.8
Japan
70.6
72.4
Finland
69.0
66.0
Iceland 1
71.0
Netherlands
69.0
63.3
New Zealand 1
70.6
New Zealand
67.7
63.0
Finland 1
70.5
France
67.3
Netherlands 1
69.6
United States
67.0
66.4
Denmark 2
69.4
Norway
65.9
67.8
Sweden
65.8
62.9
OECD (14)
65.6
Ireland
63.3
54.1
Czech Republic
61.6
62.0
Belgium 1
65.3
OECD
64.0
62.2
Ireland 1
60.6
Australia 1
41.7
Luxembourg 1
2006
2003
2000
27.5
Hungary 1
24.5
Japan 2
0
1. Programme. 2. Survey.
20
40
61.3
65.5
Denmark
38.5
Italy 1
60
80
100
Percentage
57.6
United Kingdom
50.1
Poland
0
20
2002-2007
1997-2002
40
60
80
100
Age-standardised rates (%)
5.7.3. Cervical cancer mortality, females, 1995 to 2005 (or nearest year)
Age-standardised rates per 100 000 females
18
1995
2000
2005
16
14
12
10
0
Sources: OECD HCQI Data 2009. Survival rates are age standardised to the International Cancer Survival Standards population. OECD
Health Data 2009 (cancer screening; mortality data extracted from the WHO Mortality Database and age standardised to the 1980 OECD
population). The 95% confidence intervals are represented by H in the relevant charts.
0.6
0.7
1.1
1.2
1.3
1.5
1.6
1.7
1.9
1.9
2.0
2.1
2.1
2.1
2.2
2.4
2.4
2.4
2.9
2.9
3.0
3.4
3.7
4.3
5.7
4.7
2
6.0
4
6.9
6
11.4
8
5.8.1. Mammography screening, percentage of women
aged 50- 69 screened, 2000 to 2006 (or nearest year)
5.8.2 Breast cancer five-year relative survival rate, 19972002 and 2002-2007 (or nearest period)
Netherlands 2
89.0
United States
90.5
88.6
Finland 1
86.2
Iceland
88.3
Canada
87.1
85.6
Sweden
86.1
83.8
78.1
Ireland 1
76.7
Norway 1
72.5
United States 2
United Kingdom 1
70.7
Canada 2
70.4
Japan
86.1
Finland
86.0
82.0
85.2
80.0
Luxembourg 1
63.5
Netherlands
OECD
62.2
France
82.6
Iceland 1
62.0
Denmark
82.4
76.2
New Zealand
82.1
77.0
Norway
81.9
80.5
OECD (14)
81.1
Hungary 1
60.2
New Zealand 1
60.1
Italy 1
59.6
Belgium 1
59.0
56.2
Australia 1
47.1
France 1
35.6
Czech Republic 1
2006
23.8
Japan 2
2003
2000
19.5
Slovak Republic 1
0
1. Programme. 2. Survey.
50
100
Percentage
United Kingdom
77.9
Ireland
76.2
72.2
Korea
75.5
76.9
Czech Republic
75.4
70.8
61.6
Poland
2002-2007
1997-2002
0
20
40
60
80
100
Age-standardised rates (%)
5.8.3. Breast cancer mortality, females, 1995 to 2005 (or nearest available year)
1995
2000
2005
Age-standardised rates per 100 000 females
40
30
0
1. Rates for Iceland and Luxembourg are based on a three-year average.
Sources: OECD HCQI Data 2009. S urvival rates are age standardised to the International Cancer Survival Standards population. OECD
Health Data 2009 (cancer screening; mortality data extracted from the WHO Mortality Database and age standardised to the 1980 OECD
population). The 95% confidence intervals are represented by H in the relevant charts.
5.8
10.4
11.0
16.7
19.2
19.3
19.3
19.5
19.5
19.9
20.0
20.3
20.5
20.7
20.8
21.1
21.3
21.5
22.4
22.4
23.1
23.9
24.9
25.1
25.8
27.0
24.2
10
28.4
29.5
20
5.9.1. Colorectal cancer, five-year relative survival rate, total and male/female, latest period
67.3
Japan (1999-2004)
66.0
68.7
66.1
Iceland (2003-2008)
63.2
69.2
65.5
62.0
Finland (2002-2007)
62.0
57.0
60.9
New Zealand (2002-2007)
62.3
59.6
60.7
Canada (2000-2005)
62.3
59.6
59.8
Sweden (2003-2008)
64.5
55.2
58.1
Korea (2001-2006)
57.1
59.1
58.1
Netherlands (2001-2006)
58.2
58.4
57.8
Norway (2001-2006)
59.0
56.9
57.2
OECD
57.9
56.3
57.1
France (1997-2002)
58.5
56.6
54.4
Denmark (2002-2007)
54.8
54.2
52.3
Ireland (2001-2006)
54.3
50.7
50.7
United Kingdom (2001-2006)
51.5
50.1
46.8
40
20
48.5
45.6
Czech Republic (2001-2006)
38.1
100
80
60
Age-standardised rates (%)
65.1
65.9
United States (2000-2005)
0
Female
39.3
34.7
Poland (2002-2007)
0
20
40
Male
60
80
100
Age-standardised rates (%)
5.9.2. Colorectal cancer, five-year relative survival rate,
1997-2002 and 2002-2007
65.5
62.5
United States 1
Finland 2
62.0
60.0
New Zealand
60.9
57.0
Canada 1
60.7
59.6
Sweden
60.1
57.4
Korea 3
58.1
52.3
Netherlands 3
58.1
56.9
OECD (11)
57.9
54.6
Norway 3
57.8
55.0
Denmark
54.4
50.2
Ireland 3
52.3
48.9
46.8
41.1
Czech Republic 3
0
20
2002-2007
1997-2002
40
60
80
Age-standardised rates (%)
5.9.3. Colorectal cancer mortality, 1995 to 2005 (or nearest
year)
Hungary
Czech Republic
Slovak Republic
New Zealand
Denmark
Norway
Ireland
Poland
Netherlands
Belgium
Germany
Portugal
Spain
OECD (27)
Austria
Luxembourg
Canada
United Kingdom
Japan
France
Italy
Sweden
Australia
Korea
Iceland
United States
Switzerland
Finland
Greece
Mexico
31.9
31.0
29.8
25.3
25.0
21.4
21.0
20.8
20.6
20.2
19.8
19.8
19.2
19.0
18.8
18.2
18.0
17.6
17.6
17.2
17.0
16.8
16.7
15.2
14.6
14.4
14.2
13.2
12.1
5.2
2005
1995
0
10
20
30
40
Age-standardised rates per 100 000 population
1. 2000-2005 rather than 2002-2007. 2. 1998-2003 rather than 1997-2002 3. 2001-2006 rather than 2002-2007.
Sources: OECD HCQI Data 2009. Survival rates are age standardised to the International Cancer Survival Standards population. OECD
Health Data 2009 (mortality data extracted from the WHO Mortality Database and age standardised to the 1980 OECD population). The 95%
confidence intervals are represented by H in the relevant charts.
Performance indicators and benchmarking
on care delivered in hospitals
PATH, OECD, many national projects …………
Limitations (administrative) data-bases
- Quality of coding practices
- Lack of (internationally) standardized procedure codes
- Lack of coding of secondary diagnoses
- Lack of present at admission coding
- Lack of linking via UPI’s
- Limitations Electronic Health Records
Dept Social Medicine
5.4.1. In-hospital case-fatality rates within 30 days after admission for AMI, 2007
9.6
8.1
7.6
9.9
10.7
9.1
10.9
Slovak Republic
6.6
Luxembourg (2006)
6.6
Netherlands (2005)
6.3
United Kingdom
6.1
Spain
9.2
7.7
5.3
Czech Republic
5.1
United States (2006)
5.1
Ireland
7.0
8.3
4.9
OECD
4.9
Finland
4.5
5.6
4.5
Poland
7.7
11.0
7.7
6.6
5.3
6.4
Crude rates
15
10
Rates per 100 patients
New Zealand
Norway
2.9
Denmark
2.9
3.6
5
Italy (2006)
3.2
Sweden
2.1
Iceland
0
8.7
7.1
6.0
6.7
5.8
6.5
5.6
5.5
5.1
5.4
4.9
5.0
5.2
4.9
4.9
4.2
5.6
4.6
4.5
4.5
4.5
4.4
4.0
4.3
3.7
3.2
3.3
3.4
3.0
3.1
2.7
2.9
3.0
0.9
3.3
Canada
3.3
4.6
6.6
4.8
Austria (2006)
4.2
4.0
6.9
Age-sex standardised rates
8.9
7.2
8.1
7.0
Korea
0
Female
Male
5
10
15
Age-standardised rates per 100 patients
5.4.2. Reduction in in-hospital case-fatality rates within 30 days after admission for AMI, 2003-2007 (or
nearest year)
2003
2005
2007
Age-sex standardised rates per 100 patients
12
10
8
3.7
3.4
2.9
4.7
3.9
2.9
4.9
4.2
3.2
4.8
3.7
3.3
5.3
5.2
4.2
6.0
5.2
4.5
6.9
5.7
4.5
6.3
5.2
4.7
5.8
5.2
4.9
7.7
6.6
6.1
8.1
6.6
8.3
6.6
6.5
6.2
5.1
2
3.8
4
8.8
8.5
8.1
6
0
Source: OECD HCQI Data 2009. Rates have been age-sex standardised to the 2005 OECD population (45+). 95% confidence intervals are
represented by H.
5.5.1. In-hospital case-fatality rates within 30 days after
admission for ischemic stroke , 2007
9.0
United Kingdom
Canada
7.6
Slovak Republic
7.5
Ireland
Spain
6.5
New Zealand
6.3
Czech Republic
6.2
Netherlands (2005)
5.9
Luxembourg (2006)
5.6
5.0
OECD
11.6
Sweden
3.9
Germany
3.8
Austria (2006)
3.7
Italy (2006)
3.7
Norway
3.3
3.2
Finland
2.3
Iceland
0
United Kingdom
25.5
26.0
25.2
31.0
10.7
Spain
24.2
28.2
10.8
Czech Republic
24.0
27.3
9.4
10.5
New Zealand
23.8
26.8
Canada
23.2
27.3
19.8
23.5
OECD
9.0
8.4
Iceland
19.8
Ireland
19.4
22.5
16.7
21.3
Denmark
7.0
7.3
Germany
14.5
19.7
7.4
Norway
13.7
19.9
12.8
17.2
Sweden
5.9
11.0
11.3
Korea
Age-sex standardised rates
Crude rates
10.8
13.1
Austria (2006)
5
10
15
20
Rates per 100 patients
Age-sex standardised rates
Crude rates
9.5
11.1
Finland
5.8
29.2
17.2
20.8
Italy (2006)
7.7
2.4
3.6
Korea
29.3
29.5
26.3
32.1
Netherlands (2005)
11.4
3.1
5.3
Denmark
Slovak Republic
United States (2006)
12.1
4.2
6.0
United States (2006)
30.3
32.6
Luxembourg (2006)
17.4
12.9
6.6
5.5.2. In-hospital case-fatality rates within 30 days after
admission for hemorrhagic stroke , 2007
0
10
20
30
40
Rates per 100 patients
5.5.3. In-hospital case-fatality rates within 30 days after
admission for ischemic and hemorrhagic stroke, 2007
5.5.4. Reduction in in-hospital case-fatality within 30 days
after admission for stroke, 2002-2007
Ischemic stroke
Hemorrhagic stroke
0.4
0.5
Luxembourg 1
Age-sex standardised case-fatality rates for hemorrhagic stroke (%)
16.4
New Zealand
35
R² = 0.54
LUX
30
USA
CZE
25
GBR
ESP
NZL
20
ISL
15
NOR
KOR
10
1.6
5.5
5.4
Spain
9.7
25.6
Ireland
CAN
11.6
16.5
14.1
OECD (13)
IRL
ITA
DNK
5.0
2.5
Denmark
Canada 2
SVL
NLD
1.2
6.5
Sweden
DEU
15.7
24.2
Finland
SWE
AUT
17.6
16.8
17.8
Germany 2
FIN
35.7
Netherlands 2
5
18.9
22.9
20.4
Korea 2
14.0
Austria
0
0
2
4
6
8
10
28.6
39.8
Norway
33.8
Age-sex standardised case-fatality rates for ischemic stroke (%)
0
10
20
30
40
50
% decline over period (standardised rates)
1. Based on change from 2002-2003 to 2006. 2. Based on a threeyear period only.
Source: OECD HCQI Data 2009. Rates are age-sex standardised to the 2005 OECD population (45+). 95% confidence intervals are
represented by H in the relevant charts.
Patient Safety Indicators
• Indicators based on administrative databases
• Adverse event reporting
• Safety culture
Dept Social Medicine
Indicators
– Foreign body left in during procedure (PSI 5)
– Catheter related bloodstream infections (PSI 7)
– Postoperative pulmonary embolism or deep vein thrombosis
(PSI 12)
– Postoperative sepsis (PSI 13)
– Accidental puncture and laceration (PSI 15)
– Obstetric trauma -- vaginal delivery with instrument (PSI 18)
– Obstetric trauma -- vaginal delivery without instrument (PSI
19)
24
Performance indicators in primary care
• Avoidable hospital admissions
• Lack of comprehensive administrative data-sets
Dept Social Medicine
Avoidable hospital admission rates, 2007
Asthma
COPD
Diabetic acute complications
Austria
3
United States1
CHF
Belgium
2.5
United Kingdom
Canada
2
1.5
1
Switzerland
Denmark
0.5
0
-0.5
Sweden
Finland
-1
-1.5
-2
Spain
Germany
Poland2
Iceland
Norway
Ireland
New Zealand
Netherlands3
Italy
Korea
Note: Data from Austria, Belgium, Italy, Poland, Switzerland and the United States refer to 2006. Data from the Netherlands refer to 2005.
1. Data does not fully exclude day cases. 2. Data includes transfers from other hospitals and/or other units within the same hospitals, which marginally elevate the rates. 3. Data for CHF
includes admissions for additional diagnosis codes, which marginally elevate the rate.
Source: OECD Health Care Quality Indicators Database, 2009
Patient experiences
• Service based surveys (CAHPS, Picker, CKZ)
• Population based surveys (Eurobarometer, WHO,
CWF)
• Lack of standardization
• Lack of research on validation
• Lack of research on use
Dept Social Medicine
•
•
•
•
•
•
•
•
•
Limitations National Information
Infrastructures
Mortality Statistics
Registries
Administrative Data-Bases
- secondary diagnoses
- present-at-admission coding
- unique patient identifiers
Electronic Health Records
Household and Patient Surveys
Overall: privacy and data-protection
National Information Infrastructures
• Mortality statistics
• Registries (cancer)
• Administrative
Databases
• Electronic Health
Records
• Surveys
• UPI’s/co-morbidity
• UPI’s/coding-staging
• UPI’s, present-atadmission codes,
secondary diagnoses
• Standardized secondary
data-use, privacy
concerns
• UPI’s
Hospital Level
1.2. Analysis of strategies
inter-connection
Q.I. Strategies
Ward Level
AIM
Deliveries
Appendicitis
QI Strategies
QI Strategies
QI Strategies
Outputs
Outputs
Outputs
Exploratory Factor Analysis
Strategies
- Patient Safety Systems
- TQM
- Performance Indicators
- Systems for getting Patients
Views
- Clinical guidelines
Loading
weights
.857
.822
.694
.581
Clinical
guidelines
51%
.578
Patients’
views
Health System Performance
Management
•
•
•
•
•
Whole system approach
Sub-optimization
Governance/stewardship
Incentive structure
Interconnection of strategies on performance
indicators, guidelines, safety, TQM, patient
experiences
Dept Social Medicine