Academy Health Annual Research Meeting
Chicago, June 29th 2009
Application of Patient Safety
Indicators in 16 OECD member
countries using administrative
hospital data
Saskia E Drösler, MD (*)
NS Klazinga, MD, PhD (#); DJ Tancredi, PhD (+); PS Romano, MD, MPH(+)
* Niederrhein University of Applied Sciences, Krefeld, Germany
# Academic Medical Center, University of Amsterdam, The Netherlands
+UC Davis Center for Healthcare Policy and Research, Sacramento, CA, USA
Background (1)
Health Care Quality Indicator Project of the
Organization for Economic Co-Operation and
Development (OECD) - topics
Care for chronic conditions
Care for mental disorders
Cancer Care
Care for communicable diseases
Patient Safety
 Data linkage to information on economics and resources
 Mutual learning among participating countries
 Information on health system performance for policymakers
30.06.2009
S. Drösler
2
Background (2)




Pilot in 2007 (7 countries, 12 Patient Safety Indicators)
2008: 15 Patient Safety Indicators
Calculation of rates is the responsibility of the countries
Research and development project – no public reporting
 Volunteering countries:
Belgium
Canada
Denmark
Finland
France
Germany
30.06.2009






Italy
The Netherlands
New Zealand
Norway
Singapore
Spain
S. Drösler
 Sweden
 Portugal
 UK (data from England)
 USA
3
Research Questions
 Can AHRQ PSI definitions be applied on hospital
data from an extended number of countries?
 What are the effects of internal standardization?
 Which factors compromise comparability across
countries?
 What is the test-retest reliability (6 countries, 11
indicators)?
30.06.2009
S. Drösler
4
Data Collection





Numerator and denominator populations
Age / sex stratifications
Average length of stay
Mean numbers of secondary diagnoses
Details on medical classifications and coding
rules
 Details on electronic data systems
30.06.2009
S. Drösler
5
Results – Calculation Samples
Country
Year of
Study
data
Population
collection (in millions)
Estimated
Average
ICD
annual
length of Revision in
hospital
stay (calc
use
discharges
data)
(in millions)
1,78
8
9
Belgium
2004
1,51
Canada
FY 2006/07
1,93
2,88
6,4
10
Denmark
2007
1,01
0,9
4,2
10
Finland
2006
1,21
1,07
5,4
10
France
2006
18,91
16,95
3,6
10
Germany
2006
14.2 #
16,67
7,48
10
Italy
2005
7,31
8,41
6,7
9
Netherlands
2004
2,23
1,71
4,31
9
Norway
2006
1
0,82
3,7
10
FY 2006/07
0,68
0,59
3,4
10
Portugal
2006
0,9
0,9
7,1
9
Singapore
2007
0,45
0,43
5
9
Spain
2006
3,19
4,85
7,59
9
2006
FY 2006/07
2006
1,25
11,82
32.98 #
1,50
7,75
33,51
6,05
2,92 4,6
10
10
9
New Zealand
Sweden
United Kingdom
United States
# Estimations from samples, 10% in Germany, 20% overall in the USA
30.06.2009
S. Drösler
6
Results – Variations in PSI rates
Ratio between highest and lowest PSI rate
14
99,27
Decubitis ulcer
15
75,32
Foreign body left in during procedure
16
6,29
Iatrogenic Pneumothorax
10 332,69
Infection due to Medical Care
16
93,74
Postoperative hip fracture
13
13,29
Postoperative respiratory failure
15
644,57
Postoperative pulmonary embolism (PE)
or deep vein thrombosis (DVT)
Postoperative sepsis
16
42,62
16
4950,45
Technical difficulty with procedure
16
29,56
Transfusion reaction
15
76
Birth trauma - injury to neonate
15
111,02
16
300,31
16
17,51
15
135,25
PSI
Complications of anaesthesia
Obstetric trauma - vaginal delivery with
instrument
Obstetric trauma - vaginal delivery
without instrument
Obstetric trauma - caesarean section
30.06.2009
S. Drösler
Elimination of four countries
N (countries)
Ratio - 16
countries
N (countries)
Ratio - 12
countries
10
99,25
11
17,67
12
4,42
8
136,25
12
74,84
10
12,41
11
53,38
12
9,7
12
45,06
12
29,67
11
18,34
11
23,51
12
11,13
12
11,63
11
135,25
7
Crude and stand. rates– postop. PE / DVT
1,17
1,18
10
0,42
0,41
13
0,41
0,39
5
17
0,37
0,31
0,31
6
0,29
0,30
Country ID
14
0,28
0,28
12
0,27
0,27
8
0,24
0,25
9
0,21
0,20
7
0,19
0,19
4
Age-/sex-stand. rate per 100 patients
0,14
0,14
2
Crude rate per 100 patients
0,12
0,12
11
0,08
0,07
16
0,03
0,03
3
0
30.06.2009
0,43
postop PE / DVT rate (%)
0,2
0,4
0,6
S. Drösler
0,8
1
1,2
8
Female and male rates – postop. PE / DVT
1,12
10
13
0,32
14
0,39
0,19
0,31
0,31
6
Country ID
0,41
0,40
0,38
5
0,27
0,26
8
0,27
0,28
12
0,26
0,24
9
4
0,17
0,21
0,19
0,21
7
male
0,14
0,15
2
female
0,12
0,13
11
0,07
0,08
16
3
0,45
0,37
17
1,25
postop PE / DVT rate (%)
0,03
0,02
0,00
30.06.2009
0,20
0,40
0,60
S. Drösler
0,80
1,00
1,20
1,40
9
Age categories – postop. PE / DVT
14
10
17
13
6
Country ID
4
8
7
12
5
9
2
18 - 39
40 - 64
65 - 84
>85
11
16
postop. PE / DVT rate (%)
3
0
30.06.2009
0,2
0,4
0,6
0,8
1
S. Drösler
1,2
1,4
1,6
1,8
2
10
Dependency between rates and mean number
of secondary diagnoses – postop. PE / DVT
Mean number of sec. diagnoses
6
5
4
y = 3,4475x + 1,4543
R2 = 0,4504
10
14
9
12
8
3
2
5
4
2
17
13
11
6
1
7
16
3
standardized postop PE / DVT rates (%)
0
0
30.06.2009
0,2
0,4
0,6
S. Drösler
0,8
1
1,2
1,4
11
Test-Retest-Reliability (2007 / 2008 data)
 Pearson's correlations on country level
Country
Canada
Germany
Spain
Sweden
United Kingdom
USA
n (number of PSI rates)
11
8
11
11
11
11
Pearson's cc (p≤ 0,01)
0,99
0,94
0,98
0,82
0,998
0,998
 Pearson's correlations on indicator level (n= 6 countries)
PSI Title
Foreign Body Left In During Procedure
Selected Infections Due to Medical Care
Postoperative Hip Fracture
Postoperative PE / DVT
Postoperative Sepsis
Accidental Puncture or Laceration
30.06.2009
S. Drösler
Pearson's cc (p≤ 0,01, #p≤ 0,05)
0,98
0,92
0,86 #
0,92
0,995
0,99
12
Conclusions
 Method is applicable on international data of many
countries
 Significant variations in rates
 Amount of clinical coding has impact on rates
 Age and gender populations are comparable
among countries
 Test-retest reliability is high on indicator and
country level
30.06.2009
S. Drösler
13
Implications for Policy
 More political support is needed in many countries
 Enhancement of databases
 Incentives for clinical coding
 Coding rules
 Introduction of an uniform patient identifier
 Introduction of a dx type indicator: present at
admission
 Comparative reporting of Patient Safety Indicators is planned
in the future
 Ongoing validation studies
 Further clarification of variances
30.06.2009
S. Drösler
14
Acknowledgments
• This investigation was supported by the
Organization for Economic Co-Operation and
Development (OECD).
• The countries of Belgium, Canada, Denmark,
Finland, France, Germany, Italy, the
Netherlands, New Zealand, Norway, Portugal,
Singapore, Spain, Sweden, the United Kingdom,
and the United States of America supported this
research by performing the calculations and
providing the data.
30.06.2009
S. Drösler
15