Disability free life expectancy (DFLE)
in the European Union from 1995 to 2003
using the European Community Household Panel (ECHP)
Aurore Clavel
Presentation plan
•
The project European Health Expectancy Monitoring Unit
(EHEMU)
•
Method and results of calculation for the first year
Euro-REVES: the foundation of EHEMU
•
The sustained interest in disability-free life expectancy in each country
led to a European research programme identifying reasons for the
incomparability of European results (Biomed II, 1995-1997)
•
From this point, the development of a coherent set of health
expectancies was proposed through the Health Monitoring
Programme (1997-2002)
•
The current move to standardised surveys in Europe (ECHP,
Survey on Income and Living Conditions: SILC, and the future
European Health Interview Survey) allows this development through
EHEMU
Euro-reves: A vision for Europe
Aim of European Health Expectancy
Monitoring Unit (EHEMU)
• To provide annual comparable health expectancy estimates for
all European Union countries, in association with Eurostat
• To analyse and interpret the results
• To educate the policy makers, the politicians and the public in health
expectancy as an indicator of population health
EHEMU team
From left to right :
• Emmanuelle Cambois : INED, Paris
• Carol Jagger : University of Leicester
• Aurore Clavel : Montpellier
• Herman Van Oyen : IPH, Brussels
• Geraldine Barker : University of Leicester
• Jean-Marie Robine : INSERM, Montpellier
• Isabelle Romieu : Montpellier
EHEMU workplan for Year 1
• Calculation
-Trends in disability-free life expectancy using ECHP 1994-2001
data with extrapolation for 2002-3 in relation to the new
structural indicator Healthy Life Years (HLY)
- Interrelationships between different health dimensions
using Eurobarometer 2002
• Repository
- EHEMU-calculated values
- Available information on other studies calculating Health Expectancy
• Education
- Computation manual with extension for confidence intervals
• Extension of the network
- Identifying EHEMU partners in all members states
• Dissemination
- Conception and development of EHEMU website
- Country reports
(policy and technical)
Data and Methods (1)
• Estimation of Life Expectancy (LE) and 95% CI
• Estimation of DFLE and 95% CI, using Sullivan method
 age specific probability of death
 age specific disability prevalence
• Question used
PH002 “Do you have chronic physical or mental health problem,
illness or disability?” and if Yes :
PH003 “Are you hampered in your daily activities by this physical
or mental health problem, illness or disability?”
Yes, severely
Yes, all levels
Yes, to some extend
No
Data and Methods (2)
Problems
1) Mortality and Panel rough data
• Probable data errors
• Missing data
2) Interruption of data collection
Solutions
1) Mortality and Panel rough data
• Linear imputation of age specific
probabilities according to trends
Example…
• Imputation of age specific probabilities according to observed trends
11
Expected year
10
9
8
7
6
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Year
11
10
Expected year
Example:
Female Life Expectancy
(LE) at age 65 in UK
9
8
7
6
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Year
Data and Methods (2)
Problems
1) Mortality and Panel rough data
• Probable data errors
• Missing data
Solutions
1) Mortality and Panel rough data
• Linear imputation of age specific
probabilities according to trends
Example…
• Imputation of data according to
observed trends
Example…
2) Interruption of data collection
• Linear imputation of missing probabilities of death
83
82
Expected year
81
80
79
78
77
1994
1995
1996
1997
1998
1999
2000
2001
2002
Year
83
82
81
Expected year
Example:
Female LE at birth
in Denmark
80
79
78
77
1994
1995
1996
1997
1998
Year
1999
2000
2001
2002
Data and Methods (2)
Solutions
Problems
1) Mortality and Panel rough data
• Probable data errors
• Missing data
1) Mortality and Panel rough data
• Linear imputation of age pecific
probabilities (death and disability)
• Imputation of data according to
observed trends
2) Interruption of data collection
No data for 2002 and 2003
2) Interruption of data collection
• Linear extrapolation of the disability
prevalence
Example…
• Linear extrapolation of the disability prevalence up to 2003
Ex: Male disability prevalence in Greece (65 years and older)
0,8%
0,7%
Disabilityprevalence
0,6%
0,5%
0,4%
0,3%
0,2%
0,1%
0,0%
1994
1995
1996
1997
1998
1999
Year
2000
2001
2002
2003
2004
Main results
Example: Trends in LE and DFLE at age 65, 1995-2003, Male, Austria
22
Expected year
20
LE
18
16
14
12
10
8
DFLE
6
1995
1996
1997
1998
1999
Year
2000
2001
2002
2003
Trends in LE and DFLE at birth in European countries, 1995-2003
Male
84
82
80
78
76
LE
Expected year
74
72
70
68
66
DFLE
64
62
60
58
56
54
52
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Year
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherland
Portugal
Spain
Sweden
United Kingdom
Trends in LE and DFLE at birth in European countries, 1995-2003
Female
84
82
LE
80
78
76
74
Expected year
72
70
68
66
DFLE
64
62
60
58
56
54
52
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Year
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherland
Portugal
Spain
Sweden
United Kingdom
Distribution of life and disability free life expectancy EU (14), 2003
85.00
LE
DFLE
80.00
SWE
75.00
70.00
65.00
60.00
55.00
Males
Females
sex
Trends in expected life free of disability at age 65, 1995-2003
Women
16
AUT, BEL,
DNK, ITA,
ESP, SWE
14
Expected year
12
FIN, FRA,
GBR
10
8
DEU,
GRC, IRL,
NDL, PRT
6
4
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Year
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherland
Portugal
Spain
Sweden
United Kingdom
Trends in expected life free of disability at age 65, 1995-2003
Men
13
AUT, BEL,
FIN, FRA,
DEU, IRL,
ITA, ESP
12
11
Expected year
10
GRC, NDL,
PRT, GBR
9
8
DNK,
SWE
7
6
5
4
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Year
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherland
Portugal
Spain
Sweden
United Kingdom
Trends in DFLE using the ECHP
Some conclusions
1) Life expectancy:
• Small variation in life Expectancy between these 14 members states
• increase between 1995 and 2003
2) Disability Free Life Expectancy
• Large variation between these 14 members states
• Diverging trends between 1995 and 2003:
reduction / stagnation / increase in the expected life with reported
disability while LE increases
3) Gender differences in DFLE trends in some countries
4) Gender differences in DFLE are smaller than gender differences in LE
Trends in DFLE using the ECHP
Issues and inference
 Important differences in reported disability in
the 14 european populations:
- different levels of reported disability (larger dispersion than LE)
- variation in the magnitude of the gender difference
- different trends over time
 A more elaborate analysis would include :
- a cross between national data and European values to improve
harmonisation of the instruments
- the use of different levels of severity (SILC)
…