Outcomes of Health Systems :
Towards the development of
indicators of amenable mortality
Work: Juan G. Gay
Presentation: Valérie Paris
OECD
October 9, 2009
Objectives of the project
• Explore the potential of “Amenable Mortality” to serve as an
indicator of health care systems outcomes in cross-country
comparisons.
• Assess the feasibility of inclusion of amenable mortality rates in
OECD Health data
Outcome indicators to measure health
systems performance in OECD statistics
•
•
•
•
Health status:
– Mortality: Life expectancy; Mortality rates by cause; Potential years of
life lost by cause; maternal and infant mortality
– Morbidity: perceived health status, low-weight births, dental health,
incidence of some communicable diseases, Incidence of some cancers,
and absence from work
Life styles and behaviours (that health systems may seek to influence): food
consumption, alcohol and tobacco consumption, body weight.
Quality of care (Health care quality indicator project):
– Disease specific survival rates, Avoidable hospital admissions,
– Patient safety, patient satisfaction and system’s responsiveness (in
development)
Equity in access to health care or in health status (?)
– see De Looper and Lafortune, OECD 2009
Health status:
what can be attributed to health systems?
• Current indicators:
– Either lack of specificity (e.g. Life expectancy can be influenced
by many other factors than health systems interventions)
– Or are too narrow to get a “global picture” of health systems
performance (e.g. mortality for a specific cause or HCQ
Indicators)
• Mortality “amenable to health care” could serve as an indicator for
global performance in improving health status
The concept of “Mortality amenable
to health care/systems”
• “Amenable deaths” = “deaths that should not occur in the presence
of timely and effective health care”
• “Amenable mortality” is measured by:
– Age-standardised mortality rates
– For selected causes of death: “Conditions for which effective
clinical interventions exist [that should prevent premature
deaths]” (Tobias and Yen 2009)
– In people under 75 years old (general age limit)
• With some adaptations for some diseases:
– E.g. only half of deaths due to ischemic heard diseases are
considered to be amenable to health care
– Age limits vary for some causes to take into account the fact
that health systems cannot be held responsible for deaths above
or below a certain age in certain disease categories
• Criteria for inclusion are “evidence-based”
Amenable / avoidable?
• Terms sometimes used as synonyms in the literature
• Eurostat – Atlas of mortality (2009)
– “Avoidable’ is loosely defined as important causes of death
which could be avoided by changing lifestyles or health policies”.
– Includes for instance deaths from road accident, suicides
– Age limit set at 65.
• The difference between the two concepts pertains to the
“boundaries of health systems”
– e.g. Prevent fatal home injuries or road accidents is not always
in the scope of MoH activities
– E.g. Suicides are not included in amenable mortality though
prevention of suicide is generally included in formulated health
policy objectives
• Usually, “avoidable mortality” includes more causes of death and a
unique age limit.
Methodology and data
Data
• Mortality databases from the WHO Statistical Information System (WHOSIS).
• From 1996 to 2006 (or last available year 2004-2005)
• Population structure: OECD population structure 1980.
• Lists of causes published by Nolte & McKee (2008) and Tobias & Yeh (2009)
• Limitations:
– Switzerland and Turkey excluded because of data limitations
– Minor modifications to the original Tobias & Yeh list were done to fit the
particular grouping of codes used in the WHO database (only in ICD9
codes):
• Exclusion of deaths from Thyroid Cancer because they are integrated
in a much larger category in the WHO database.
• Asthma included in the Chronic Obstructive Pulmonary Diseases
category (all ages <75 included).
Method
• Standardize Mortality Rates (SMR) over 100,000 people for specific causes of
death in specific age groups (<75 years).
Analyse
Analyse
• Level and trend in amenable mortality
• Analysis by gender
• Comparison of results obtained from the two lists of “amenable causes”
• Disaggregated analysis according to partition proposed by Murray & Lopez
(1996):
– Transmittable, maternal and perinatal causes of deaths.
– Non transmittable diseases.
Nolte & McKee (2008) and Tobias & Yeh (2009)
Causes of deaths
Nolte & McKee (2008)
Tobias & Yeh (2009)
Infectious diseases Tuberculosis
Intestinal Infections (other than typhoid,
diphtheria) <14
Typhoid, diphtheria, tetanus, septicaemia,
poliomyelitis, osteomyelitis
Whooping cough & measles <14
Measles – 1-14
Tuberculosis,
Selective invasive bacterial infections
(incl. malaria, meningitis, infections of the
skin)
Neoplasms
Colorectal cancer,
Melanoma of skin, nonmelanotic skin
cancer, Breast cancer
Cervical cancer and uterine cancer
Bladder cancer
Thyroid cancer
Hodgkin’s disease,
Leukaemia < 45
Benign tumours
Colorectal cancer,
Malignant neoplasm of skin
Breast cancer
Cervical cancer and uterine cancer (<45)
Neoplasm of the testis
Hodgkin’s disease,
Leukaemia < 45
Endocrine,
Thyroid disorders
nutritional and
Diabetes mellitus < 50
metabolic diseases
Diseases of the
Epilepsy
nervous system
Thyroid disorders
Diabetes (type 2) - 50% of deaths
Epilepsy
Nolte & McKee (2008) and Tobias & Yeh (2009)
Causes of deaths
Nolte & McKee (2008)
Tobias & Yeh (2009)
Diseases of the
circulatory system
Rheumatic heart diseases <45
Ischemic heart diseases – 50% of deaths
Cerebrovascular diseases
Hypertensive diseases
Rheumatic heart diseases
Ischemic heart diseases - 50% of deaths
Cerebrovascular diseases – 50% of
deaths
Diseases of the
genitor-urinary
system
Nephritis and nephrosis
Benign prostatic hyperplasia
Nephritis and nephrosis
Obstructive uropathy and prostatic
hyperplasia
Diseases of the
respiratory
system
All respiratory diseases (excl.
pneumonia/influenza) . 1-14
Pneumonia/influenza
Chronic Obstructive Pulmonary disease
>45
Asthma < 45
Diseases of the
Peptic ulcer
digestive system Appendicitis
Abdominal hernia
Cholelithiasis and cholecystitis
Peptic ulcer disease
Acute abdomen, appendicitis, intestinal
obstruction, cholecystitis / lithiasis,
pancreatitis, hernia
Perinatal mortality
Maternal deaths
Birth defect
Perinatal deaths (excluding stillbirths)
Complications of the perinatal period
Congenital cardiovascular anomalies – 1-14
External causes
Misadventures to patients during surgical
and medical care
Results (Tobias & Yeh)
Amenable Mortality all causes, both males and females.
SMR per 100,000 people, 1996 to 2006
250
200
150
100
50
0
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
AUS
AUT
BEL
CAN
CZE
DEU
DNK
ESP
FIN
FRA
GBR
GRC
HUN
IRL
ISL
ITA
JPN
KOR
LUX
MEX
NLD
NOR
NZL
POL
PRT
SVK
SWE
USA
2006
Amenable mortality - 2006 or last entry
HUN
MEX
SVK
POL
CZE
USA
PRT
NZL
DNK
GBR
FIN
IRL
DEU
CAN
NLD
AUT
NOR
AUS
KOR
ESP
LUX
SWE
ISL
GRC
ITA
FRA
JPN
0
20
40
60
80
Nolte List - 2006 or last entry
100
120
140
Tobias List - 2006 or last entry
160
180
200
Reduction in Amenable mortality between 1996
and 2006 (or last available year)
IRL
FIN
CZE
ISL
DNK
ITA
NLD
GBR
DEU
AUT
NOR
AUS
NZL
GRC
KOR
LUX
SWE
JPN
SVK
FRA
POL
HUN
CAN
PRT
ESP
USA
MEX
Nolte list
Tobias list
0
5
10
15
20
25
30
35
40
45
% Reduction in amenable mortality
Males
300
Results by gender
Amenable Mortality all causes
(Tobias & Yen)
SMR per 100,000 people, 1996 to
2006
180
160
250
140
Females
200
180
120
160
150
140
100
120
80
100
100
80
60
60
50
40
40
20
20
0
0
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
0
1996
1997
1998
1999
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
2000
2001
2002
2003
2004
2005
2006
AUS
AUT
BEL
CAN
CZE
DEU
DNK
ESP
FIN
FRA
GBR
GRC
HUN
IRL
ISL
ITA
JPN
KOR
LUX
MEX
NLD
NOR
NZL
POL
PRT
SVK
SWE
USA
Non transmissible
Results, by disease category
Amenable Mortality all causes
SMR per 100,000 people, 1996
to 2006
200
180
160
200
140
180
160
120
140
100
Transmissible
50
45
40
35
30
25
120
80
100
60
80
20
15
10
40
60
20
40
5
0
0
20
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
0
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
AUS
AUT
BEL
CAN
CZE
DEU
DNK
ESP
FIN
FRA
GBR
GRC
HUN
IRL
ISL
ITA
JPN
KOR
LUX
MEX
NLD
NOR
NZL
POL
PRT
SVK
SWE
USA
Main findings
•
Clear declining tendency in amenable mortality.
•
Differences among best performing countries have drastically decreased in
the last ten year.
•
France, Japan, Sweden and Island have constantly demonstrated better
results throughout the last decade.
•
Eastern European countries and Mexico have systematically performed less
good than the rest of OECD counties.
•
US is performing significantly below the rest of OECD countries excluding
Mexico and Eastern Europe.
– Surpassed by counties like Finland, New Zealand, Great Britain, Ireland and
Denmark that were experiencing higher Amenable Mortality in 1996
– No general trend of reduction on Amenable Mortality was identified in the US
since 1999 (except for non-transmissible diseases).
•
New Zealand, Denmark and Great Britain have also experienced a constant
decrease in amenable mortality, yet a gap between them and the rest of
best performing OECD countries persist.
Limitations
• Dissimilar diagnostic practices of death certification and use of ICD
codes across countries.
• Definition of the causes of death that can be considered amenable
to health care is expected to vary over time
• Definition of age limits is expected to vary over time
• By definition, AM does not take into account:
– Improvements in survival that do not allow people to go beyond 75 years
(AIDS/VIH?)
– Improvements in the quality of life: Is not an appropriate indicator to assess
the performances of health care services, whose primary intend is to
improve the quality of life, with low impact on mortality. E.g.: Mental care
is virtually not taken into account
• Lists of causes of deaths amenable to health care have been
modest in taking into account deaths that could be avoided from
changes in life-styles (abusive consumption of tobacco or alcohol)
Comparisons with other outcome indicators
• Life-expectancy
– Takes into account all causes of death
• Potential years of life lost
– Include all causes of mortality, including external causes (road
accidents, suicide, falls, etc…)
– Age limit: 70 years (for all causes)
– Is the some of all years lost between age of death and 70 years
(death at 50 « weights » half less than death at 30, which is not
the case in amenable mortality).
Life expectancy and amenable mortality
Amenable mortality (2006 or last entry)
200
180
HUN
160
MEX
SVK
140
120
POL
CZE
USA
PRTBEL
NZL
DNK UK
FIN
IRL
DEU
CAN
NLD
AUTNOR
AUS
ESP
LUX
KOR GRC SWE
ISL
ITA
FRA
100
80
Amenable mortality contributes
to general mortality by 10%
(France) to 18% (Mexico,
Hungary)
60
40
20
JPN
y = -12.502x + 1075.4
0
70
72
74
76
78
80
Life Expectancy (2006 or last entry)
82
84
PYLL and amenable mortality
Amenable mortality (2006 or last entry)
200
180
HUN
y = 0.0251x - 7.2491
160
SVK
140
120
POL
CZE
100
UK
NZL
NOR DNK
NLD IRL
FIN
DEU
CAN
AUT
ESPAUS
LUX
SWE
KOR
ISL
GRC
ITA
FRA
JPN
80
60
40
MEX
USA
PRT
20
0
0
1000
2000
3000
4000
5000
PYLL (2006 or last entry)
6000
7000
8000
Conclusions
• Amenable Mortality is a practical and effective indicator that could
be useful in the comparison of the performance of health care
systems across OECD countries.
• AM offers the potential to go further in the identification potential
weaknesses of health systems (by categories of diseases)
• Inclusion in OECD health data requires the choice of a list
• AM is only an indicator of outcome. It should be related to resources
invested in health care to really assess health systems performance
(efficiency)