The social context of prevention

1
The social context of prevention

• Finn Diderichsen MD PhD
• University of Copenhagen
• Institute of Public Health
• February 3, 2004

2
What is prevention ?

• It is what interacts with wealth to produce
  health

3
Transforming wealth into health. GDP and Life
expectancy 1960-99
Lifeexpectancy
GDP in USD
Source World Health Chart www.whc.ki.se
4
Differential effectivenes of health policy ?
5.9 years per 10.000 (SE)
Life expectancy
6.2 years per 10.000 (UK)
3,3 years per 10.000 (DK)
GDP in USD
5
The impact of public policies on health
development. Improvements in life expectancy in
UK over six decades 1901-60 (Preston et al 1992)
Years
6
Denmark still lags behind in health development
but compared to Russia

• Life
• expectancy
• 1970-2001

7

• Mortality in heart disease, cancer
• and injuries.
• Denmark
• and EU
• 1970-2000
• Age standardized per 100.000

8

• Knowledge and money is not enough.
• Structure, culture and policy modifies the effect
  
• on public health.

9
What are the public health priorities ?

• Is it about eternal life and happiness?
• Is it about peoples actual freedom and capability
  to live the life they value ?
• (Amartya Sen Development as Freedom
  1999).
• Disability, handicap and early death as the major
  obstacles to that freedom.

10
Measuring disability and early death Burden of
disease measurement (WHO)

• FOR EACH DISEASE
• DALY Years of Life Lost Years Lived with
  Disability
• number of deaths life expectancy at age of
  death (YLL)
• incidence duration disability weight
  (YLD)

11
Burden of disease in EU and the World 2000 DALY
per 100.000 (WHO 2003)
12
Years lost in disability vs. early death
DALY per 100.000
13
What disease groups generate the gap across
socio-economic groupsThe six most important in
Sweden Slope index of inequality -DALYs per
100.000 (Ljung, Diderichsen 2004)
14
Burden of disease measurement tells us that

• The poor countries are suffering equally from
  the diseases of affluence in addition to the
  diseases of poverty
• The increasingly important role of disability
  vs. mortality (now 53 of DALYs in Europe)
• The increasingly dominating role of mental
  health problems, also in the health divide

15
Theory of prevention

• Primary prevention acts on causes of diseases and
  injuries.
• Without knowledge on cuases
• no prevention.
• Therefore
• Theory of causation is crucial.

16
The importance of causal thinking

• Prevention is about causes of disease
• and about causes of causes
• One disease has many causes
• Causes act in chains and mediates the effect of
  each other
• Causes interact with each other

17
Distal and proximal causesMurray and Ezzati 2003
18
The causal web of CHDMurray and Ezzati 2003
19
Mediation the causal chains where the occurence
of one cause depends on another

• The effect of social position on disease might be
  mediated through working conditions,
  psychosocial factors and health behaviour
• All are causes eventuelly to be intervened upon.

Social position
Skill develop-ment at work
Self-efficacy
Health behaviour
Health outcome
20
Causal thinking.

• Most of those exposed to a cause do not get ill.
  Something else is needed for illness to occur
• Therefore causes are never independent
• they interact

21
Interaction

• Five contributing causes A-F constitutes one
  sufficient cause.
• If there exists cases caused by A B in
  combination, A and B interacts and their effects
  will be synergistic (non-additive)

22
Interaction when the effect of one cause
depends on another

• The effect of smoking depends on something else
• Interaction as a source of ecological bias

23
The health divide

• Growing educational inequality in mortality
• Total mortality per 1000.
• Men and women 30-59 years
• Denmark 1981-95
• (Mackenbach, Kunst et al 2003)

24
The social clustering of riskfactors SUSY. NIPH
2000
25
Social groups are differentially susceptible to
the effects of smoking SHEEP-study (Hallqvist,
Diderichsen et al 1998)
Interaction Synergy index (3.3-1)/((1.1-1)(1
.9-1))2.3 (95CI 1.2-4.5)
26
Interaction Occupation alcoholKohort of
conscripts N49.323Hemmingsson, Diderichsen et
al. Soc Sci Med 1998
Synergy index 2.26 (95 CI 1.21-4.53)
27
The relevance of interaction

• Some causes are easier to intervene against
  than others (e.g. genes, social position).
• When two causes interact we can remove some of
  the effect of one (less amenable) cause by
  intervening against the other.
• Targeting clinical advice to the most susceptible
  individuals

28

• What determines the choice of cause in a chain to
  intervene against ?
• Whether it is distal or proximal to the outcome?
• The technical and political opportunities for a
  change ?
• The impact on health (average and distribution) ?

29
The potentials of prevention

• Estimating the burden attributable to past
  exposure
• Estimating the burden avoidable by future reduced
  exposure

30
Attributable and avoidable burden
Mathers et al 2003
31

• Burden of disease attributable to 10 major
  riskfactors
• EU 2000
• Percent ()
• of DALYs
• (Ezzati et al. Lancet 2002)

of DALYs
32
Limitations of the WHR-2002 -table

• No social and economic causes
• More vague symptoms (mental, allergic and
  musculoskeletal) are not included
• When epidemiological evidence is scarce effects
  are omitted
• Attributable fraction is strongly sensitive to
  the estimated size of the effect.

33
The cost-effectiveness of prevention

• Several policies and interventions are evaluated
  in terms of both effects and costs.
• A few examples

34
Tobacco control cost and effect WHO Lancet 2003
35
Other examples of prevention cost and effect
WHO-CHOICE 2003
36

• 5.0 USD is 0.2 of the Danish health care budget
  in 2001 !

37
EU Common Agricultural Policy What prevention
has to work against!

• CAP spending per year (45 bill euro)
• including
• Destruction of quality fruit and vegetables 117
  mill euro
• Subsidies to fatty dairy products 1.6
  bill euro
• Subsidies to tobacco production 965
  mill euro

38
PATHWAYS TO HEALTH INEQUALITY 5 concepts
(1-5) and 5 policy entry-points (A-E)
5.Context Policy
A
1. Social position
B
2. Specific cause
C
E
3. Disease or injury
D
4. Social consequences
39
1st entrypoint A Social stratification. B
Differential exposure. C Differential
susceptibility. D Differential consequences
E Generating social context
Context Policy
A
Social position
B
Specific cause
C
E
Disease or injury
D
Social consequences
40
Modifying social stratification- National Swedish
targets

• Target 1.3 Allocating resources to local schools
  to compensate effects of segregation
• Target 4.1 Ensure possibilities for life long
  learning, adult education, vocational training
  etc.

41
2nd entrypoint A Social stratification. B
Differential exposure. C Differential
susceptibility. D Differential consequences
Context Policy
A
Social position
B
Specific cause
C
Disease or injury
D
Social consequences
42
Tackling differential exposure National Swedish
targets

• Target 1.1 Reduced poverty for families with
  children with family policies
• Target 12 Reduced tobacco smoking among
  low-income groups by tax-policy and brief advice
  by GPs

43
3rd entrypoint A Social stratification. B
Differential exposure. C Differential
susceptibility. D Differential consequences
Context Policy
A
Social position
B
Specific cause
C
Disease or injury
D
Social consequences
44
Differential susceptibility

• The clustering of social and behavioral causes
  among underprivileged groups
• Interaction between causes in the same pathway
  increases the vulnerability to single causes

45
Tackling differential susceptibility Swedish
targets

• Target 5.2 Increasing control over work as
  demands are growing (will decrease susceptibility
  to effects of high demand)
• Target 3.1 Family policies that provide
  preconditions for close emotional attachment
  between infants and parents.

46
Lessons (1)

• We have several epidemiological instruments for
  setting priorities and realistic targets
• Knowledge alone does not transform wealth into
  health. Prevention out of context does not work.
  
• For some diseases with large burdens, causes are
  largely unknown (e.g. depression)

47
Lessons (2)

• We have to adapt to a future of prevention with
  many causes with small individual effects
• But small effects may imply large public health
  effects when exposure is prevalent
• Effects will often be vague symptoms rather than
  mortal diseases
• Even the strong ones (tobacco, asbestos, alcohol,
  physical inactivity) are still (increasingly)
  prevalent
• So information to individuals about individuals
  risk will be even more insufficient in the future

48
Lessons (3)

• The growing knowledge of proximal biological
  (genetic) causes make our causal knowledge more
  precise
• An effective prevention has however to be based
  on a knowledge on the (more distal) context where
  exposures are generated and mediated
• Our epidemiological knowledge about structure,
  culture and policy as causes is very limited.
• Our designs and methodologies to evaluate these
  effects can be greatly improved

49
Lessons (4)

• We need to improve our theories and methods for
  dealing with complex causal structures such as
  mediation, interaction
• We need to study populations as systems where
  people influence each other rather than (as now)
  as heaps of isolated individuals.

50
Do we need a population laboratory in with BBH as
a medical ressource center ?

• Where we closely can study the mechanisms in the
  social etiology of disease
• Where the indivudal and contextual consequences
  of disease can be studied
• Where preventive policies and interventions can
  be implemented and evaluated
• Where the contextual prerequisites of prevention
  can be evalutaed and modified
• Where theory and methods in public health
  training can meet society and population
• Where modern epidemiology and health economy can
  be applied in health planning and ressource
  allocation