Social Inequality, Psychosocial Factors and CHD:

1
Social Inequality, Psychosocial Factors and
CHD What Have We Learned from the Research in
Rich Countries? John Lynch Department of
Epidemiology, Biostatistics and Occupational
Health McGill University, Montreal,
Canada PURE - Dubai, January 2006
2
What have we learned from social epidemiology?
The Social Gradient in CHD
CHD Rate / 100K
Social Advantage
Social Disadvantage
3
The First Whitehall Study
4.0
3.2
RR
2.1
Rose and Marmot. Br Heart J (1981)
4
The social class difference was partly explained
by known coronary risk factors men in the lower
grades smoked more and exercised less, they were
shorter and more overweight, and they had higher
blood pressures and lower levels of glucose
tolerance. Most of the difference, however,
remains unexplained. It seems that there are
major risk factors yet to be identified,
Rose and
Marmot (1981, p.13)
5
if these aspects of lifestyle account for less
than a third of the social gradient in mortality,
what accounts for the other two thirds? The
second question occupies the rest of this book.
(p. 45 of 320)
Marmot (2004)
6
Implications (1)

• There must be other mechanisms at the individual
  and contextual level that generate social
  gradients in CHD such as stress, job control,
  autonomy, social participation, neighborhood
  effects, social capital, etc.
• i.e., a range of psychosocial factors

7
Psychosocial factors in social epidemiology
Stress Sense of Coherence Hardiness Optimism John
Henryism Job Strain Ways of Coping Anxiety Self
esteem Resilience Social cohesion Social capital
Type A Behavior Pattern Cynical
Distrust Anger/Anger-in/Anger-out/ Hostility Socia
l Isolation Social Support Control Sense of
Control Mastery Hopelessness Depression
? CHD
8
Implications (2)
2. Interventions focused on health behaviors and
conventional risk factors are unlikely to
appreciably reduce the social gradient in health.
For instance, a 1998 JAMA study that has been
cited more than 200 times (Web of Science),
stated that, Thus, public health policies and
interventions that exclusively focus on
individual risk behaviors have limited potential
for reducing socioeconomic disparities in
mortality. (p. 1707)
Behaviours do not matter
Behaviours matter but they are just not the big
story
The downplaying of behaviours and their
physiological correlates in understanding social
inequalities in CHD influences research,
funding, interventions and policy.
9
Social Inequality
Traditional Risk Factors
CHD
INTERHEART PAR 75
10
What have we learned from CHD epidemiology?

• Conventional risk factors smoking,
  hypertension, dyslipidemia and diabetes do
  explain most CHD in populations
• INTERHEART study (2004) show PAR for 4
  conventional risk factors is 76
• Emberson, Whincup, et al (2005) in BRHS show
  smoking, blood pressure and cholesterol account
  for 81 PAR (adjusted for regression dilution
  bias)
• Stamler (1992,1999) and Greenland (2003) show
  that 75-100 of all CHD cases occur among those
  exposed to at least one conventional risk factor

11
A Paradox?
So how can it be that the factors accounting for
most CHD in a population do not seem to account
for most of the social gradient in CHD, when the
social gradient simply emerges from sub-grouping
the population according to some indicator of
social position?
12
A Paradox?
1000 cases of CHD
Most educated 100
Middle educated 300
Least educated 600
13
An Illustration

• Population sample of 2,682 Finnish men Kuopio
  IHD Study
• Stratified the population into lower and higher
  risk groups based on standard clinical
  definitions - current smoking, hypertension,
  dyslipidemia and prevalent diabetes
• 34.7 were current smokers, 58.7 had
  hypertension, 42.7 had dyslipidemia, 6.5 had
  diabetes, and 84.9 had at least one of these
  risk factors
• Fatal (ICD9 codes 410-414) and nonfatal CHD,
  classified according to MONICA criteriia
• There were 425 CHD events (108 fatal, 317
  non-fatal) during an average follow-up of 10.5
  years.

Lynch, Davey Smith, Harper, Bainbridge. J
Epidemiol Community Health (in press)
14
Accounting for Cases of CHD in this Population

• Of the 425 total cases - 402 (94.6) occurred
  among men exposed to at least one of the four
  conventional risk factors
• 70 of cases occurred in men with at least 2
  risk factors
• PAR 70
• Similar to INTERHEART (PAR 75) and with US
  cohorts where 90 of cases occur among those
  exposed to at least one conventional risk factor
• Thus these results are consistent with the idea
  that
• conventional risk factors DO EXPLAIN the vast
  majority of CHD cases

15
Accounting for the Social Gradient in CHD in this
Population

• Social gradient in CHD comparing high to low
  education is RR 1.90
• After adjustment for conventional risk factors
  RR 1.68
• This corresponds to a 24 reduction in the
  excess RR
• Thus these results are consistent with the idea
  that
• conventional risk factors DO NOT EXPLAIN the
  social gradient in CHD

16
What would population levels and social gradients
in CHD look like if there were no conventional
risk factors in this population?
17
Whole KIHD Population N 2,682
Low Risk Segment of KIHD Population N 404
(15.1)
18
Excess Risk (whole pop) 88 per 1000 Excess
Risk (low risk pop) 25 per 1000 Make it a
low risk pop 72 reduction in the excess risk
of social inequality in CHD
194
RR 1.8
143
CHD Risk / 1000
106
66
60
RR 1.6
41
Low risk population
Low risk population
Low risk population
19

• In a low risk population with no smoking,
  hypertension, dyslipidemia and diabetes
• Relative social gradient in CHD remains
• RR 1.8whole vs 1.6low risk
• Absolute social gradient is reduced by 72
• ER 88whole vs 25low risk per 1,000

20
Which social inequality is better?
Original social inequality in CHD
Give the disadvantaged the same risk factor
levels as the most advantaged
CHD Rate / 100K
If we intervened to substantially reduce risk
factors in all social groups
Social Disadvantage
Social Advantage
21

• Within an absolute risk framework, there is no
  paradox between the observations from CHD and
  social epidemiology. Conventional risk factors do
  account for most CHD and for most of the absolute
  social gradient in CHD (72 of the excess risk).
• However, when explaining relative social
  gradients in CHD, the apparent paradox may arise
  that the factors which explain most cases of CHD
  do not explain the relative social gradient.
• Adjustment for conventional risk factors only
  reduced the relative social gradient in CHD by
  24. This is normally interpreted to mean that
  most of the effects of social inequality on CHD
  do not work through mechanisms linked to
  conventional risk factors.

22
What does it take to be a confounder?

• The extent to which a 3rd variable reduces the
  RR in an exposure-outcome association (an
  indication of confounding of the association)
  depends on the relative distributions of the
  exposure over strata of the 3rd variable and the
  strength of its association with the outcome.
• In this case there is a more extreme relative
  distribution of education over strata of
  hopelessness than over strata of the conventional
  risk factors (partly because the prevalence of
  conventional risk factors is high), and so
  hopelessness/depression appears to be a stronger
  confounder (in this case interpreted as a
  mechanism) of the association between education
  and CHD.

23
Depression
Education
? 28
Social Gradient RR1.9
? 24
Conventional Risk Factors
CHD
on the basis of the adjusted RR analysis we would
normally interpret hopelessness/depression as an
important mechanism in the causation of the
social gradient in CHD
24
How important are these 2 mechanisms?

• intervene to eliminate hopelessness / depression
• ? reduces the relative social gradient 28
• ? eliminates 14 of CHD cases (confounded
  estimate)
• intervene to eliminate smoking, hypertension,
  dyslipidemia and diabetes
• ? smaller reduction in the relative social
  gradient 24
• ? eliminate 90 of CHD cases and 72 of the
  absolute social gradient

25

• 72 reduction in ER is an underestimate because
  low risk population is not really low risk
• No differences in blood pressure and have higher
  BMI
• but low risk population has ex-smokers, higher
  cholesterol and LDL, and are 5 cm shorter

26

• Deciding on the importance of various risk
  factors for elucidating the mechanisms behind
  social gradients in CHD cannot be done on the
  basis of relative comparisons alone.
• As Geoffrey Rose commented
• Relative risk is not what decision-taking
  requires relative risk is only for researchers
  decisions call for absolute measures.
  
  
  
  p. 19, Strategy of Preventive Medicine.

27
Black White Inequality in Infant Mortality
over the 20th Century, USA
Relative Inequality
Black
White
28
Age-specific Mortality Differences between the
Richest and Poorest 20 of the Worlds Population
Poorest 20
RR
Mortality Rate per 1000
RR
Richest 20
Gwatkin (2000)
29

• Explaining and reducing relative social
  gradients in CHD is a legitimate and important
  focus of research and intervention because it can
  help reveal novel CHD risk factors and mechanisms
  that are unevenly distributed across social
  groups.
• However, explanations for relative social
  gradients need to be understood within the
  context of what causes most cases of CHD.
  Otherwise they may deflect attention from the
  most important population-level causes.
• Rose if everyone in the population smoked we
  would be finding that radon or asbestos were the
  most important causes of lung cancer

30

• In populations where the prevalence of
  conventional risk factors is high, it is possible
  that there are no or small social gradients in
  those risk factors such that they cannot account
  for relative CHD differences across social groups
  but contribute substantially to the absolute risk
  of CHD in all social groups.
• In the Whitehall Study there are small social
  differences in blood pressure and no social
  differences in total cholesterol
• So reducing major, high prevalence risk factors
  from the population may have little effect on the
  relative social gradient but a large effect on
  the absolute social gradient as indicated by a
  reduction in the excess risk.

31

• Importantly, this also means that whatever
  proximal (CRP) and distal factors (social
  capital) are identified as causes of the relative
  social gradient in CHD, if their behavioral and
  biological mechanisms do not involve conventional
  risk factors then they probably account for a
  small proportion of CHD cases.

32
Social Inequality
Control Depression Stress
?
Traditional Risk Factors
CHD
INTERHEART PAR 80
33
Income Inequality, Poverty and Heart Disease
1900-2000, USA
poverty
income inequality
heart disease
34
Race-specific Voting Participation in
Presidential Elections and age-adjusted,
all-cause mortality, USA, 1968 - 1998
White Voting
Black Mortality
All-cause mortality per 100,000
Voting in Presidential elections
White Mortality
Black Voting
Year
Lynch and Davey Smith (2003)
35
Sex-Specific Smoking Trends and Heart Disease,
1900-1998, USA
Male Heart Disease
Per Capita Consumption
AADR
Male Smoking
Female Heart Disease
Female Smoking
36
Sex-Specific Smoking Trends and Income
Inequality, 1900-1998, USA
Male Smoking
Income Inequality
Female Smoking
Gini
Per Capita Consumption
37
Sex-Specific Cholesterol Trends and Income
Inequality, 1980-2002, USA
Male Cholesterol
Gini
Female Cholesterol
Total Cholesterol Trends from Minnesota Heart
Survey Arnett et al. Circulation (2005)
38
Conclusion

• We should re-affirm that smoking, hypertension,
  dyslipidemia and diabetes are the most important
  causes of CHD in populations and of social
  gradients in CHD.
• If our concern is to reduce the overall
  population health burden of CHD and the
  population health burden of CHD inequalities,
  then reducing conventional risk factors will do
  the job.
• Eliminating the social gradient in CHD
  attributable to non-traditional risk factor
  mechanisms will make a modest contribution to
  improving overall population health because the
  magnitude of the between social group CHD
  differences is small relative to the total risk
  variation in the population and to that due to
  traditional risk factors.

39
Conclusion
We should increase our efforts to find ways to
influence the multiple pathways from
international, national and local policy through
to individual behaviour that will reduce
conventional risk factors among current and
future generations in richer and poorer countries.
40
Individual and Population Level
Causation Issues for PURE
41
Heart Disease
Steyn, et al Circulation (2005)
All Stroke
42
Major Risk Factors
Social Inequality
CHD
Psychosocial Factors
Reporting Tendency
43
Thank you
44
Triangulating Epidemiology
Populations Sub-groups Individuals Biology Genetic
s Disease
45
Which social inequality is better?
1) Original population levels and social
inequality in CHD
CHD Rate / 100K
Social Advantage
Social Disadvantage
46
Implications for Research on Human Growth and
Development

• in terms of CHD and probably many other diseases
  as well, we should focus on the developmental
  influences (political, economic, social,
  psychological, biological, genetic) on population
  levels and social inequalities in the major
  health behaviors in rich and poor countries

47

• Important to reiterate Geoffrey Rose
• The causes of individual cases of CHD.
• The positive predictive value for individual risk
  prediction is very low because lots of people
  with conventional risk factors dont get CHD
• what makes people susceptible to the risk
  factors?
• Vs

48
Risk of CHD and Lung Cancer by Smoking Status
CHD (fatal or non-fatal)
Lung Cancer Death
Current or Former Smoker
172 / 1000
15 / 1000
Never Smoker
12 / 1000
0
49

• Vs
• The causes of population levels of CHD
• Almost all cases of CHD do have the conventional
  risk factors and so population levels almost
  completely accounted for 3-4 conventional risk
  factors
• Thus eliminating the risk factors eliminates the
  disease

50
If causes can be removed, then susceptibility
ceases to matter



Geoffrey Rose
51
Extensions

• Understand why a relative social gradient
  (RR1.6) remains even in a low risk population
• 2. Given the differential distribution of
  exposure to at least 1 risk factor across
  educational groups, is the exposure distribution
  enough to account for the greater disease burden
  generated among the low educated?
  
• Greater exposure or greater susceptibility?
• exposed to 1 CHD Risk per 1,000
• High ed. 77 65
• Mid ed. 84 83
• Low ed. 89 128

52
(1997)
53
(No Transcript)