Presented in the January 2015 meeting of the COST Action "Adapting European health systems to diversity" (ADAPT)

1. Applying intersectionality
in health and healthcare
Some views from social epidemiology
Davide Malmusi
Agència de Salut Pública de Barcelona
ADAPT meeting, Bologna, 22 January 2015

2. Conceptual framework: determinants of health inequalities
Commission to Reduce Social Inequalities in Health in Spain

3. • Some issues on inequality in healthcare by socioeconomic
position (SEP) and gender
• Intersectionality in health and in policy evaluation: guide
and examples
• Migrant status and intersectionality: an example
• Summing up
Outline

4. • Self-ratings of health capture true inequalities in
health by SEP – or they even underestimate them1,2
• GPs disagree more often on low SEP patients’ health
ratings – tending to overestimate their health3
Evidence on SEP inequality in quality of
care - within public systems (1)
1. Bago d’Uva T, et al. Differential health reporting by education level … Int J Epidemiol. 2008;37:1375-83.
2. Beam Dowd J, Todd M. Does Self-reported Health Bias ... J Gerontol B Psychol Sci Soc Sci. 2011;66:478-89.
3. Kelly-Irving M, et al. Do general practitioners overestimate the health ... Soc Sci Med. 2011;73:1416-21.
Patient’s education level GP agrees GP under-rates GP over-rates
High education 75% 11% 14%
Medium education 74% 11% 16%
Low education 50% 12% 37%

5. • Among the chronically ill, low SEP patients perceive
lower quality of doctor-patient relationship4
• Coronary risk scores overestimate mortality risk –
much more in higher than lower social classes5
Evidence on SEP inequality in quality of
care - within public systems (2)
4. Vonneilich N, et al. [Social inequality and perceived quality...]. Gesundheitswesen. 2011;73:211-6.
5. Ramsay SE, et al. Prediction of coronary heart disease... Eur J Cardiovasc Prev Rehabil. 2011;18:186-93.

6. • Poverty screening – and referral6
• Socioeconomic factors in risk algorithms7
• Consciousness of own bias, positive discrimination8
• Resource allocation and quality incentives in
deprived areas9
• Population-wide vs opportunistic programs10
SEP inequality in quality of care:
some steps in the good direction
6. Bloch G. Poverty: a clinical tool for primary care in Ontario. Revised Nov 2013. Available at http://ocfp.on.ca
7. Fiscella K, et al. Adding socioeconomic status to Framingham scoring… Am Heart J. 2009;157:988-94.
8. Bærøe K, Bringedal B. Just health: On the conditions for acceptable... J Med Ethics. 2011 ;37:526-9.
9. Kiran T, et al. The association between quality of primary care... J Epidemiol Community Health. 2010;64:927-34.
10. Palència L, et al. Socio-economic inequalities in breast and cervical cancer... Int J Epidemiol. 2010;39:757-65.

7. • GPs more likely to
overestimate
women’s health
than men’s3
• But are women’s
and men’s health
self-ratings equally
valid?
Gender inequalities in quality of care
3. Kelly-Irving M, et al. Do general practitioners overestimate the health ... Soc Sci Med. 2011;73:1416-21.

8. Gender inequalities in health:
Perception or real illness?

9. • Women live longer but (in most countries) report
poorer health than men
• Is this a matter of perception or illness behaviour?
Gender inequalities in health:
Perception or real illness?

10. Data source: Spanish
National Health
Survey 2006
Gender inequalities in health:
Perception or real illness?

11. Data source: Spanish
National Health
Survey 2006
Gender inequalities in health:
Perception or real illness?

12. • Women live longer but (in most countries) report
poorer health than men
• Is this a matter of perception or illness behaviour?
NO -> They really suffer from more (non-life-threathening
but limiting and painful) chronic conditions
Gender inequalities in health:
Perception or real illness?

13. • As healthcare systems and professionals, we prioritise
action on life-threathening diseases
• Conditions that don’t kill, but cause lot of suffering and
disability in large shares of the population, fall behind
• Reorienting healthcare towards these conditions may
be also good for gender equity
Gender inequalities in health:
What does this matter for healthcare?

14. • Some issues on inequality in healthcare by SES and gender
• Intersectionality in health and in policy evaluation: guide
and examples
• Migrant status and intersectionality: an example
• Summing up
Outline

15. Intersectionality: a quick guide

16. The analysis of health inequalities and
health equity impacts of policies is
usually centered on social class / socio-
economic position.
Other social relations that generate
health inequalities are usually more
neglected – or treated separately.
Intertwined mechanisms of power
relations that interact among them: need
for intersectional approaches.
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide

17. Intersections between axes create complex
social locations that are more central to the
nature of social experiences than any single
axe of inequality.
Groups in society are affected by their
position in multiple systems of power and
oppression, a “matrix of domination” that
changes over time and place and in
different institutional domains.
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide

18. The health effects of an intervention,
context, condition… differ depending on
one’s position according to the
intersection of several axes / dimensions
of inequality / power relations.
When designing/implementing
interventions, targeted policies can be as
ineffective as general ones in addressing
multiple identities and power relations.
Tools include multistrand/equality/power
mainstreaming and Intersectionality-
Based Policy Analysis.
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide

19. The health effects of an intervention,
context, condition… differ depending on
one’s position according to the
intersection of several axes / dimensions
of inequality / power relations.
When evaluating interventions (or testing
associations) we have to consider that
populations are not uniform (nor binary
or categorical either), and go beyond the
simple “whether” it works, to understand
“for whom” (and “why” and “how”)
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide

20. Intersectionality: a quick guide
Example 1. Family composition and health
Ruiz-Cantero MT et al. A framework to analyse gender bias… JECH 2007
Model adjusted by sex and
social class:
Larger household size
slightly associated with
poorer self-rated health
Cohabiting with a >65 aged
person slightly associated
with better health
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.

21. Model stratified by sex and adjusted by social class:
Household size and cohabiting with elderly strongly associated
with poorer self-rated health among women but not among men
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide
Example 1. Family composition and health
Ruiz-Cantero MT et al. A framework to analyse gender bias… JECH 2007

22. Model stratified by sex and social class:
Household size and cohabiting with elderly strongly associated
with poorer self-rated health only among manual women
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide
Example 1. Family composition and health
Ruiz-Cantero MT et al. A framework to analyse gender bias… JECH 2007

23. Intersectionality: a quick guide
Example 2. Crisis and mental health
Bartoll X et al. The evolution of mental health in Spain… Eur J Public Health 2014
Prevalence (%) of poor mental health (GHQ-12>2)
The “zero change” in the whole population masks a worsening in
men and improvement in women.
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
2006/07 2011/12 Prevalence Ratio
Total 19,6 19,8 1,01
Men 14,7 16,9 1,15**
Women 24,6 22,7 0,92*
* p<0,05 ** p<0,01 *** p<0,001

24. Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
2006/07 2011/12 Prevalence Ratio
Men 14,7 16,9 1,15**
15-24 11,3 11,2 0,98
25-34 15,2 16,1 1,05
35-44 15,0 18,7 1,24*
45-54 14,8 19,1 1,29**
55-64 16,9 17,3 1,02
* p<0,05 ** p<0,01 *** p<0,001
Intersectionality: a quick guide
Example 2. Crisis and mental health
Bartoll X et al. The evolution of mental health in Spain… Eur J Public Health 2014
Prevalence (%) of poor mental health (GHQ-12>2)

25. Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
2006/07 2011/12 Prevalence Ratio
Men 14,7 16,9 1,15**
Social class I 11,8 12,2 1,04
II 14,6 15,4 1,06
III 15,9 15,4 0,97
IV 14,8 18,2 1,23**
V 15,3 18,6 1,21
* p<0,05 ** p<0,01 *** p<0,001
Intersectionality: a quick guide
Example 2. Crisis and mental health
Bartoll X et al. The evolution of mental health in Spain… Eur J Public Health 2014
Prevalence (%) of poor mental health (GHQ-12>2)

26. Within men, the worsening is concentrated among the middle-
aged and the manual social class, and stronger in immigrants.
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
2006/07 2011/12 Prevalence Ratio
Men 14,7 16,9 1,15**
Born in Spain 14,7 16,4 1,11*
Foreign-born 14,5 19,3 1,33*
* p<0,05 ** p<0,01 *** p<0,001
Intersectionality: a quick guide
Example 2. Crisis and mental health
Bartoll X et al. The evolution of mental health in Spain… Eur J Public Health 2014
Prevalence (%) of poor mental health (GHQ-12>2)

27. Intersectionality: a quick guide
Example 3. Urban renewal and health
Mehdipanah R et al. The effects of an urban renewal project… JECH 2014
Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.

28. Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide
Example 3. Urban renewal and health
Mehdipanah R et al. The effects of an urban renewal project… JECH 2014

29. Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Intersectionality: a quick guide
Example 3. Urban renewal and health
Mehdipanah R et al. The effects of an urban renewal project… JECH 2014

30. Palència L, Malmusi D, Borrell C. Incorporating intersectionality in policy evaluation analysis. A quick guide.
Effects also varied by neighbourhood: clearer health
improvement in the periphery (intervention with strong residents’
participation, including subway, elevators, escalators,) but not in the
inner city (where the program was one among the many changes going on)
Qualitative studies are also key to study intersectionality: in an
inner city neighbourhood, age and immigrant status shaped the
perception on changes occurring - a senior group reported mainly
discontent with gentrification and immigration, an immigrant-background
youth group was satisfied with new public spaces and youth centers
(Mehdipanah et al., Health Place 2013; Mehdipanah et al., Soc Sci Med
2015)
Intersectionality: a quick guide
Example 3. Urban renewal and health
Mehdipanah R et al. The effects of an urban renewal project… JECH 2014

31. • Some issues on inequality in healthcare by SES and gender
• Intersectionality in health and in policy evaluation: guide
and examples
• Migrant status and intersectionality: an example
• Summing up
Outline

32. Intersectionality: my first (?) encounter

33. Intersectionality in health status in Catalonia
Objectives: To propose a health-equity-based
classification of migration types. To analyse inequalities in
self-rated health in Catalonia in the intersections of
migration type, gender and social class. To analyse the
contribution of socioeconomic conditions to migration-
related health inequalities in Catalonia.
Methods: Cross-sectional analysis of two 2005-06
population surveys (Health and Living Conditions).
Population aged 25 to 64. “Age-adjusted” prevalences
and logistic regressions.

34. Intersectionality in health status in Catalonia
Migration type internal/international, rich/poor,
recent/less recent based on birthplace and year of arrival:
• Catalonia (Local-born)
• Rest of Spain
– High tertile of regional development in 1981
– Medium or low tertile of development
• Foreign countries
– Very highly developed countries (UN HDI) in 2006
– Rest of the world, divided based on year of arrival:
• Recent immigrants (since 2000)
• Less recents (until 1999)

35. Intersectionality in health status in Catalonia
2 dimensions: Migration type and gender
Odds Ratio of fair/poor self-rated health by migration type
Adjusted by age (Model 1) + class (M2) + standards of living (M3)
Women Men

36. Intersectionality in health status in Catalonia
Social class based on current, last or partner’s occupation:
• I: higher-level professionals, managers, directors of large
companies
• II: medium-level professionals and directors of small companies
• III non-manual: administrative workers, clerks, safety and
security workers
• III manual: self-employed and supervisors in manual occupations
• IV-a: skilled manual occupations
• IV-b: semi-skilled manual occupations
• V: unskilled manual occupations

37. Intersectionality in health status in Catalonia
Sample by social class and migration type (women)
1192
880
171
108
88
33
102
342
146
33
17
3
60
176
139
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
No manual Manual qualificat Manual no qualificat
Estranger-pobre
Estranger-ric
Espanya-pobre
Espanya-rica
Catalunya

38. Intersectionality in health status in Catalonia
3 dimensions: Gender, social class and migration
Age-adjusted prevalence of material deprivation

39. Intersectionality in health status in Catalonia
3 dimensions: Gender, social class and migration
Age-adjusted prevalence of fair/poor self-rated health
The two bars for the same category show results in the two surveys

40. Summing up
Gender and class also determine inequities in quality of care
People’s positions of power arise from (time- and place-specific)
matrices of domination according to multiple axes
In quantitative studies, multi-stratified analysis can reveal the
specific groups that are disadvantaged or whose health is
affected by a policy, condition… – and give us hints of why, which
can be further explored with qualitative studies

41. sophie-project.eu
@sophieproject
info@sophie-project.eu
dmalmusi@aspb.cat
Thank you!
Gràcies!
Grazie!
Photos: Roberto Brancolini,
Roberto Malaguti, Fredi Roig