Developing and validating an InequityinHealth index IHI based on Millennium Development Goals

1
Developing and validating an Inequity-in-Health
index (IHI) based on Millennium Development Goals
The 12th World Congress on Public Health.
Istanbul, Turkey

• Javier Eslava-Schmalbach, MD, MSc, PhD
• Helman Alfonso, MD, MSc, PhD
• Hernando Gaitán, MD, MSc
• Henry Oliveros, MD, MSc
• Carlos Agudelo. MD, MSc
• National University of Colombia
• Curtin University and University of Western
  Australia
• San Rafael Hospital . Colombia
• Funded by Colciencias

2
Inequity

• The term inequity has a moral and ethical
  dimension. It refers to differences which are
  unnecessary and avoidable but which are also
  considered unfair and unjust. So, in order to
  describe a certain situation as being
  inequitable, the cause has to be examined and
  judged to be unfair in the context of what is
  going on in the rest of society (Whitehead,
  2000)

3
Gini Coefficient and Lorenz Curve
4
Limitations

• A single coefficient must be obtained for each
  health condition considered
• The Gini coefficient changes depending on the way
  a population is sorted
• Equally bad health

5
Gini Coefficient and Lorenz curve
Gini0.23
Gini0.16
6
Millennium Development Goals

• Signed by 189 States in 2000. They represent
  agreed-upon goals for life and were aimed at
  improving world conditions by 2015. They included
  several areas concerning health

7
Objective

• This study was aimed at developing and validating
  a new Inequity in Health Index using the
  indicators proposed for monitoring the progress
  of the Millennium Development Goals

8
Perspective

• The perspective of equity-in-health as being
  equal outcomes in health in equal populations
• The MDG imply that unequal outcomes are avoidable
  and unnecessary
• Value judgments about outcomes
• vulnerable population, avoidable outcomes
  (unjust and unfair) inequity

9
Methodology

• Design ecological study
• Variables were selected from
• Millennium Development Goals.
• Human Development Report, 2005
• United Nations databases
• World Development Indicators 2005
• Variables were selected if they were individually
  registered in more than 40 of total countries

10
Disparity measurement

• Attributable fraction
• Negative outcomes
• (country X best country)/country X
• Positive outcomes
• (best country country X)/best country
• Which percentage of outcome is explained by
  living in country X and not in the referent
  country
• AF 1 higher
  disparity

11
Development of the index
1. Principal factor analysis 2. Principal
component analysis
Area
12
Methodology
1
13
Index Reliability- Validity

• Sorting the variables in three different ways
  (Spearman and Kendal concordance coefficient)
• Internal consistency (Cronbachs alpha)
• Criteria validity human development index, human
  poverty index, health gap indicator, probability
  of dying before reaching 40, life expectancy and
  Gini coefficient (Spearman)
• Discriminant validity (Kruskal Wallis test)
• Sensitivity to change (Sign test)
• Stata 8.2 and Excel

14
Results
15

• Variable Med 95 CI Min
  Max
• Children Underweight 14 11 - 17
  1.0 48
• Children mortality 30 22
  39 3.0 284
• Maternal Mortality rate 110 83 -
  146 0.0 2000

16
(No Transcript)
17
Principal factor analysis. Eigenvalue and
cumulative variance
18
Internal consistency
19
(No Transcript)
20
(No Transcript)
21
MDG-based Inequity in Health Index
World Inequity-in-Health Index (177 countries),
2003 (area0.3033p)
IHI, Democratic Republic of Laos (region1), 2003
22
(No Transcript)
23
Conclusion

• The novelty of this proposed index lies in
  building a bi-dimensional composite allowing
  inequity in health to be graphically and
  quantitatively estimated in countries, regions
  and around the world.

24
Thanks