Longitudinal Analysis of the Relationship between Migration and Health Status

1
Longitudinal Analysis of the Relationship between
Migration and Health Status

• Study of Adult Population of Indonesia

Salut Muhidin, Dominic Brown Martin
Bell 4th International Conference on
Population Geographies 12 July 2007, Hong Kong
2
Whats New?

• Some studies have been done on the link between
  migration (M) health (H). Among others
• UK (Bentham 1988 Boyle et al. 1999 2001
  Dorling 1998)
• USA (Findley 1988 Kington et al. 1998)
• NL (Verheij et al. 1998)
• Australia (Larson et al. 2005)
• The studies are applied in the context of
  developed countries. YET, it is still little
  known in the developing countries. One of its
  main reason is data limitation.
• The ideal design for testing the M-H relationship
  requires life histories data, with appropriate
  information on background characteristics at
  different points in the life cycle
• Fact Indonesia has now a longitudinal data which
  cover information on migration and health. IFLS
• The contribution here
• Investigating the relationship M-H in the context
  of a developing country
• Using the available longitudinal data, i.e. IFLS

3
Research Question

• Is there any relationship between migration and
  health in the context of Indonesia?
• Q1 Do migrants differ from non migrants in terms
  of
• health and socioeconomic status?
• Q2 Does the probability of migration depend upon
  
• the health status accounting for
  socioeconomic
• variables?

Health
Migration
4
Side 1 Migration

• Determinant of Migration
• It is strongly related to particular personal
  traits and some important life events e.g.
  education, marriage and separation, job related,
  and retired (elderly). Age regularities in
  migration (Rogers and Castro, 1980)
• Dimension Migration
• Time Permanent - temporary (Intention to stay)
• Distance short - long
• Geographic Internal and international
  (urban-rural)

5
Side 2 Health

• Health has multi dimensions
• It has been linked to many factors physical,
  mental, and social well-being, genotype and
  phenotype, gender and place of residence.
• Health measures
• General Health Status (GHS)
• Physical ability (ADL)
• Chronic illness
• Mental Health, or
• Health related behaviors, etc.

6
Data Source

• Indonesia Family Life Survey (IFLS)
• Longitudinal survey
• 3 waves 1993, 1997, and 2000
• Organizer
• RAND, University of Indonesia and Gadjah Mada
  University
• Coverage
• 13 provinces (83 population of Indonesia)
• 7,224 HH (Base in 1993)
• 6,820 HH (94 were re-interviewed)
• 33,081 people (Base in 1993)

7
Data Structure
IFLS-1
IFLS-3
IFLS-2
Stayed or Moved away
Stayed or Moved away
Health Status 1997
Health Status 2000
Health Status 1993
All respondents
Re-tracked respondents
8
Data Structure
Health info 12,985
IFLS 1993 N33,081
Age 15 21,630
Health 93 Migration 93-97 N 12,985
Age 15lt 11,451
Migration info 21,630
IFLS 1997 (MH93) N12,985
Health info 11,495
Traced 12,366
Health 97 Migration 97-00 N 11,495
Died (454) No traced (165)
Migration info 12,366
9
Variable Migration

• Definition of Migration
• It is based on the status of leaving (staying)
  or changing their usual residence as recorded at
  the baseline (previous) survey ? Current
  Migration
• IFLS2 Migration 1993-1997
• IFLS3 Migration 1997-2000
• Type of Migration
• Short Distant (inter village and district) and
  Long Distant
• Information on migration characteristics
  (age, destination and reasons) of those who have
  moved was also collected.

10
Variable Health Status

• General Health Status (GHS) Self reported
• GHS was generated from a question In
  general, how is your health at this time? The
  answers were (a) Very healthy, (b) somewhat
  healthy and (c) somewhat unhealthy or (d)
  unhealthy.
• Activity of Daily Living (ADL) Reported
  observed
• ADL was constructed by using nine questions
  if the respondent could do (was capable of)
  certain daily activities. The answers were three
  possibilities easily, with difficulty, and
  unable to do. It includes three functions
• (1) mobility (to walk 5 kilometers to bow,
  squat, and kneel to stand up from sitting in a
  chair or from sitting on the floor),
• (2) personal care (i.e. to dress and to go to the
  bathroom without help)
• (3) home occupation (i.e. to carry a heavy load
  to sweep and to draw a pail of water).

11
ResultsProportion of Migration
Current
Note GHS (General Health Status) and ADL
(Activity of Daily Living)
12
Current
13
Current
14
Models

• Model 1 ? Selectivity
• What is the probability of migration with
  respect to the health status (does migrant has
  better health?).
• ? Migration(93-97) f (Health 93)
• ? Migration(97-00) f (Health 97)
• Model 2 ? State Dependency
• What is the probability of migration with
  respect to the current and previous health
  status.
• ? Migration(97-00) f (Health 93, Health 97)
• ? Migration(97-00) f (Health 97) among Healthy
  Pop.93
• Logit Regression Model
• the dependent variable is (1) Migration or
  (0) No migration

15
Model 1A Selectivity
Migration(93-97) f (Health 93)
Without Control Variable
With Control Variables
Yet significances are washed out by covariates
16
Model 1B Selectivity
Migration(97-00) f (Health 97)
Without Control Variable
With Control Variables
Yet significances are washed out by covariates
17
Model 2A Dependency
Migration(97-00) f (Health 93, Health 97)
18
Model 2B Dependency
Migration(97-00) f (Health 97) among Healthy 93
19
Covariates

• Age Groups 15-19, 20-24.60
• Sex Male (1) Female (0)
• Education Primary, Secondary, Tertiary
• Employment Working (1)
• Expenditure 21-40, 41-60, 61-80, 81-100
• Marital Status Union (1)
• Birth Place Urban (1)
• Current Residence Java-Bali (1)

Age Sex
Education
20
Conclusion

• Longitudinal data (IFLS survey) offers the
  possibility
• To assess the relationship Health Migration in
  Indonesia
• To evaluate the selectivity dependency
• In the context of Indonesia
• The relationship between Health and Migration
  tends to be positive
• People with good health status (ADL in
  particular) are more likely to be positively
  associated with migration (Mig 97-00 in
  particular)
• YET, the significances are often washed out by
  other socio-economic covariates.
• Age Separation Young Older
• Data Focus on IFLS2 IFLS3
• Health Measurement

21
Discussion

• Measurement of Health
• Measurement of Migration
• Different Result?
• More Questions
• Health Changes Improved, Deteriorated, Stable
• Does health improve or deteriorate with
  migration?
• Changes in socio-economic variables
• employment status, marital status, income
• Relationship Migration ? Health Status

22
Thank You