Socioeconomic Differentials in Mortality among the Oldest Old in China

Document Sample
Socioeconomic Differentials in Mortality among the Oldest Old in China Powered By Docstoc
					Socioeconomic Differentials in
Mortality among the Oldest Old
           in China

           Haiyan Zhu
             Yu Xie

      University of Michigan
Background


   Well documented inverse relationship:
      Higher SES        Lower mortality

   However, magnitude varies by age
Yu Xie (1994). “Log-multipliticative Models for Discrete
Time, Discrete Covariate Event-History Data”
   SES differentials in mortality diminish with
    age (Convergence hypothesis)?
    – The oldest old are detached from economic
      activities and thus immune from some causal
      mechanisms (e.g., job hazards, work stress)
    – Biological determinants, rather than social
      determinants, play a predominant role
    – The oldest old are a select group with respect to
      unobserved health traits
Research Questions:

   Does SES still affect mortality among
    the oldest old Chinese?
   Do SES differentials in mortality
    disappear beyond an old age (e.g., 80,
    90, or 100)?
Importance

   Previous studies have not specifically
    studied these questions
   This study focuses on these questions
    using a sample of oldest old in China
Data and Measures

   The Chinese Healthy Longevity Survey:1998
    and 2000 cohort data
   SES: composite of education and urbanity
    –   Illiterate, rural
    –   Illiterate, urban
    –   Literate, rural
    –   Literate, urban
   Covariates: age, sex, ethnicity, region, self-
    rated health, ADL, and “time-interval”
    – “time-interval”: time since the first interview
Analyses

   Three alternative target populations
    80 and older, 90 and older, and 100 and
     older
   Logit, discrete time analysis
Descriptive Statistics
  Table 1 Distribution of SES and Mortality by SES, female
Female                                Number         Percentage   Death (%)

80+

               1. Illiterate, rural   3114           59.0         21.4

               2. Illiterate, urban   1486           22.3         21.5

               3. Literate, rural     236            8.7          11.9

               4. Literate, urban     426            10.0         15.0
90+

               1. Illiterate, rural   2279           60.2         40.8

               2. Illiterate, urban   946            27.9         38.5

               3. Literate, rural     115            4.9          41.8

               4. Literate, urban     181            6.9          40.6

100+

               1. Illiterate, rural   1264           61.0         60.4

               2. Illiterate, urban   424            31.4         60.9
               3. Literate, rural     50             2.5          52.2

               4. Literate, urban     68             5.1          52.2

       Note: numbers are unweighted; percentages are weighted.
Descriptive Statistics
(Continued)
  Table 2. Distribution of SES and Mortality by SES, male
Male                                  Number         Percentage   Death (%)

80+

               1. Illiterate, rural   958            26.7         27.1

               2. Illiterate, urban   356            6.4          30.0

               3. Literate, rural     1167           41.2         23.5

               4. Literate, urban     1062           25.8         21.1
90+

               1. Illiterate, rural   554            29.4         51.3

               2. Illiterate, urban   197            9.2          53.5

               3. Literate, rural     586            36.4         45.4

               4. Literate, urban     419            25.0         34.8

100+

               1. Illiterate, rural   164            33.6         69.3

               2. Illiterate, urban   63             15.3         72.0
               3. Literate, rural     153            31.5         65.3

               4. Literate, urban     78             19.6         53.4

       Note: numbers are unweighted; percentages are weighted.
Results
Table 3. Effects of SES on Mortality (controlling ethnicity and region)

                                             Mortality
Variable                           80+         90+            100+

Female                          .657***      .762**        .789**
Age (time-varying)              1.092***     1.055***      1.035


Interval                        1.014**      1.017***      1.030***
SES
  Illiterate, urban             .997         .937          .981
  Literate, rural               .773*        .912          .932
  Literate, urban               .708**       .731*         .657**

* p<.05; ** p<.01; *** p<.001
Results (Continued)
Table 4. Effects of SES on Mortality (controlling health, ethnicity and region)
                                               Mortality
  Variable                          80+          90+           100+


  Female                          .622***      .690***       .716***


  Age (time-varying)              1.080***     1.041**       1.023


  Interval                        1.017**      1.022***      1.036***


  SES

   Illiterate, urban              1.022        .930          .939


   Literate, rural                .795        .936          1.029

   Literate, urban                .711**       .717*         .601**
  Self-rated Health               1.649***     1.612***      1.473***
  ADL                             1.690***     1.824***      1.906***
   P<0.1; * p<.05; ** p<.01; *** p<.001
             time1   time2   time3
1st cohort      80      81      82


2nd cohort    81       82      83



3rd cohort     82       83     84
Summary

   SES differentials in mortality persist,
    using all three operationalizations of
    “old age”
   Convergence was not supported
   Future studies

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:5
posted:8/30/2012
language:English
pages:14