Page 15 - IJPS-3-1

Gu D, Feng Q, Sautter JM and Qiu L

3 Results

3.1 Description of Urban Life Exposure among the Sample

Table 2 presents the distribution of study variables among the analytical sample of
older adults recruited in the CLHLS 2002–2014. In this sample, 12.7% were born in
an urban area and 39.1% were living in an urban area at the time of the first CLHLS
interview. When the respondents were classified with both time points, 58.9% were
born and remained in a rural area (rural-rural), 28.4% moved from a rural area to an
urban area (rural-urban), 10.6% were born and remained in an urban area (urban-urban),
and 2.1% moved from an urban area to a rural area (urban-rural). These urban-to-rural
older adults with early-life-only urban exposure were possibly people who moved to
the countryside to avoid war and never returned to urban areas. It is also possible that
some of them were urban youth migrating to rural areas for the call of government in
the 1950s and 1960s.
When PLO was further incorporated into the four types of classification, we found
that about half of respondents had no urban life exposure, 6.1% had full urban
exposure, and about 20% had late-life-only exposure (rural-urban, r-PLO). Other types
of urban life exposure had low representation: 4.3% had mid-life-only exposure (rural-
rural, u-PLO), possibly industrial workers in township/village enterprises; around 8.6%
had mid-late-life exposure (rural-urban, u-PLO), possibly those who moved to an
urban area in their early occupational career; about 1.5% had early-life-only exposure
(urban-rural, r-PLO); 0.6% had early-mid-life exposure (urban-rural, u-PLO); and
about 4.5% had early-late-life exposure (urban-urban, r-PLO), possibly those who
assumed an agricultural job in a suburb.
About 21.7% of the analytical sample reported migration during the life course, ever
moving to another county/city. For both lifetime-exposure and lifetime-no-exposure
respondents, most had no migration experience. Overall, in a relative scale, the mid-
late-life exposure older adults had the highest proportion of migration (60%), as
compared to only 12% among the no exposure older adults.
3.2 Association between Exposure to Urban Life and Mortality Risk
Table 3 presents relative hazards of mortality and 95% confidence intervals (CIs) for
urban versus rural environment at birth and at the first CLHLS interview, stratified by
age group and gender. Except for women aged 65–79 (Model II in the upper panel of
Table 3), those born in an urban area had no advantages in mortality over those born
in a rural area. Urban residence at the first CLHLS interview was associated with
lower risk of mortality for men and both sexes combined in each age group when only
demographics were controlled for (Model I), yet the relative mortality risk was not
significant when other covariates were added (Model II).
Table 4 further presents the relative hazards of mortality for the four-type
classification of urban life exposure with rural-rural as reference. Among older
adults aged 65 –79, rural-to-urban migrants had 13% lower mortality risk when only
demographics were controlled for, but the relative mortality risk was not significant
when other covariates were further controlled for. Residential change from urban
to rural areas (urban-to-rural migrants) was associated with increased mortality risk
compared to rural-rural. The increased risk was more robust in the oldest-old than in
the younger old adults, and more robust in women than in men. Although the urban-
urban respondents had a lower mortality risk compared to rural-rural for the two
age groups and two sexes, the association was generally not significant at p < 0.05
except that women at ages 65–79 had a 37% (p < 0.05) higher mortality risk when
socioeconomic status, family/social support, health behaviors, and baseline health were
taken into consideration.
Table 5 further incorporates PLO and migration into urban exposure, representing the
eight-type and fourteen-type classifications, respectively. The upper panel shows that
compared to no exposure, those with mid-late-life exposure had 9% lower mortality

International Journal of Population Studies 2017, Volume 3, Issue 1 9