Page 43 - IJPS-1-1

Danan Gu

Thirdly, in women, when compared to Type I (the concordance of not-OSA & not-SSA
type), Type IV (the concordance of OSA & SSA type, the third row in Panel C) had 25–71%
lower risk of mortality. The corresponding figures in men were 43–56%. These reduced
ratios were significant except for male centenarians.
Fourthly, in the case of OSA (Panel D), the survival difference between SSA and
not-SSA (i.e., Type IV versus Type III) was non-significant for both women and men
across four age groups, which is consistent with the Kaplan-Meier findings in Figures 2
and 3 where no factors were controlled for.
Fifthly, for those who were SSA, the difference in mortality risk between OSA and
not-OSA (i.e., Type IV versus Type I) was significant in some age groups (Panel E). For
women, the difference was significant in octogenarians and centenarians, whereas for men
it was significant in octogenarians and nonagenarians.
Sixthly, in Panel F the subtype of SSA & OSA was associated with lower mortality risk
for both women and men in all age groups (except women ages 65–79) compared to all
other types combined. In women, compared to their counterparts who were either not-OSA
or not-SSA or both not-OSA & not-SSA (i.e., Types I, II, and III were combined into one
category), those who were OSA & SSA (Type IV) had a 47–70% lower ratio of mortality
in three age groups above age 80. In men, Type IV was associated with 32–50% lower
ratio of mortality.
4. Discussion

Prior research suggested that objective measurements and self-rating tools can be used
together to refine the classification of successful aging (Cernin, Lysack, and Lichtenberg,
2011; Pruchno, Wilson-Genderson, Rose et al., 2010). In this study, with a large nationally
representative sample focusing on the oldest-old, we showed how objectively and subjec-
tively measured indicators of successful aging can be used jointly to construct different
mortality-predictive subtypes among Chinese older adults, a group that is understudied in
the existing literature of successful aging. Subjectively measured successful aging indica-
tors reflect respondents’ own feelings and likely include some conditions that are unob-
served by researchers or medical personnel. Therefore, the distinctive subtypes from the
joint classification of objectively and subjectively measured indicators of successful aging
are expected to have added value to better predict subsequent mortality.
Literature has shown that objectively-defined and subjectively-rated successful aging
measurements could differentiate survivorship in the overall population whether they are
used alone or both are simultaneously presented in the model (Stenholm, Koster, Valkei-
nen et al., 2015; Brown, Thompson, Zack et al., 2015; Diener and Chan, 2011); our results
confirm these findings. However, our approach has extra value for distinguishing subpo-
pulations in terms of mortality risk. Specifically, our findings revealed that when older
adults were not-OSA, there was a significant difference in mortality risk between SSA and
not-SSA (i.e., Type II versus Type I). This is also true that when older adults were not-SSA,
there was a significant difference in mortality between OSA and not-OSA (i.e., Type III
versus Type I). The difference in the latter scenario was likely greater than the difference
in the former scenario. However, when older adults were OSA, there was no difference in
mortality risk between SSA and not-SSA (i.e., Type IV versus Type III); and when older
adults were SSA, the lower mortality risk associated with OSA versus not-OSA (i.e., Type
IV versus Type I) was not universal across age groups.
Overall, the findings of the current study about differential mortality risk across sub-
types of successful aging suggest that the added value of concordance and discordance
between OSA and SSA is important not only for those who are both OSA and SSA, but

International Journal of Population Studies | 2015, Volume 1, Issue 1 37