Recent Research Findings

Environmental factors have significant impact on healthy longevity in China. Using CLHLS 2002 and 2005 longitudinal individual and community datasets, multi-level and multivariate modeling analyses show that higher employment rate in the community reduces one’s risk of ADL disability, cognitive impairment, frailty, and mortality by 20-45% (p<0.05~p<0.001), while higher per capita GDP and lower illiteracy rate decreases one’s risk of cognitive impairment by 30-40% (p<0.05~p<0.001). Analyses show that air pollution increases risks of frailty, ADL disability, and cognitive impairment by 10-25% (p<0.05~p<0.001), whereas more rainfall reduces likelihood of ADL disability and self-reported poor health. Furthermore, too low or too high seasonal temperature increases risks on some poor health outcomes and mortality by 30-50% (p<0.01~p<0.001). None of the statistically significant effects of community variables on health and survival can be explained by individual factors, indicating the robustness of the environmental effects on healthy longevity (Zeng, Gu, Purse, Hoenig 2007, presentation at the 2007 GSA annual meeting, Nov 16-20, San Francisco).

CLHLS study denies the traditional son preference in China. The analysis of data from CLHLS 2002 and 2005 waves clearly demonstrates that having daughter(s) is beneficial at old ages in China, with respect to enjoying better filial piety of and relationship with children, care provided by children, and also maintaining a better cognitive capacity and reducing mortality risk, while controlling for various confounding factors. Such daughter-advantages are mostly more profound among the oldest-old as compared to younger elderly, and more profound in rural areas than in urban areas. We conclude that the publicly predominant perception of son preference in rural China, which leads to action of aborting the girl fetus to insure having at least one son under low fertility, is not a rational choice of peasants’ own interests for better old age care (Zeng and Vaupel 2007).

Frailty index and quality of dying at oldest-old ages. Based on 13,717 older adults from the 2002 and 2005 waves of the CLHLS, we examine the association between frailty and type of death. Four types of death were classified: Type I (less than 30 days bedridden and no suffering), Type II (less than 30 days bedridden and suffering), Type III (30 or more days bedridden and no suffering), and Type IV (30 or more days bedridden and suffering).  Multinomial analyses showed that higher levels of frailty significantly increased the relative risk ratios (RRR) of mortality for all types of death. Men with the highest levels of frailty were most likely to experience a Type IV death (RRR 8.70, 95% CI 6.31-12.00), whereas women with the highest levels of frailty were most likely to experience a Type III death (RRR 11.53, 95% CI, 17.84-16.96). Regardless of frailty, centenarians and nonagenarians were most likely to exhibit a Type I death , and adults aged 65-79 and 80-89 were more likely to experience a Type II death. Adjusting for compositional differences had little impact on the link between frailty and type of death for both sexes and age groups. The association between frailty and type of death differs by sex and age. Health scholars and practitioners need to consider sex and age heterogeneity in frailty to provide more effective measures to reduce preventable suffering prior to death. This study provides new information about the association between frailty and type of death that is largely unknown in most developing countries (Gu, Dupre, Warner, and Zeng 2007).

Association of body mass index with healthy longevity in China. This study is based on unique data from the Chinese Longitudinal Healthy Longevity Survey conducted in 2002 and 2005 (n=15,972, including 3,142 centenarians, 3,747 nonagenarians, 4,238 octogenarians, and 4,845 seniors aged 65-79, covering 85% of the total population of China). Controlling for various confounding variables including health status one year ago, the cross-sectional binary and ordered logistic regression analyses indicate that (1) physical health measured by self-reported health, IADL, and Frailty Index is best among those with BMI 25-29.9 (P<0.000), followed in order by those with BMI 30+, 18.5-24.9, and <18.5 (P<0.000-0.069); (2) the higher the BMI, the better cognitive function measured by MMSE (P<0.000), but an opposite and not-significant estimate was obtained for ADL. Multivariate survival analyses show that old men with BMI 30+ have the lowest risk of mortality; men with BMI 18.5-24.9 and 25-29.9 are 20% and 30% less likely to die, respectively, as compared to those with BMI <18.5 (p<0.000). The effects for men are slightly weakened but still significant after controlling for SES, family/social support, health practice, and health conditions. However, as compared with the lowest BMI, the protective effects of higher BMI on mortality among old women are only around 10-15%  (p<0.05) and they are not statistically significant after controlling for various confounders. Using the largest sample of the oldest-old, we also compared the impacts of BMI on health and mortality between the oldest-old aged 80+ and younger elderly aged 65-79 (Zeng, Gu, Liu 2007, Poster at the 2007 GSA annual meeting, Nov 16-20, San Francisco).

Left handedness has no significant disadvantage in longevity. Using the Chinese Longitudinal Healthy Longevity Survey (1998, 2000, 2002, and 2005; n=21,000+; age 65+), we examined factors associated with left-handedness, and estimated consequent effects on mortality. Across all ages, and after controlling for numerous covariates, we found no clear pattern of longevity due to left-handedness. However, compared with elderly Chinese women (3.5%, 95% CI 3.2-3.8%), older Chinese men have a higher prevalence of left-handedness (4.0%, 95% CI 3.7-4.4%), consistent with previous studies in other countries. A novel finding of this study is that factors associated with left-handedness appear to be gender-specific. On average, non-Han ethnic minorities are more likely to be left-handed (OR=1.64, 95%CI 1.32-2.03), and we also observed associations between SES and handedness (e.g., OR=0.84, 95% CI 0.72-0.98 for those with 1+ years of schooling as compared with those without any schooling). Geographic variation in the prevalence of left-handedness across China is also substantial. Survival analyses show that lefthanders in China do not appear to have a shorter lifespan (HR=1.06, 95%CI 0.94-1.20 for women and HR=1.11, 95%CI 0.97-1.27 for men), confirming earlier findings at younger ages.  Further work will evaluate related hypotheses that lefthanders have a higher incidence of chronic medical conditions and disability (Gu, Purser, and Zeng 2007, presentation at the 2007 GSA annual meeting, Nov 16-20, San Francisco).