Page 94 - 南京医科大学自然版

第46卷第8期
·1200 · 南 京 医 科 大 学 学 报 2026年8月

grouping cutoff at 14：participants with scores>14 were lipidemia，rheumatism/rheumatoid disease，chronic ne⁃
classified into the high psychological resilience group phritis，and hepatitis. Additional covariates included
（exposure group），while those with scores≤14 were as⁃ self ⁃ rated quality of life（categorized as very good or
signed to the low psychological resilience group（con⁃ good，average，bad or very bad），obesity index（ABSI，
trol group）. BRI，WHtR，and BMI），and an activity score. The ac⁃
1.2.2 Outcomes tivity score encompassed housework，fieldwork，garden⁃
The primary outcome of this study was all⁃cause ing，reading，raising domestic animals/pets，playing
mortality. Mortality status（death or survival）was ascer⁃ cards/mahjong，watching TV/listening to the radio，and
tained based on data from the Wave 7 or Wave 8 follow⁃ organized social activities. Frequency was scored as fol⁃
up surveys. In the CLHLS，survival status for living lows：almost every day=4；weekly（but not daily）=3；
participants was confirmed through direct interviews. monthly（but not weekly）=2；sometimes（but not month⁃
For deceased participants or those unable to respond ly）=1；never=0.
due to severe health conditions，information was ob⁃ Dietary diversity was evaluated based on the con⁃
tained via proxy interviews. These proxies provided de⁃ sumption frequency of nine food groups：fresh vegeta⁃
tails regarding the date of death and circumstances pre⁃ bles，fruits，meat，fish，eggs，bean products，milk/dairy，
ceding death. Survival time was calculated as the inter⁃ nuts，and tea. A score of 1 was assigned if a food item
val from the baseline survey to the date of death for de⁃ was consumed at least once per week（options：“almost
ceased participants，or to the date of the last completed every day”，“quite often”，“not every day”，“not every
follow ⁃ up interview for censored participants（survi⁃ day but at least once per week”）；otherwise，a score of
vors）. Dates of death were verified using proxy reports 0 was assigned. Finally，basic activities of daily living
and，where available，death certificates. （BADL）were assessed across six domains：bathing，
1.2.3 Covariates dressing，toileting，indoor mobility，continence，and eat⁃
Covariates included baseline age，sex，ethnicity ing. Scoring was based on assistance needs：no assis⁃
（Han or ethnic minority），residence（rural or urban）， tance=3；assistance in one domain=2；assistance in
marital status（married or others），occupation，insur⁃ more than one domain=1.
ance status（insured or not，possession of any of the 1.3 Statistical analysis
eight major insurance types was considered insured）， Variables with missing（<30% missingness）were
living arrangement（living alone vs. others），smoking imputed using the hot⁃deck imputation method，and all
status（never，former or current），alcohol consumption subsequent analyses were performed on the imputed
（never，former or current），sleep duration（<6 h，6-8 h， dataset. Baseline characteristics were compared be⁃
or > 8 h），and depression symptoms. Depression symp⁃ tween the low and high psychological resilience
toms were assessed using two questions：① Have you groups. Continuous variables were presented as mean
felt sad，blue，or depressed for two weeks or more in and standard deviation（x ± s）if normally distributed，
the last 12 months? ② Have you lost interest in most or as median and interquartile range［M（P25，P75）］if non⁃
things，like hobbies，work，or similar activities? Partici⁃ normally distributed. Categorical variables were sum⁃
pants answering“yes”to either question were classi⁃ marized as frequencies and percentages［n（%）］. Group
fied as having depressive symptoms. differences were assessed using the independent sam⁃
We also accounted for multimorbidity（presence ples Student’s t⁃test，Wilcoxon rank⁃sum test，or the
2
of ≥2 chronic conditions），including hypertension，dia⁃ Chi⁃square（χ ）test，as appropriate .
betes，heart disease，stroke，or cardiovascular disease， Restricted cubic spline（RCS）analysis was used
chronic lung diseases（bronchitis，emphysema，pneumo⁃ to visualize the non⁃linear relationship between psycho⁃
nia，asthma），tuberculosis，cataract，glaucoma，cancer， logical resilience and all⁃cause mortality. The associa⁃
gastric/duodenal ulcer，Parkinson’s disease，arthritis， tion between psychological resilience（high or low）and
dementia，epilepsy，cholecystitis/cholelith disease，dys⁃ all ⁃ cause mortality risk was evaluated using univari⁃