Stability and change in structural social relations as predictor of mortality among elderly women and men

Citation
R. Lund et al., Stability and change in structural social relations as predictor of mortality among elderly women and men, EUR J EPID, 16(12), 2000, pp. 1087-1097
Citations number
47
Categorie Soggetti
Envirnomentale Medicine & Public Health","Medical Research General Topics
Journal title
EUROPEAN JOURNAL OF EPIDEMIOLOGY
ISSN journal
03932990 → ACNP
Volume
16
Issue
12
Year of publication
2000
Pages
1087 - 1097
Database
ISI
SICI code
0393-2990(2000)16:12<1087:SACISS>2.0.ZU;2-2
Abstract
In a follow-up study of 70-95years old women and men (n = 911) we studied t he association between change and stability in three structural aspects of social relations (contact frequency, contact diversity, cohabitation status ) from 1986-1990 and mortality after the next four years in 1994. Women age d 70-74 years who developed low contact frequency or developed small contac t diversity showed significantly higher mortality, adjusted ORfreq: 3.78 (1 .08-13.20), adjusted ORdiv: 3.79 (1.24-11.58). Women aged 70-74 years with continuously low contact frequency showed an increased mortality compared t o women constantly experiencing high contact frequency, adjusted OR: 2.75 ( 1.04-7.26). A tendency in the same direction for sustained small contact di versity was found, adjusted OR: 1.98 (0.70-5.61). Among women aged 75+ year s no impact of frequency and diversity was demonstrated, whereas continuous ly living alone was a significant predictor of mortality, when compared to women continuously living with somebody, adjusted OR: 2.57 (1.29-5.09). In men, we found a significantly increased mortality among those who developed high contact frequency and developed large contact diversity ORfreq: 3.91 (1.02-14.94) and ORdiv: 6.04 (1.30-28.03). In summary, we found rather larg er age differences in the strength of the association between change in str uctural social relations and mortality. Furthermore, the associations seeme d stronger among women than men, which may however mainly be explained by t he small number of men in our cohort.