LAUSR

OBJECTIVES Predictive effect of self-rated health (SRH) on mortality in older adults has been observed. The purpose of the study was to analyze this association in Poles aged 65+. METHODS Data were obtained from the nationwide, multidisciplinary PolSenior project, conducted in a representative sample of older population. The study group comprised 4049 respondents (48.0% women) without significant cognitive deficit. SRH was measured using Visual Analog Scale. The analysis included selected socio-economic, health status and life-style factors. Mortality data were retrieved from the state registry. RESULTS During 5-year period, 414 women (21.4%) and 672 men (31.8%) have died, including 17.5% of women and 26.6% of men with good, 21.6% and 32.9% with fair, 36.2% and 55.3% with poor SRH, respectively. Kaplan-Meier survival curves for SRH revealed significant differences for both genders. Univariate Cox regression analysis revealed significant hazard ratios (HRs) for mortality among women and men with poor compared to good SRH [2.48 (1.83-3.37); 2.62 (2.04-3.36), respectively] and those with fair compared to good SRH [1.29 (1.03-1.60); 1.29 (1.10-1.52), respectively]. Age-adjusted HRs for mortality were significant between groups with poor and good SRH [women: 1.98 (1.46-2.68), men: 2.06 (1.60-2.64)]. Multivariate Cox proportional hazard regression model including revealed significant HRs for mortality between women with poor and good SRH [1.67 (1.06-2.64)]. CONCLUSIONS SRH was associated with mortality in both genders. After adjustment for age, this relationship was maintained in respondents with poor compared to good SRH. Inclusion of potential confounders demonstrated that SRH was an independent predictor of mortality only in women.