Skip to main content
. 2015 Jul 9;15:78. doi: 10.1186/s12877-015-0074-4

Table 4.

Hazard ratios for mortality, by baseline sociodemographic characteristics and health conditions of those residents with valid mortality data

Baseline variable No. of subjects a (%) No. of person-years No. of deaths Hazard ratio b (95 % CI)
Overall 699 (100) 4,134.6 598
Age (years)
 65–74 95 (12.7) 774.9 68 1.00 (reference)
 75–84 287 (39.9) 1,904.7 248 1.48 (1.11–1.99)
 ≥85 317 (47.4) 1,455.0 282 2.19 (1.52–3.16)
Sex
 Women 386 (75.7) 2,343.0 331 1.00 (reference)
 Men 313 (24.3) 1,791.6 267 1.11 (0.94–1.31)
Type of facility
 Public 401 (47.0) 2,348.4 357 1.00 (reference)
 Subsidized 72 (8.0) 369.3 60 1.14 (0.90–1.44)
 Private 226 (45.0) 1,416.8 181 0.84 (0.66–1.08)
No. of chronic conditions
 0–1 151 (21.6) 1,038.6 125 1.00 (reference)
 2–3 290 (42.0) 1,851.0 245 0.99 (0.77–1.26)
 ≥4 258 (36.4) 1,245.0 228 1.21 (0.96–1.54)
Functional dependency
 Independent 187 (22.1) 1,465.7 154 1.00 (reference)
 Mild/moderate 316 (47.0) 1,910.9 270 1.37 (1.14–1.63)
 Severe/total 179 (28.4) 688.9 157 2.27 (1.78–2.89)
 Unknown 17 (2.5)
Cognitive impairment
 Unimpaired 297 (41.6) 2,069.6 246 1.00 (reference)
 Mild/moderate 169 (26.8) 798.1 145 1.38 (1.05–1.81)
 Severe 79 (14.0) 283.1 70 2.11 (1.45–3.05)
 Unknown 154 (17.6)
Self-rated health
 Very good/good 336 (48.1) 2,309.3 274 1.00 (reference)
 Fair 197 (26.4) 1,137.0 174 1.28 (1.07–1.53)
 Poor/very poor 95 (13.5) 443.8 85 1.58 (1.17–2.12)
 Unknown 71 (12.0)
Physician-rated health
 Very good/good 329 (42.7) 2,291.5 275 1.00 (reference)
 Fair 222 (31.1) 1,059.7 196 1.41 (1.15–1.73)
 Poor/very poor 69 (10.3) 289.6 61 1.55 (1.02–2.35)
 Unknown 79 (15.9)
Objective health c
 Good 40 (4.4) 367.3 29 1.00 (reference)
 Intermediate 199 (31.2) 1,376.1 161 1.27 (0.84–1.91)
 Poor 351 (51.6) 1,631.5 310 1.83 (1.27–2.63)
 Unknown 109 (12.8)

a Unweighted sample counts and weighted percentages based on the underlying population distribution

b Hazard ratios and 95 % confidence intervals (CIs) were obtained from Cox models, with years from the baseline interview as the time scale, adjusted for baseline age, sex, and type of facility, taking into account the stratified cluster sampling and the different selection probabilities

c Residents presenting with 0–1 chronic conditions, functionally independent in basic activities of daily living, and having unimpaired cognition were assigned to good objective health; those presenting with ≥ 4 chronic conditions, severe/total functional dependency, or severe cognitive impairment were assigned to poor objective health; and the remaining residents presenting with intermediate severity levels were assigned to fair objective health