Table 2. Association Between Characteristics of Nursing Home Residents With Advanced Dementia and Their Proxies and the Proxy’s Perception That the Resident Had Less Than 6 Months to Livea.
| Characteristic | Assessment Intervals With Characteristica (n = 2526), No. (%) |
Assessment Intervals in Which Proxy Estimated Resident Had <6 Months to Live (n = 309), No. (%) |
Unadjusted Odds Ratiob for Proxy-Perceived Prognosis (95% CI) | |
|---|---|---|---|---|
| With Characteristic Present | With Characteristic Absent | |||
| Resident | ||||
| Age >87 y (median) | 1178 (46.6) | 164 (6.5) | 145 (5.7) | 1.34 (0.96-1.88)c |
| Female | 2110 (83.5) | 248 (9.8) | 61 (2.4) | 0.78 (0.50-1.21) |
| White | 2272 (89.9) | 279 (11.1) | 30 (1.2) | 1.05 (0.56-1.96) |
| Alzheimer disease | 1813 (71.8) | 216 (8.6) | 93 (3.7) | 0.90 (0.62-1.31) |
| Chronic obstructive pulmonary disease | 264 (10.5) | 43 (1.7) | 266 (10.5) | 1.46 (0.92-2.32) |
| Congestive heart failure | 387 (15.3) | 52 (2.1) | 257 (10.2) | 1.14 (0.71-1.81) |
| Diabetes | 454 (18.0) | 56 (2.2) | 253 (10.0) | 1.01 (0.66-1.56) |
| TSI = 0 | 1315 (52.1) | 165 (6.5) | 144 (5.7) | 1.06 (0.76-1.49) |
| BANS-S >21 (median) | 996 (39.4) | 129 (5.1) | 180 (7.1) | 1.12 (0.80-1.56) |
| Any hospital transfer in prior 3 mod | 100 (4.0) | 17 (0.7) | 292 (11.6) | 1.50 (0.86-2.60) |
| Proxy | ||||
| Age >61 y (median)e | 1186 (47.4) | 158 (6.3) | 149 (6.0) | 1.20 (0.86-1.69) |
| Female | 1613 (63.9) | 224 (8.9) | 85 (3.4) | 1.57 (1.10-2.24)c,f |
| Child of resident | 1565 (62.0) | 200 (7.9) | 109 (4.3) | 1.15 (0.80-1.65) |
| Asked their opinion about goals of care by a nursing home clinician | 1126 (44.6) | 183 (7.2) | 126 (5.0) | 1.96 (1.52-2.54)c,g |
Abbreviations: BANS-S, Bedford Alzheimer Nursing Severity-Subscale (range, 7-28; higher scores indicate greater functional disability); TSI, Test for Severe Impairment (range, 0-24; lower scores indicate greater cognitive impairment).
Analyses were done at the level of assessment intervals, which included all baseline and follow-up resident/proxy assessment intervals (n = 2526). Resident medical chart reviews and proxy interviews were done at baseline and quarterly for up to 12 months. Static variables were brought forward from baseline. Proxy’s perception of prognosis and other dynamic variables (eg, goals of care discussion, hospital transfers) were ascertained from each assessment period. Proxies stated the resident had less than 6 months to live at 12.2% of all baseline and follow-up assessment intervals (309 of 2526).
Unadjusted and adjusted odds ratio accounted for clustering among resident/proxy dyads using generalized estimating equations.
Variables that were significant at P < .10 in bivariable analyses and entered into the multivariable model.
Hospital transfer included hospitalization or emergency department visit.
Age missing for 24 proxies.
Adjusted odds ratio, 1.55 (95% CI, 1.09-2.20).
Adjusted odds ratio, 1.94 (95% CI, 1.50-2.52).