Older adults may have substantial changes in health over time. We examine the stability of older adults′ utilities for hypothetical health states of dependency in Activities of Daily Living(ADLs) over a year as a function of change in health status.

126 medical outpatients(mean age=75, 65% female, 35% non-white) reported utilities at Time 1 and one year later(Time 2) for current health(CH) and for hypothetical health states of dependency in 6 ADLs, a combination of all 6 ADLs(All6), a combination of 3 ADLs while living in a Nursing Home(NH) or their Own Home(OH), and Walking. Total number of medical conditions was added at Time 1 and 2. Difference scores for baseline and follow-up utilities were computed.

51 participants reported more medical conditions(worse health) , 44 reported fewer medical conditions(improved health), and 31 reported no change. Time 2 CH utilities were .84 for no change, .83 for improved health, and .77 for worse health. Time 2 mean utilities for single ADL states ranged from .81(Bathing) to .70(Toileting) and were .52 for All6, .53 for NH, .66 for OH, and .73 for Walking. Mean differences ranged from .005(Bathing) to .111(Walking) and were .045 for All6, .015 for NH, .043 for OH, and .027 for Walking.

Analyses of variance revealed significant effects for the lowest rated ADLs, Toileting, F(2,123)=2.93, p&lt.05 and Walking, F(2,121)=3.43, p&lt.05. Significant effects were also found for All 6, F(2,122)=4.125, p&lt.05, NH, F(2,121)=5.40, p&lt.01, and OH, F(2,121)=4.13, p&lt.05. Tukey's post-hocs revealed no difference between no change and improved health. Worse health showed a significantly greater difference over time than no change for All 6, NH, and OH, p&lt.05 and a trend toward greater difference for Toileting, p=.064. Improved health differences were significantly greater than worse health for Walking, p&lt.05.

Change in health over time appears to affect older adults′ health preferences when rating even hypothetical health states described identically over time. Worsened health status may color an individual's valuation of hypothetical health states. Policy analyses drawing on utilities should ensure that a broad range of health status was included in the population providing the utilities. In individual medical decision-making tasks, health professionals should ensure that patients have a clear grasp of the outcome scenarios that is not unduly influenced by their own recent changes in health.