2007 HSR&D National Meeting Abstract
3039 — The Stability of Self-Reported Preferences in VA Nursing Home Residents
Housen P (VA COE) , Shannon G
(VA COE), Orlando M
(RAND), Simon B
(VA COE), Cadogen M
(UCLA School of Nursing), Sohn L
(VA COE), Jones M
(RAND), Buchanan J
(Harvard University), Saliba D
(VA COE, RAND)
Understanding preferences for daily routines aids post-acute providers and nursing facilities in individualizing care plans to improve quality of life for functionally dependent patients. However, few empirically tested tools for assessing preferences exist, and the stability of these preferences has not been established. Previous pilot tests of a Preference Assessment Tool (PAT) showed acceptable response and completion rates among post-acute and long-stay cognitively intact and moderately impaired residents. A VA National Steering Committee subsequently requested longitudinal follow-up with the PAT to determine whether preferences should be assessed more often than at admission only.
Participants were residents of two Southern California VA NHs. We used the PAT, a 24-item interviewer administered survey, to establish baseline preferences (n=142). 72-hour re-test was performed, and the PAT survey was re-administered within six months of initial interview to all residents from the baseline sample who remained in the facilities. We used Cohen’s kappa statistic and percentage agreement as criteria for item-level stability. Cognitive status was assessed using the Cognitive Performance Scale (CPS), a measure highly correlated with the Mini-Mental Status Examination.
74 residents were available for the 3-6 month follow-up interview. Participants in the subset were average age 75 years (S.D.=11 years) and 27% non-white. 72% were cognitively intact, or only mild to moderately impaired (CPS<=3). 72-hour reliability testing showed that 19 of the 24 items were at least moderately reliable (kappa>=0.41; agreement>=75%). However, only 11 of the 24 items were stable over the longer period of several months. Preferences related to the domain of resident autonomy were particularly unstable; only 2 of 7 items relating to perceived control over daily routines such as bedtime, bath time, dressing, and appearance remained constant.
Our findings support the hypothesis that routine care preferences are subject to change. Fluctuations may reflect changes in clinical condition or accommodation to functional limitations and facility environment.
The results suggest resident preferences should be systematically assessed more frequently than the current practice of at admission only. Next steps should include testing the instrument and variables associated with change in self-reported preferences at facilities in other regions.