Decision-making capacity to execute a health care proxy: development and testing of guidelines

To evaluate the reliability and validity of guidelines to determine the capacity of nursing home residents to execute a health care proxy (HCP).

A cross-sectional study.

A 750-bed not-for-profit nursing home located in New York City.

A random sample of 200 nursing home residents: average age, 87; 99% white; 83% female; average length of stay, 3.05 years; mean Mini-Mental State Exam (MMSE) score, 15.9.

Demographic characteristics (Minimum Data Set (MDS)); function and cognitive status (Institutional Comprehensive Assessment and Referral Evaluation (INCARE)); Reisberg Dementia Staging; MMSE; Minimum Data Set-Cognitive Performance Scale (MDS-COGS)); an investigator-developed measure of a nursing home resident's capacity to execute a health care proxy (Health Care Proxy (HCP) Guidelines.)

The internal consistency of the decision-making scales in the HCP Guidelines, paraphrased recall and recognition, reached acceptable levels, alphas of .85 and .73, respectively. Interrater reliability estimates were .92 and .94, respectively, for the recall and recognition scales; test-retest reliability estimates were .83 and .90. The discriminant validity of these scales is promising. For example, the MMSE correlation was .51 with the Recall scale and .57 with the Recognition scale. Of residents with severe cognitive impairment (MMSE < 10), 71% completed 50% or more of the scaled items in the HCP guidelines and 95% consistently named a proxy.

Seventy-three percent of testable residents, approximately three-quarters of whom were cognitively impaired, evidenced sufficient capacity to execute an HCP. Of residents with severe cognitive impairment, the HCP guidelines are potentially useful in identifying those with the capacity to execute a HCP. The guidelines are more predictive than the MMSE in identifying residents able to execute a HCP.