Predictors of advance directive restrictiveness and compliance with institutional policy in a long-term-care facility

To ascertain factors influencing the level of advance directives selected by nursing home residents or surrogates and the time delay to documentation of these choices in the medical record after implementation of a facility-wide policy.

Longitudinal cohort study of nursing home residents followed from date of advance directive policy initiation or time of admission for a maximum of 21 months from study commencement.

A 315-bed multilevel nursing home.

Four hundred twenty-four nursing home residents (mean age 85, 74.9% female, 96.1% white).

Level of advance directive status chosen--full code, do not resuscitate (DNR) or palliative care only--and date documented in the medical record.

Factors predictive of restricted advance directives (DNR or palliative care) included age greater than 85 years (P = 0.025), documented use of a surrogate decision maker (P = 0.001), low physical function (P less than 0.001), low cognitive function (P less than 0.001), and having a nursing home-employed physician (P = 0.001). These results were confirmed using logistic regression models. Median time to directive documentation decreased from 54 days for residents admitted in the first quarter to 1 day for residents admitted in the fourth quarter of the year following initiation of an advance directive policy.

In logistic models, nursing home-employed physicians were more likely to write restricted advance directive orders than community-based physicians even after controlling for resident age, cognitive status, and physical function. In addition, implementation of a formal nursing home advance directive policy can shorten time to physician documentation of resident advance directive status.