Rate of use and determinants of withdrawal of care among patients with subarachnoid hemorrhage in the United States.

Abstract

The use of "withdrawal of care" and impact upon outcomes among patients with subarachnoid hemorrhage (SAH) is not well studied. To identify the rate and determinants of "withdrawal of care" among SAH patients. We determined the frequency of "withdrawal of care" and compared the demographic, clinical characteristics, and in-hospital outcomes among patients with SAH stratified by use of "withdrawal of care." "Withdrawal of care" during hospitalization was instituted in 8912 (3.4%) of the 266,067 patients with SAH. In the stepwise logistic regression, age >65 (odds ratio [OR] 4.5, 95% confidence interval [95% CI] 3.3-6.1), women (OR 1.2, 95% CI 1.0-1.3), African American (OR 0.7, 95% CI 0.5-0.8), Hispanic ethnicity (OR 0.4, 95% CI 0.3-0.6), renal failure (OR 1.6, 95% CI 1.2-1.9), intracerebral hemorrhage (OR 2.0, 95% CI 1.7-2.4, All Patient Refined Diagnosis-Related Groups severity score of extreme loss of function (OR 40.1, 95% CI 6.0-270.7), All Patient Refined Diagnosis-Related Groups severity score of severe loss of function (OR 15.0, 95% CI 2.1-103.8), insurance status of private health maintenance organization (OR 0.7, 95% CI 0.5-0.9), and hospital region south United States (OR 0.7, 95% CI 0.5-0.8), were significant predictors of "withdrawal of care" among patients with SAH. In-hospital mortality was significantly greater, but mean hospitalization charges and length of stay were significantly lower among those with "withdrawal of care." Although "withdrawal of care" was effective in limiting hospital charges and resource use, caution is needed to avoid disproportionately high mortality. The prominent relationship between race/ethnicity, insurance status, and hospital location with "withdrawal of care" raises concerns that factors other than severity of disease influence decision making.