[Referred medical patients not admitted to the Intensive Care Unit: prevalence, clinical characteristics and...

Information on the outcomes of patients who were refused to the ICU is limited. The aims of this study were to compare the clinical characteristics of patients who were admitted with those of patients who were refused to the ICU and to identify clinical parameters associated with triage procedures.

Observational prospective cohort study. The following data were collected using a standard questionnaire: comorbidities, acute illness, vital status, laboratory data and APACHE II score. The end-points of interest were admission to the ICU and vital status at hospital discharge.

A total of 455 patients were studied; 254 (56%) were admitted and 201 (44%) were not. The main reason for the refuse of admission was the lack of ICU beds (82%). Patients who were not admitted had a higher mortality (85% vs. 61%; p < 0.001). In multivariable analysis, the following variables were associated to non-admission [odds ratio, (95% confidence interval)]: metastatic cancer [5.6(1.7-18.7)], arterial systolic pressure < 90 mmHg [5.2(3.0-8.8)], age > 70 years [4.0(2.4-6.5)], hepatic cirrhosis [3.7(1.8-7.6)], and Glasgow coma scale < 5 [3.6(1.9-6.9)]. The variables associated with ICU admission were: mechanical ventilation [0.5(0.3-0.7)] and acute coronary syndromes [0.1(0.03-0.6)].

Refusal of ICU admission is frequent and generally as a consequence of ICU beds shortage. Patients who were not admitted had a higher mortality. Clinical characteristics associated with the refusal of admission were identified suggesting that they are used in clinical decision-making for ICU triage.