Objective: To explore the concept of futility by asking clinicians for estimates of survival and admission decisions for an intensive care unit patient with little chance of survival, and to compare these estimates with results from an intensive care database.
Design: Questionnaire based on the presenting features of a genuine patient. It asked for estimated hospital survival, decision on intensive care admission, resuscitation status and importance of family views. Analysis of a regional intensive care database.
Setting: Physicians working in British intensive care units.
Participants: We received 169 replies, 146 from consultants.
Measurements and results: Median estimated hospital survival was 5%; 60% of consultants and 76% of trainees would have admitted the patient, with 9% and 14%, respectively, prepared to perform further cardiopulmonary resuscitation. Among those estimating survival probability as less than 1%, 17.2% would have admitted the patient. Family opinions were vital to 4.3% of respondents and unimportant to 9.8%. There were 251 patients in the database with similar physiological derangements. Their observed hospital mortality was 91%. At intensive care admission an admitting physician assessed 111 of these patients as 'expected to die'. Mortality in this group was 99.1% (one survivor).
Conclusions: Experienced intensivists did not agree on estimated survival. Even when estimates agreed, admission decisions varied. Database analysis suggested that clinical judgement is relevant when assessing the risk of dying. Lack of consensus on survival estimates and admission decisions suggests that it would be difficult to achieve agreement on appropriate use of intensive care resources and on what constitutes futile treatment.