Quality of resuscitation orders in general medical patients.

BACKGROUND Documentation of resuscitation status in hospitalized patients has relevance in the management of cardiopulmonary arrest. Its association with mortality, Length Of hospital Stay (LOS) and the patients' primary diagnosis has not been established in general medical inpatients in hospitals in Australia and New Zealand. AIM To investigate the association of resuscitation orders with in-hospital mortality and LOS in a range of diagnoses, adjusting for severity of illness and other covariates. DESIGN Retrospective study. METHODS The admission notes of 1681 medical admissions to four tertiary care teaching hospitals across Australia and New Zealand were reviewed retrospectively for frequency and nature of resuscitation documentation and its association with mortality, LOS and primary diagnosis. RESULTS Resuscitation orders were documented in 741 patients (44.7%). For the 232 patients with a Not For Resuscitation (NFR) order, the in-hospital mortality rate was higher than in control patients (14% vs. 1.2%, P<0.005). The mortality rate remained significantly higher in the NFR group after propensity matching of the controls for age and co-morbidity (14% vs. 5%, P<0.005). The death-adjusted LOS for the NFR group was also significantly higher compared to the control patients (9.7 days vs. 4.7 days, P<0.005) and this difference remained after propensity matching (9.7 days vs. 7.7 days, P<0.05). Those patients with a primary diagnosis of respiratory tract infection or cardiac failure were more likely to be documented NFR compared to those with cellulitis or urinary tract infection. CONCLUSIONS The documentation of NFR in a patient's admission notes is associated with increased in-hospital mortality and LOS. This is only partly explicable in terms of these patients' greater age and co-morbidity.