Send to

Choose Destination
Resuscitation. 2004 Aug;62(2):137-41.

Association between clinically abnormal observations and subsequent in-hospital mortality: a prospective study.

Author information

Department of Intensive Care, Dandenong Hospital, P.O. Box 478, Dandenong, Vic. 3175, Australia.



Patients with unexpected in-hospital cardiac arrest often have an abnormal clinical observation prior to the arrest. Previous studies have suggested that a medical emergency team responding to such patients may decrease in-hospital mortality from cardiac arrest, but the association between any abnormal clinical observation and subsequent increased mortality has not been studied prospectively. The aim of this study was to determine the predictive value of selected abnormal clinical observations in a ward population for subsequent in-hospital mortality.


Prospective data collection in five general hospital ward areas at Dandenong Hospital, Victoria, Australia.




During the study period, 6303 patients were admitted to the study areas. Of those, 564 (8.9%) experienced 1598 pre-determined clinically abnormal events and 146 of these patients (26%) died. The two commonest abnormal clinical events were arterial oxygen desaturation (51% of all events), and hypotension (17.3% of all events). Using a multiple linear logistic regression model, there were six clinical observations which were significant predictors of mortality. These were: a decrease in Glasgow Coma Score by two points, onset of coma, hypotension (<90 mmHg), respiratory rate <6 min(-1), oxygen saturation <90%, and bradycardia >30 min(-1). The presence of any one of the six events was associated with a 6.8-fold (95% CI: 2.7-17.1) increase in the risk of mortality.


Six abnormal clinical observations are associated with a high risk of mortality for in-hospital patients. These observations should be included as criteria for the early identification of patients at higher risk of unexpected in-hospital cardiac arrest.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center