Risk assessment in upper gastrointestinal haemorrhage: implications for resource utilisation.

AIMS: Acute upper gastrointestinal (GI) bleeding is a common and serious problem with an associated mortality of some 10%. It is desirable, both for optimising outcomes and for the efficient use of resources, that some form of risk assessment be made early and management be stratified accordingly. A risk scoring system was recently proposed and validated by Rockall and colleagues from the UK. We wished to assess its validity in a New Zealand setting.
METHODS: 565 consecutive patients treated for acute upper gastrointestinal bleeding at Wellington Hospital between 1988 and 1991 were the subject of a major prospective study. All patients were retrospectively assigned a score of 0-7 based on presentation criteria as suggested by Rockall and colleagues. The score is a composite one having regard to age, haemodynamic variables on presentation and associated serious co-morbidity. Correlation was sought between the score and the in hospital mortality risk.
RESULTS: The overall 30-day mortality was 11%. Of the 60.5% of patients with a total score of less than 4 ('low risk'), the group mortality was 3.2%. Of the 39.5% of patients with a total score of 4-7 ('high risk'), the group mortality was 22.4%.
CONCLUSIONS: The scoring system appears as valid in a New Zealand patient population as in the UK. We suggest that this scoring system be adopted in hospitals for patients with acute upper GI bleeding to efficiently direct the use of 'intensive care' type facilities to those most at risk, and thereby optimise management.