Background: Adults hospitalized with acute, nonvariceal upper GI hemorrhage can be accurately stratified according to their risk of subsequent adverse outcomes by using the Rockall score. Low-risk patients (Rockall score less-than-or-equal 2) may be candidates for early discharge.
Methods: Cases were identified with ICD-9-CM codes for calendar years 1997 and 1998. Medical record data to determine patient Rockall risk score, health care resource utilization, and adverse outcomes were abstracted with standardized forms.
Results: Fifty-three of 175 (30%) cases had Rockall scores < or =2. As predicted, those patients with Rockall scores < or =2 had a low risk of adverse outcomes with only 2 of 53 (4%) meeting criteria for recurrent bleeding as defined by the "Rebleed" variable, and no mortality. These low-risk patients had a mean hospital stay of 2.6 plus minus 2.1 days; 49% were admitted to an intermediate or intensive care unit bed and 57% were given H2 receptor antagonists intravenously.
Conclusions: The proportion of patients admitted with acute, nonvariceal, upper GI hemorrhage with Rockall Scores < or =2 was substantial. Adverse outcomes were rare. In contrast, the level of health care resource utilization appeared high. The Rockall score has potential as a clinically based concurrent decision rule to improve the quality of care by finding those patients less likely to require intensive health care services.