Article Text

Download PDFPDF
OC-023 Can we predict length of stay in patients admitted with acute upper gastrointestinal bleeding?
  1. V Jairath1,
  2. R Logan2,
  3. S Hearnshaw3,
  4. S Travis4,
  5. M Murphy5,
  6. K Palmer6
  1. 1NHS Blood and Transplant Unit, John Radcliffe Hospital, Oxford, UK
  2. 2Division of Epidemiology, University of Nottingham, Nottingham, UK
  3. 3Department of Gastroenterology, Royal Infirmary, Newcastle, UK
  4. 4Department of Gastroenterology, John Radcliffe Hospital, UK
  5. 5NHS Blood and Transplant, John Radcliffe, Oxford, UK
  6. 6GI Unit, Western General Hospital, Edinburgh, UK


Introduction Acute upper gastrointestinal bleeding (AUGIB) is a common cause of admission to hospital in the UK with an annual incidence of 50–150 per 100 000 cases and a stubbornly high mortality of around 10–14%. Thus AUGIB represents a huge economic burden to the NHS. Enabling clinicians and bed managers to predict the likely length of stay of those patients admitted could aid triage to an appropriate hospital area and facilitate the multidisciplinary team in discharge planning. We analysed if length of hospital stay can be predicted from the clinical and/or full Rockall score.

Methods Further analyses were performed on patients included in the 2007 UK Comparative Audit of Upper Gastrointestinal Bleeding and the Use of Blood.1 In this study prospective data were collected electronically on consecutive patients presenting to all UK hospitals with AUGIB between 1 May and 30 June 2007.

Results Data were available on 6750 patients across 208 UK hospitals. We restricted this analysis to new admissions with AUGIB (5550/6750). A clinical Rockall score was available for 5550 patients and a full (post endoscopy) Rockall score for 4095 patients. We analysed the median length of stay by clinical and full Rockall score for all new patients admitted with AUGIB. The median length of stay for a clinical Rockall score of 0–1 was 3 days. The median length of stay for a full Rockall score of zero was 2 days- each one point increase in the full Rockall score corresponded to one extra day in hospital. Non-parametric, two-tailed Spearman coefficient calculation demonstrated the correlation between Rockall score and length of stay was highly significant for both the clinical score (r=0.291, p<0.01) and full score (r=0.313, p<0.01).

Conclusion The Rockall score was originally developed as a tool to predict mortality from AUGIB. In this retrospective analysis, we have shown that both the clinical and full Rockall score can be used to predict length of hospital stay for patients presenting with AUGIB. We have also demonstrated a median length of hospital stay of 3 days for patients with a clinical Rockall score of 0–1 which could be deemed excessive given recent further evidence supporting discharge followed by early outpatient endoscopy for such low risk patients.2

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.