Send to

Choose Destination
Lancet Gastroenterol Hepatol. 2017 Sep;2(9):635-643. doi: 10.1016/S2468-1253(17)30150-4. Epub 2017 Jun 23.

Derivation and validation of a novel risk score for safe discharge after acute lower gastrointestinal bleeding: a modelling study.

Author information

National Health Service (NHS) Blood and Transplant, Oxford, UK; National Heart and Lung Institute, Imperial College, London, UK.
Division of Epidemiology and Biostatistics, Western University, London, ON, Canada; Department of Medicine, University of Western Ontario, London, ON, Canada. Electronic address:
Department of Interventional Radiology, Oxford University Hospitals, NHS Foundation Trust, Oxford, UK.
Department of Colorectal Surgery, Oxford University Hospitals, NHS Foundation Trust, Oxford, UK.
Department of Gastroenterology, Charing Cross and Hammersmith Hospitals, Imperial College Healthcare, NHS Trust, London, UK.
National Health Service (NHS) Blood and Transplant, Oxford, UK; Department of Haematology, Oxford University Hospitals, NHS Foundation Trust, Oxford, UK; Department of Transfusion Medicine, University of Oxford, Oxford, UK.
Centre for Statistics in Medicine, Nuffield Department of Orthopaedics, Botnar Research Centre, Oxford, UK.



Acute lower gastrointestinal bleeding is a common reason for emergency hospital admission, and identification of patients at low risk of harm, who are therefore suitable for outpatient investigation, is a clinical and research priority. We aimed to develop and externally validate a simple risk score to identify patients with lower gastrointestinal bleeding who could safely avoid hospital admission.


We undertook model development with data from the National Comparative Audit of Lower Gastrointestinal Bleeding from 143 hospitals in the UK in 2015. Multivariable logistic regression modelling was used to identify predictors of safe discharge, defined as the absence of rebleeding, blood transfusion, therapeutic intervention, 28 day readmission, or death. The model was converted into a simplified risk scoring system and was externally validated in 288 patients admitted with lower gastrointestinal bleeding (184 safely discharged) from two UK hospitals (Charing Cross Hospital, London, and Hammersmith Hospital, London) that had not contributed data to the development cohort. We calculated C statistics for the new model and did a comparative assessment with six previously developed risk scores.


Of 2336 prospectively identified admissions in the development cohort, 1599 (68%) were safely discharged. Age, sex, previous admission for lower gastrointestinal bleeding, rectal examination findings, heart rate, systolic blood pressure, and haemoglobin concentration strongly discriminated safe discharge in the development cohort (C statistic 0·84, 95% CI 0·82-0·86) and in the validation cohort (0·79, 0·73-0·84). Calibration plots showed the new risk score to have good calibration in the validation cohort. The score was better than the Rockall, Blatchford, Strate, BLEED, AIMS65, and NOBLADS scores in predicting safe discharge. A score of 8 or less predicts a 95% probability of safe discharge.


We developed and validated a novel clinical prediction model with good discriminative performance to identify patients with lower gastrointestinal bleeding who are suitable for safe outpatient management, which has important economic and resource implications.


Bowel Disease Research Foundation and National Health Service Blood and Transplant.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center