A Proposed New Mortality Prediction Tool For Emergency Laparotomy (NMPT)
S Mohamed (1), R Khoju (2), Z Sallami (1), T Atcha (1)
(1) Department of Surgery, Bronglais General Hospital, Aberystwyth, Wales, UK
(2) Department of Anaesthesia, Bronglais General Hospital, Aberystwyth, Wales, UK
• 2nd Patient Report of NELA reported that the estimate of risk of
death provided by P-POSSUM tends to overestimate risk above
Try to find a more accurate peri-operative risk stratification tool to
predict 30-day mortality for those with P-POSSUM expected
mortality of more than 15%.
Data collection and Development of the tool:
Our local emergency laparotomy data from December 2013 to
November 2015 were retrieved from NELA database.
These data were used to calculate the Surgical Outcome Risk Tool
(SORT) for each patient using NCEPOD Application.
We developed a new mortality prediction tool (NMPT) by calculating
the average of SORT and post-operative P-POSSUM for each patient.
We compared the performance of these tools; NMPT, SORT, pre-and
post-operative P-POSSUM with respect to discrimination and calibration.
The new tool NMPT was validated in patients whose P-POSSUM
expected mortality is more than 15%.
Stata was used for
ROC curves and AUROC calculation to evaluate discrimination and
Estimate the P value for the Hosmer–Lemeshow χ2 test and calibration
• This proposed new mortality prediction tool (NMPT) has a good
discrimination and calibration.
• It is better calibrated in patients with P-POSSUM expected mortality
more than 15%.
• Giving the limitation of a small cohort, further studies are required to
assess its performance in larger cohort of patients