BACKGROUND Scoring systems such as POSSUM and P-POSSUM have been developed to help predict mortality and morbidity in patients. The ratio of observed-to-predicted (O/P ratio) mortality and morbidity has been used as a performance indicator to compare different procedures, clinicians or hospitals. The aim of this study was to assess the predictive value of POSSUM compared with P-POSSUM in patients undergoing laparotomy in Queen Elizabeth Hospital, Sabah. METHODS 381 patients over the age of 12 undergoing general surgical laparotomy between 1 May 2006 and 30 April 2007 were prospectively enrolled. RESULTS In general, POSSUM over-predicted mortality compared with P-POSSUM (O/P ratio: 0.366 versus 0.721). P-POSSUM was still poor at predicting mortality in the lowest and highest risk groups (O/P ratios: 0 and 0.438). Both systems over-predicted mortality in trauma (O/P ratios: POSSUM, 0.306; P-POSSUM, 0.459), younger patients (O/P ratios: POSSUM, 0.325; P-POSSUM, 0.622) and non-ICU patients (O/P ratios: POSSUM, 0.171; PPOSSUM, 0.421). P-POSSUM was significantly better than POSSUM in other age groups and ICU patients. In terms of morbidity, POSSUM was able to predict morbidity quite accurately with O/P ratio of 0.746 and performed equally well across the subgroup. POSSUM was poor in predicting morbidity in young patients (O/P ratio: 0.652) and non-ICU patients (O/P ratio: 0.543). CONCLUSION P-POSSUM is a better overall predictor of mortality in patients undergoing laparotomy in this hospital compared to POSSUM. POSSUM is fairly accurate in predicting morbidity. However, further refinement is needed to improve its predictive value in specific areas, and so increase its utility in our local setting.