Evaluation of the possum and p-possum system of surgical audit in patients undergoing Laparotomy at Kenyatta national hospital
Abstract
The Physiological and Operative Severity Score for the enumeration of Mortality and Morbidity
(POSSUM)and its Portsmouth predictor modification (P-POSSUM) are auditing tools designed
to compare surgical outcomes independent of case mix. They use the patients' physiological and
operativedata to predict morbidity and mortality. In low risk patients however, POSSUM over
predictsmortality. The Portsmouth predictor modification (P-POSSUM) was developed to
overcomethe over prediction of mortality by POSSUM. Evaluation of POSSUM and PPOSSUMsystems
of surgical audit has mainly been done in the U.K. hospitals.
Aims: to determine the applicability of POSSUM and P-POSSUM as audit tools in patients
undergoing laparotomy at Kenyatta National Hospital (KNH).
Methods: A prospective analysis of 166 consecutive patients undergoing laparotomy at KNH,
between October 2006 and February 2007, was done. All data (12 physiological and six
operative factors) collected using a data sheet were analysed. Predicted mortality rates were
calculated using the POSSUM and P-POSSUM equations for mortality, while predicted
morbidity rates were calculated using the POSSUM morbidity equation. The primary outcome
measure was the actual 30- day mortality rate. Secondary outcome measures were the morbidity
seen. Comparisons were then made between the predicted and the observed morbidity and
mortality rates using both linear and exponential methods of analysis. The observed to expected
morbidity and mortality (0: E) ratios were calculated for each risk group. The overall 0: E ratio
for each method of analysis was then obtained. Receiver-operator characteristic curves were
designed to determine the relative accuracy of each approach at prediction. The Hosmer-
Lemeshow goodness of fit test was also calculated.
Results: The observed mortality was 8 patients of 166 patients (4.8%) with an actual morbidity
rate of 52%. The predicted POSSUM mortality was 17.1% while the predicted P-POSSUM
mortality score was 7.4%. The predicted POSSUM morbidity score was 48.1 %. P-POSSUM
showed a statistically significant ability to predict post operative mortality (0: E=0.67:1) while
POSSUM over predicted mortality (0: E= 0.2: 1). POSSUM morbidity score appeared accurate
in predicting actual morbidity (0: E=1 :1). Area Under Curve (AUC) of Receiver Operator
Curves (ROC) of POSSUM mortality and morbidity equation were 0.7440 and 0.6758
respectively while that of P-POSSUM mortality equation was 0.7775 showing good
discrimination of all the three scoring systems