The Physiological and Operative Severity Score for enUmeration of Mortality and morbidity (POSSUM) scoring system has been proposed as an accurate predictor of death which takes account of case mix. It appears to overpredict death, and may have drawbacks which prevent accurate individual or subgroup analysis. An alternative system, the Portsmouth predictor equation for mortality (P-POSSUM) may have overcome these problems, but its apparent advantage could be related to inappropriate analysis of POSSUM predictions.
Some 312 patients having arterial procedures were studied. POSSUM and P-POSSUM scores were used to predict death and compared with actual outcomes. The observed:predicted (O:E) mortality ratios were calculated by two methods for each of the scoring systems. First the analysis devised by the inventors of POSSUM was used and second the method devised for P-POSSUM.
The O:E ratio for POSSUM using its recommended ‘exponential’ method of analysis was 1·14, but it was 0·59 if the P-POSSUM ‘linear’ analysis was used. The O:E ratio for P-POSSUM using its correct method of analysis was 0·89, but it was 0·67 if the method of analysis devised for POSSUM was used.
The O:E ratios for POSSUM and P-POSSUM were close to unity when the appropriate analysis was performed. Both POSSUM and P-POSSUM overpredicted death if the incorrect analysis was used. © 1998 British Journal of Surgery Society Ltd