POSSUM in Colorectal Surgery

Dermot Burke, The General Infirmary at Leeds

Powerpoint File

You cannot be a modern practicing surgeon in the UK without being aware of the concept of audit. Audit has been a central part of surgical life for approximately 20 years. At first it was used to influence local practice, then, with NCEPOD, to influence national practice. Now, with the publication of performance league tables, comparisons are made between both trusts and surgical units. Yet these comparisons rely upon raw data with no allowance made for patient case-mix. This results in inappropriate comparisons and the potential to make inaccurate inferences. The practice of risk analysis attempts to remedy this by allowing for variability in case-mix.

The APACHE (Acute Physiology And Chronic Health Evaluation) score was an early attempt to categorise patients and stratify their risk of mortality. This was modified to become the APACHE II, and again to become the APACHE III score. However these scores were developed using data from both medical and surgical patients who had been admitted to intensive care. Their calculation relies upon data collected from a large number of variables over a 24 hour period. They take no account of operative severity. Using APACHE scores as risk assessment for surgical patients therefore, meant that it was difficult to gather the data in the first place and that no account was paid to the magnitude of the surgical insult. For these reasons it was thought that a method of assessing surgical risk, which was both simple to calculate and took account of the operative findings, was needed.

The POSSUM (Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity) score is such a method. POSSUM was developed by Copeland et al and described in 1991(1). They identified 62 patient variables that they considered important and, using multivariate analysis, reduced these to 18 significant variables. Prospective analysis of these variables on 1372 patients (833 elective, 539 emergency) gave rise to the POSSUM equations for mortality and morbidity. These equations were to be used to enable the surgeon to predict the overall mortality and morbidity in patients who had undergone surgery. It is important to note that that the prediction applies to the overall group (e.g. 0-10% predicted mortality in 300 patients) and does not give an individual patient’s risk. Subsequent analysis of data from Portsmouth suggested that mortality was overpredicted by the original POSSUM equation, particularly in patients at low risk. The Portsmouth group suggested a reworking of the original POSSUM mortality equation using their own data, but using the same variables as in the POSSUM equation. This became known as the P-POSSUM equation(2). This, the P-POSSUM equation, predicts mortality only, the Portsmouth group deciding that the definition of morbidity was too uncertain to allow a proper analysis. It appears to be a better predictor of mortality than the original POSSUM equation.

Subsequent application of the POSSUM and P-POSSUM equations to patients with general surgical(3) and gastrointestinal(4) conditions respectively confirm that these equations, when applied to a group of surgical patients, have a good predictive value for morbidity and mortality for the group, not for the individual. The papers by Sagar et al(3) and Tekkis et al(4) exemplify the value of the original work by Copeland, as they compare results between surgical teams. Sagar et al(3) compared surgical units in different hospitals. These units had quite different crude mortality and morbidity rates. Although numbers in this study are small, the observed:expected mortality ratios in both hospitals were similar when adjustment was made using POSSUM. Similarly Tekkis et al(4) compared results between gastrointestinal surgeons within the same unit and found that the use of P-POSSUM allowed better comparisons to be made. This paper suffers slightly form the collection of POSSUM data retrospectively – it cannot be certain to be accurate. However, the concept – of comparing outcome while allowing for case-mix – is what Copeland et al had in mind originally.

Much of the literature regarding the use of the POSSUM equations has looked at general surgical patients – the group originally studied by Copeland et al. Since his publication in 1991, subspecialisation has proceeded rapidly within general surgery. The paper by Tekkis et al(5) restricts the study population to patients with colorectal disease and finds that the POSSUM equations work well in predicting outcome. Subsequent work has suggested that there should be a separate risk prediction score for each subspecialty arena of surgery e.g. V-POSSUM (vascular), O-POSSUM (oesophageal). Thus we have an attempt at a risk prediction score specific to colorectal surgery (both benign and malignant) the CR-POSSUM. This is a very recent innovation and has not been widely tested. It is likely that the measurement of CR-POSSUM, or something similar, will become standard practice for colorectal surgeons.

POSSUM equations have therefore been shown to be quite accurate in predicting surgical risk in a population of patients. Presently it would be better to use POSSUM to assess morbidity risk and P-POSSUM to assess mortality risk. The future will certainly see the assessment of a colorectal specific POSSUM. This will, therefore, be very useful as the practising colorectal surgeon is concerned, as this approach seems to be a simple to use in daily colorectal practice. Should disaster strike and the “men in grey suits” knock on our door, demanding to know why our results are so poor, then production of the appropriate POSSUM data (P-POSSUM, CR-POSSUM, son of CR-POSSUM) should suggest that it is the case-mix to blame……..not our surgery………or not!

PRACTICALITIES

Calculating POSSUM

Calculate P-POSSUM and CR-POSSUM directly at http://www.riskprediction.org.uk

Downloading POSSUM

Download ACCESS 97 database from www.edu.rcsed.ac.uk/lectures/lt1.htm

Download e-POSSUM for use on Palm PDA and Pocket PC at www.gaspalm.co.uk/

EQUATIONS (WHERE R=RISK OF MORTALITY)

POSSUM I[R/(1-R)] = -7.04 + (0.13*physiology score) + ( 0.16*operative severity score)

P-POSSUM I[R/(1-R)] = -9.065 + (0.1692*physiology score) + (0.1550*operative severity score).

REFERENCES

1 Copeland GP, Jones D, Walters M

POSSUM: a scoring system for surgical audit.

Br J Surg. 1991; 78: 355-360

2 Prytherch DR, Whiteley MS, Higgins B, Weaver PC, Prout WG, Powell SJ

POSSUM and Portsmouth POSSUM for predicting mortality. Physiological and Operative Severity Score for the enUmeration of mortality and morbidity

Br J Surg. 1998; 85: 1217-1220.

3 Sagar PM, Hartley MN, Mancey-Jones B, Sedman PC, May J, Macfie J

Comparative audit of colorectal resection with the POSSUM scoring system

Br J Surg. 1994; 81: 1492-1494

4 Tekkis PP, Kessaris N, Kocher HM, Poloniecki JD, Lyttle J, Windosr ACJ.

Evaluation of POSSUM and P-POSSUM scoring systems in patients undergoing gastrointestinal surgery.

Br J Surg. 2003; 90: 640-345

5 Tekkis PP, Kocher HM, Bentley AJ, Cullen PT, South LM, Trotter GA, Ellul JPM.

Operative mortality rates among surgeons: comparison of POSSUM and P-POSSUM scoring systems in gastrointestinal surgery.

Dis Col Rect 2000; 43: 1528-32