Risk score in surgery

Risk score in surgerySeptic Severity Score (by Elebute & Stoner)
Mannheim Peritonitis Index (MPI);
Peritonitis Index Altona II (PIA II);

All these scores are reflect the early developments and I think it fair to say that they are no longer applicable. The MPI and SSS have attracted quite some interest in Europe and South America. Patients can be scored easily and there will be some prognostic information from the scores. However they are insensitive, score things that we now know are less relevant, and the results must be calibrated in each institution, i.e. they are not universal. They are also subjective and time-sensitive which means that patients have to be prospectively scored and one cannot easily score patients retrospectively because of the subjective classification of some of the variables. They can provide a stratification among patients but there is no prognostic equation attached. So they are a deficient stratification tool and not an outcome model.

This in contrast to APACHE II which is universal, time-insensitive, objective, and has an equation that turns scores into an outcome model. APACHE has all this but it is not a perfect tool. There was once some hopes to find a score that prospectively dicotomize the patients into those that will eventually survive and those that will die. However the world of peritonitis is not like that. The best we can hope to do is estimates of an increasing risk of eventual death or survival. So the outcome model will only provide you with two estimates; i.e. a patient with 30% risk of mortality will at the same time also have a 70% chance of surviving. What the surgeon does with these two estimates is up to him and has nothing at all to do with the score.

The APACHE II has several major imperfections which I will shortly give you. For peritonitis there is no diagnostic weight for the outcome model equation. Attempts have been made to calculate such a factor but not very successfully because other factors which are not prefectly accounted for came into play. The first of these are age points. If you are of old age and suffer severe peritonitis the patient is very likely to die. The original age points are an underestimate of the risk. They should probably be doubbled to account for the increased risk in old people.

Second, many patients with peritonitis have several days before they are treated in the intensive care unit, for instance all with postoperative peritonitis. The time course changes the patient risk substantially and that has to be acoounted for. In the general APACHE system there was a problem about how to describe how patients were selected for ICU treatment. In the APACHE III system there are variables such as if patients are admitted from ER, OR, floor, or other hospital, as well as the number of days elapsed since the patient was admitted to hospital.

Third, the patient response to treatment in the first days after the operation for peritonitis is very important and not at all accounted for in the APPACHE II but is in APACHE III. If patients are scored every day with a new APACHE score it will describe how the patient changes but unfortunately one must use a new equation for each day to calculate how the risk is changeing. If you wish you can use dayly scores for stratification but you cannot calculate the outcome because those equations are not available unless you buy the APACHE III system. The reason for this is partly commercial (APACHE corporation) and partly because the risk estimates changes also with how medicine develops. So the equations have to be changed continuously. When you use APACHE II you compare yourself with treatment results in the United States. Many today are happy that they perform better than the APPACHE II tells but the reason is purely that they are 15 years later.

One way of dealing with changeing risk is to recognize that the worst thing that happens to the patient with peritonitis is the development of multi-organ failure. The MOF score according to Goris, not perfect, but has attracted some considerable interest to describe the postoperative course of patients with peritonitis.

Perhaps the best you can do is this: Use APACHE II, which I know you have, and score dayly, or for instance day1, 3, 5, 7. 10 and 14. Score same days with the Goris score. This will give you a good stratification of patients. If you would like to develop an outcome model the best is probably to accumulate a trainee set of patients, between 50-100 patients, and use these to develop equations of your own. There will have to be at least two, one for Day 1 and one for Day 5 which also have the Goris score. If you think that is too much I could propose to use the original APACHE II equation, use the weight 1 for peritonitis, double the age points, but for Day 5 you would anyway need to calculate an addition with the Goris score. Check that the outcome estimates fit in new patients.

If you would like a better score of MOF than the Goris I suggest you look up a paper by John Marshall in Crit Care Med.

One Response
  1. Retty

Leave a Reply

Your email address will not be published. Required fields are marked *