Question 18

Regarding prediction scoring systems in the ICU.

Discuss one commonly used example under the following headings: Components, advantages, disadvantages, and its use in the ICU.
(100% marks)

[Click here to toggle visibility of the answers]

College answer

Aim. To explore the candidate knowledge of APACHE, SAPS, MOPS and use in ICU practice.

Key sources include: Paper 2009.1 Q11 - comparing APACHE with SOFA. 2005.2 SQ 4 - principles of Illness severity scoring systems in the critically ill pt. Systems in use daily in ICUs outcomes comparisons. CanMEDS scholar.

Discussion: Scoring systems used to predict mortality and other outcomes are universal in Australian and NZ ICUs. Trainees/SIMGs should be encouraged to have a working knowledge of those commonly used in the Binational ICU registry and assessment of risk severity in research. A demonstrated knowledge of any of the ICU prediction scoring systems gave an expert pass.

Given the wording of the question, other prediction scoring systems used in the ICU such as Child-Pugh scores would have answered the stated question. Examiners gave credit as the answers corresponded to the question asked.

The GCS is not in itself a predictive scoring system and was marked incorrect as it did not address the question. Other ways to improve the answer detail required included exploration of scoring systems derivation as part of the explanation of the components.


Whereas the  Question 11 from the first paper of 2009 asked the candidates to compare APACHE with SOFA, this question asked for any predictive scoring system, which theoretically could mean that the candidates could have used a range of familiar systems. This is more similar to  Question 4 from the second paper of 2005 which wanted to know about the general principles of illness severity scoring systems (where, interestingly, the GCS was an acceptable response, and in fact formed a part of the college answer). The bottom line is, there are multiple options (there's several versions of the APACHE system, for example), and so it would be better to pick one system here, and dissect it using the provided headings. Arbitrarily, the ancient 1991 APACHE III-j* system is selected here, for no reason other than the fact that (at the time of writing, on a freezing July day in 2023) it is used by ANZICS CORE**, which makes it easier to answer the last part of the question - because what would a CICM trainee be able to say about the TISSSAPS IIMPM II or POSSUM systems under that heading?


Components of APACHE III-j

  • Major medical and surgical disease categories
  • Acute physiologic abnormalities in the first 24 hours (17 variables)
  • Age
  • Preexisting functional limitations
  • Major comorbidities
  • Treatment location immediately prior to ICU admission

Advantages of APACHE III-j

  • Of APACHE-III-j specifically:
    • Easy to collect data (computerised information systems can calculate the score automatically without much human input)
    • Well-validated and internationally familiar
    • Makes comparison between health services easier
    • Is an improvement on the APACHE-II score, as some of the physiological variables have been re-weighed
  • Of such scoring systems in general:
    • Can be used to standardise quality assurance studies and research
    • Can help perform comparisons between health services
    • Usually user-friendly and dependent on variables which are already being collected for patient care purposes

Disadvantages of APACHE III-j

  • Of APACHE III-j specifically:
    • Does not incorporate frailty (but APACHE-IV does!)
    • Older, and drifting in calibration in terms of predictive value (last validated in 2003!)
    • Not a sequential score (thus, cannot be used to track response to therapy)
  • Of these systems in general:
    • Generally these are poor predictors of individual patient outcome
    • They are susceptible to coding errors, particularly where one variable is subjective (eg. the "diagnosis" category in the APACHE score)
    • There is a variation in recording of data - not everyone is equally accurate at filling out the forms, and computerised systems can record spurious readings uncritically
    • There are differences in patient groups which influence "illness severity" which are not measured by the scoring system
    • Some data goes missing
    • Delay to ICU admission affects the initial score 
    • Outcomes may not be related to ICU alone - the whole hospital is involved

Use of APACHE III-j in the ICU

  • Used in ICUs around the world to predict mortality for critically ill patients
  • Permits benchmarking worldwide, including the analysis of endpoints in trials
  • Allows the standardisation of illness severity scoring, permitting comparison between patient populations in trials
  • Allows administrators to assess illness acuity in health services and to use this information to determine the allocation of resources and staff

A lot of these advantages and disadvantages were extracted from the excellent 2008 paper by Shann et al.


* The "j" in APACHE III-j is the 10th iteration of the APACHE-III model. 

** The ANZICS CORE now uses ANZROD (since 2014) as the predictive model for mortality, but they still collect APACHE-III data.

Knaus, William A., et al. "The APACHE III prognostic system: risk prediction of hospital mortality for critically III hospitalized adults." Chest 100.6 (1991): 1619-1636.

Shann, F., et al. "Critical care outcome prediction equation (COPE) for adult intensive care." Critical Care and Resuscitation 10.1 (2008): 35-41.