Illness severity scoring systems

Junior ICU staff are expected to fill out these cumberome forms, and nobody ever explains to them the hidden meaning behind this seemingly purposeless activity. However, there is method to this madness. Pragmatically, filling out APACHE forms ensures that the unit you work in will remain accredited for training. The ability to calculate SMRs on the basis of APACHE data helps compare your unit to other similar units. And you can use the data in clinical trials, retrospective audits, in mortality prediction, funding allocation and many other things besides.

Traditionally, we suck at answering questions about this boring topic. Question 11 from the first paper of 2009 asked the candidates to compare APACHE with SOFA, and 87% of us failed. Question 4 from the second paper of 2005 was more fair, asking generally about principles behind illness scoring systems and their relationship to outcome - still only 29% were able to score passing marks. Writing in 2017, the younger version of this author theorised that, though these questions have not been seen for many years, a local recurrence would be inevitable, and would likely lead many candidates to their doom; and this speculation was vindicated when scoring systems appeared again as Question 18 from the cursed first paper of 2023.

The objective of this chapter is to arm the candidates with enough basic information to score a passing mark. There are of course various scoring systems which are disease specific (eg. the Glasgow Coma Scale, the MELD score for liver disease) and which have defined uses in predicting prognosis and allocating treatment. The ICU scoring systems are somewhat different. Life In The Fast Lane has an excellent summary of ICU scoring systems, which is directed at the CICM Fellowship Exam candidate.

Here are some links to the seminal articles which describe these systems:


Why do we need ICU scoring systems?

ICU scoring systems collect data about the whole patient, with all their multisystem problems, and render this complex picture down into a single numerical value - their "illness severity score". So, if somebody asks, "how sick was that man?" you may be able to answer: "22".

How are these scoring systems useful?

  • They (try to) predict outcome and length of stay
  • They can be used to compare predicted and observed outcome
  • They stratify patients for clinical trials, according to disease severity
  • They assess ICU performance
  • They allow resources to be allocated to ICUs according to the illness severity of their patients
  • They allow a comparison of ICUs

What is an "ideal" scoring system?

LITFL gives the following as a list of qualities for the "ideal" ICU scoring system. Their

According to them, the ideal scoring system should have the following characteristics :

  • Scores calculated on the basis of easily / routinely recordable variables
  • Well calibrated and validated
  • A high level of discrimination
  • Applicable to all patient populations in ICU
  • Can be used in different countries, health systems or patient cohorts
  • The ability to predict mortality, functional status or quality of life after ICU discharge


  • considers co-morbidities
  • considers organisational aspects
  • provides a common language for discussion
  • method to evaluate critical care practice and process
  • allows ability to compare groups in clinical trials

In his 2010 review of scoring systems, Jean-Louis Vincent gives this list of "ideal" features:

  • Simple and inexpensive
  • Routinely available in all ICUs
  • Reliable (intra and inter-observer)
  • Objective (that is, observer independent)
  • Specific to the function of the organ in question
  • Therapy independent
  • Sequential (available at ICU admission or shortly thereafter and then at fixed periods of time)
  • Not affected by transient, reversible abnormalities associated with therapeutic or practical interventions
  • Reflect acute dysfunction of the organ in question but not chronic dysfunction
  • Reproducible in large, heterogeneous groups of ICU patients
  • Reproducible in several types of ICUs from different regions of the globe
  • Abnormal in one direction only
  • Using continuous rather than dichotomous variables

What does an ICU scoring system measure?

Most of them measure physiological variables.

eg. how low the hemoglobin, how high the blood presure etc.

Some of them measure interventions, eg. how much noradrenaline did the patient require, and whether they ended up mechanically ventilated or on dialysis

Most of the abovementioned variables are given a score, weighing their "severity". The weighting is derived from logistic regression analysis of large demographic data sets, which reveal the association of these physiological variables with mortality.

The APACHE score

APACHE stands for Acute Physiology, Age and Chronic Health Evaluation (I-IV).

  • APACHE II is the most commonly used one
  • 12 variables are measured
  • Scores range from 0 to 71
  • Derived from histrical data set

The APACHE-III was intoduced in 1991, and APACHE-IV in 2006. Each is an improvement (in terms of prognostic accuracy) but for some reason these are not widely used in Australia.

The APACHE system assesses three factors which influence ICU survival: chronic disease, patient reserve, and severity of acute illness. It takes into account such afctors as whether the admission was elective or not, and whether the admission was post-operative or nonsurgical.

The physiological component of the system is based on the most abnormal scores within the first 24 hours. Though it is not possible to predict individual patient outcomes, standardised mortality ratios can be used on large patient populations.

The SOFA score

SOFA stands for Sequential Organ Failure Assessment .

  • 6 organ systems are scored according to their function
  • The degree of organ support is taken into account
  • Used to analyse secondary endpoints in clinical trials

There is a defined score of 1-4 for each organ system, which is collected daily. This not a predictive model- there are no mortality algorithms here. A higher SOFA score can be said to relate to increased mortality, but there is no mathematical model to help us figure out exactly how the total score relates to survival.

A comparison of SOFA and APACHE scoring systems

LITFL has a good table which describes the differences between these two scoring systems.

It is derived from the CICM Fellowship exam question (Question 11 from the second paper of 2009). With minor modifications, it is reproduced below:


A Comparison of the SOFA and APACHE Scoring Systems



Basic premise

ICU mortality depends on three domains:

  • Premorbid health
  • Severity of illness
  • Patient's physiological reserve

Thus, if one can quantify these domains, one may be able to predict mortality on the basis of such measurements.

Degree of organ dysfunction is related to acute illness. Originally designed with sepsis in mind, but subsequently validated in other disease states.

Measured parameters

Heuristic groupings of 12 physiologic variables, Glasgow Coma Score (GCS), age, and chronic health evaluation status.

6 domains of organ system function

Measurement collection

Worst score within the first 24 hours

Daily measurement of

Unique features

Incorporates chronic illness, emergency admission, age, surgical vs non-surgical admission, and cardiorespiratory arrest

Incorporates the use of organ system support sug as vasopressors and dialysis


0 to 71

0 to 24

Mortality prediction

The risk of hospital death is computed by combining APACHE II score with Knaus' 
weighted coefficient for different types of disease entities. A score of 25 represents a predicted mortality of 50% and a score of over 35 represents a predicted mortality of 80%.

SOFA does not predict mortality, and the original authors intended it to be used as a means of reproduceably describing a sequence of complications in the critically ill.

That said, higher SOFA scores are in fact associated with increased mortality.

Prognostic value

APACHE is a poor predictor of individual patient outcome.

One can monitor response to therapy by the change of daily SOFA scores

Some studies have tested these scales head-to-head. An Indonesian study found that SOFA scores predict mortality better than APACHE-II in surgical ICU patients. The Canadians found that APACHE had good ROC chacateristics and predicted mortality resembled observed mortality, but in a population with a mean APACHE score of 16. An audit of a large Scottish database however has revealed that all scoring systems generally fail, and generate mortality predictions which are completely different to the observed mortality.

The less popular illness severity scoring systems

TISS: Therapeutic Interventions Scoring System

  • 76 variables (interventions and treatments)
  • Collected daily
  • Indicates nursing and medical workload
  • Does not indicate severity of illness
  • Most useful for accountants

SAPS: Simplified Acute Physiology Score

  • SAPS 1 only looked at physiology, and was used by French ICUs
  • SAPS 2 added chronic health conditions, and was used in Europe and North America
  • SAPS 3 had 20 variables and was used worldwide

MPM: Mortality Prediction Models

  • MPM measures variables at admission and in the first 24 hours
  • It calcuates the risk of in-hospital death on the basis of these variables, using a logistic regression model.
  • MPM II was based on the same historical data set as SAPS 2 and predicts mortality at 24, 48 and 7 hours.

POSSUM: = Physiological and Operative Severity Score for the enumeration of Mortality and Morbidity

  • 12 physiological parameters for surgeons
  • Used by surgeons as a risk adjustment tool
  • Different subspecialties have their own: V-POSSUM is for vacular surgeons, Cr-POSSUM is for colorectal, etc

Problems of using severity scoring systems

  • There is a variation in recording of data - not everyone is equally accurate at filling out the forms
  • There are differences in patient groups which influence "illness severity" which are not measured by the scoring system
  • Some data goes missing
  • Some data may be doctored in pursuit of funding
  • Outcomes may not be related to ICU alone - the whole hospital is involved


Young, Paul, et al. "End points for phase II trials in intensive care: Recommendations from the Australian and New Zealand clinical trials group consensus panel meeting." Critical Care and Resuscitation 15.3 (2013): 211. - this one is not available for free, but the 2012 version still is:

Young, Paul, et al. "End points for phase II trials in intensive care: recommendations from the Australian and New Zealand Clinical Trials Group consensus panel meeting." Critical Care and Resuscitation 14.3 (2012): 211.

Suter, P., et al. "Predicting outcome in ICU patients." Intensive Care Medicine20.5 (1994): 390-397.

Martinez, Elizabeth A., et al. "Identifying Meaningful Outcome Measures for the Intensive Care Unit." American Journal of Medical Quality (2013): 1062860613491823.

Tipping, Claire J., et al. "A systematic review of measurements of physical function in critically ill adults." Critical Care and Resuscitation 14.4 (2012): 302.

Gunning, Kevin, and Kathy Rowan. "Outcome data and scoring systems." Bmj319.7204 (1999): 241-244.

Woodman, Richard, et al. Measuring and reporting mortality in hospital patientsAustralian Institute of Health and Welfare, 2009.

Vincent, J-L. "Is Mortality the Only Outcome Measure in ICU Patients?."Anaesthesia, Pain, Intensive Care and Emergency Medicine—APICE. Springer Milan, 1999. 113-117.

Rosenberg, Andrew L., et al. "Accepting critically ill transfer patients: adverse effect on a referral center's outcome and benchmark measures." Annals of internal medicine 138.11 (2003): 882-890.

Burack, Joshua H., et al. "Public reporting of surgical mortality: a survey of New York State cardiothoracic surgeons." The Annals of thoracic surgery 68.4 (1999): 1195-1200.

Hayes, J. A., et al. "Outcome measures for adult critical care: a systematic review." Health technology assessment (Winchester, England) 4.24 (1999): 1-111.

RUBENFELD, GORDON D., et al. "Outcomes research in critical care: results of the American Thoracic Society critical care assembly workshop on outcomes research." American journal of respiratory and critical care medicine 160.1 (1999): 358-367.

Turnbull, Alison E., et al. "Outcome Measurement in ICU Survivorship Research From 1970 to 2013: A Scoping Review of 425 Publications." Critical care medicine (2016).