Scoring Systems for Comparison of Disease Severity in Intensive Care Unit Patients

Published on 10/03/2015 by admin

Filed under Critical Care Medicine

Last modified 22/04/2025

Print this page

rate 1 star rate 2 star rate 3 star rate 4 star rate 5 star
Your rating: none, Average: 0 (0 votes)

This article have been viewed 1687 times

Chapter 89 Scoring Systems for Comparison of Disease Severity in Intensive Care Unit Patients

Benoit Misset, MD , Islem Ouanes, MD

1 What are severity scores?

Scoring systems have been developed to compare the severity of disease among patients in intensive care units (ICUs). These scores are established at admission or during the ICU stay, and they are widely used in the ICU; other scores were recently developed to be used at ICU discharge to predict worse outcome after discharge (early death or ICU readmission). The scores include the assessment of several physiologic parameters that have been documented to play an independent role in predicting hospital death. The scores presented in this chapter are only those used to assess general disease severity. Other scoring systems are used to assess particular organ function (i.e., Acute Lung Injury Score for evaluation of acute respiratory distress syndrome and Sequential Organ Failure Assessment [SOFA] or to assess resource and workload use (i.e., Therapeutic Intervention Scoring System [TISS]).

Scores at ICU admission

2 Which scores are used for assessing the general severity of disease at ICU admission?

The three most frequently used systems are the:

image Acute Physiology and Chronic Health Evaluation (APACHE)

image Simplified Acute Physiology Score (SAPS)

image Mortality Predictive Model (MPM)

The most recent versions of each system are the APACHE IV (2006), the SAPS III (2005), and the MPM III (2007).

3 Why were scores to assess general disease severity at ICU admission developed?

image To assess performance of the ICU. The ICU patient is a medical or a surgical patient who has either acute failure of one major vital function or a high risk for development of such failure. Because the mortality rate of ICU populations is usually high and varies widely depending on patient admission policies, an objective assessment of the patients’ general disease severity is necessary to ensure that the mortality rate in an ICU is consistent with the overall severity of its patient population at admission. The ratio between observed and predicted mortality, called the standardized mortality ratio, is the simplest way to assess the performance of an ICU. It allows comparisons among mortality rates of various ICUs or the mortality rates documented in one ICU over time.

image To assess the patient’s risk for death. The scores give an objective evaluation that helps the clinician confirm the severity of the patient’s illness. However, these scores cannot be used to make decisions about individual patients (e.g., withdrawal of support).

image To compare or match populations in clinical studies. In randomized, controlled studies the scores have been used to confirm that the populations obtained by randomization had a similar disease severity at admission to the ICU. In case-control studies, the scores have been used to match the control to the case patients.

4 How were scores assessing general severity at ICU admission constructed?

Scores were constructed in large, multicenter, prospective populations. The variables were selected and weighed by consensus of panels of experts (first version of SAPS in 1984, first version of APACHE in 1981, and APACHE II in 1985) or through multiple logistic regression analyses used for most recently developed scores (SAPS II and III, APACHE III and IV, and MPM II and III) to determine whether the parameters were independent predictors of hospital death. The tested variables include age, worst values over the first 24 hours of ICU admission for certain acute physiologic abnormalities (e.g., sodium, potassium, partial arterial oxygen tension, urine output, Glasgow Coma Scale), category at admission (medical or surgical patient), and several underlying diseases (e.g., metastatic cancer, acquired immunodeficiency syndrome). The MPM system also includes several therapeutic items (e.g., number of venous lines).

The SAPS III is based on a more complex methodology. It has the advantage of being based on a worldwide population and of giving a larger place to prior health status and to circumstances of admission in addition to the physiologic imbalance at ICU admission. MPM0, MPM III, and SAPS III are collected entirely at admission to the ICU (i.e., within 1 hour), which reduces potential suboptimal care in the first day of the ICU in the assessment of severity.

The most recent version of APACHE score is the APACHE IV developed in 2006 with use of a database of more than 100,000 patients admitted to 104 ICUs in 45 hospitals in the United States in 2002-2003, and remodeling the APACHE III score.

MPM III is a recent update of MPM0 (2007) with use of a database of 124,885 patients from 135 ICUs. This recent score uses 16 variables including three physiologic parameters obtained within 1 hour of ICU admission to estimate mortality probability at hospital discharge.

5 How were scores assessing general disease severity at ICU admission validated?

All of these models were validated in the initial studies in a subset of patients that was not used for construction of the scoring system. The performance of the scoring systems was considered adequate if they showed good discrimination in predicting hospital mortality and had a good calibration for the entire population under investigation. Discrimination in predicting hospital mortality is assessed with receiver operating characteristics (ROCs): the higher the area under the ROC curve, the more discriminative the test. The calibration in the entire population is measured with the goodness-of-fit test: the observed mortality must not be statistically different from the expected mortality in population deciles of equal probability intervals. The lower the Hosmer-Lemeshow H statistic value and the higher the corresponding p value, the better the calibration.

6 Which scores have been validated adequately?

The APACHE I and II, SAPS I, and MPM I have not been constructed or validated with the current accepted methodologic standards. The SAPS II, MPM II, and APACHE III scores have been shown to have good discrimination and calibration in large multicenter studies.

image The SAPS II is well validated. The score needs to be updated with more recent ICU populations.

image The MPM II is well validated and has the advantage of being the only score available at ICU admission rather than at 24 hours after admission. This advantage is made possible because the score includes some therapeutic items (e.g., venous lines, drainage systems). The MPM II score also needs to be updated with more recent ICU populations.

image The APACHE III is well validated and updated regularly, but its use is limited by the fact that clinicians must pay to know and use its equation for calculating death probability.

image The APACHE IV and the MPM III are well calibrated, and they have a good discrimination; they were validated in a multicentered study of 11,300 ICU patients from California showing that APACHE IV had better discrimination and longer data extraction time than MPM III. MPM III was also validated in 55,459 patients from 103 ICUs, 25 of which did not participate in the original development.

image The SAPS III is well validated and updated as it was published in 2005, that is, 12 years after the most recent among the other ones. Unlike the APACHE III, its construction details were diffused to the entire scientific community. It appears to be a good candidate for an international benchmark, and its use is free of charge.

See Box 89-1.

Box 89-1 Structure of the recent simplified acute physiology score iii

Box I

Age

Comorbidities

Length of stay before ICU admission

Intrahospital location before ICU admission

Use of major therapeutic options before ICU admission

Box II

ICU admission: planned or unplanned

Reason(s) for ICU admission (nine categories)

Surgical status at ICU admission (none, scheduled, emergency)

Anatomic site of surgery (six categories)

Acute infection at ICU admission (nosocomial, respiratory)

Box III

Glasgow Coma Scale score

Total bilirubin

Body temperature

Serum creatinine level

Heart rate

White blood cell count

Serum pH

Platelet count

Systolic blood pressure

Blood oxygenation

See also www.saps3.org.

Scores over the ICU stay

7 Why were scores assessing disease severity over the ICU stay developed?

The scores measuring daily severity were developed to improve the prediction of an individual patient’s death (already provided by scores at admission), to assess the activity and the performance of ICUs, and to match patients in clinical investigations. These scores are particularly useful as inclusion criteria in randomized studies in which patients are entered into the investigation several days after admission. These scores are also used as matching criteria in case-control studies addressing the attributable morbidity of ICU-acquired events, such as nosocomial infections.

8 Which scores have been developed for assessing severity over the ICU stay?

Various scores have been developed on ICU samples of various sizes. The scores developed on the largest ICU populations include the:

image Organ System Failure score (OSF)

image Organ Dysfunction and Infection score (ODIN)

image Logistic Organ Dysfunction Score (LODS)

image Sequential Organ Failure Assessment score (SOFA)

image Multiple Organ Dysfunction Score (MODS)

9 How were the scores assessing severity over the ICU stay constructed?

Sequential scores measure the number and/or the intensity of organ dysfunction. By contrast with the SAPS, MPM, and APACHE scores, they do not take age, category of admission, or underlying diseases into account, because these items do not change over the ICU stay.

The OSF, ODIN, SOFA, and MODS scores were constructed empirically. The OSF and ODIN scores assess the number of organ dysfunctions, and the SOFA and MODS scores also assess the intensity of organ dysfunction. The MODS construction included testing of its validity, reproducibility, and sensitivity of each test item between observers.

The choice and the weight of each LODS item are derived from a multiple logistic regression model, with use of hospital mortality as the dependent variable. The construction was made from the data collected during the first day in the ICU.

10 How were they validated?

Discrimination for predicting mortality is better in multivariate models when the evolution of the score account and its initial values are taken into account. This principle was demonstrated with the LODS score.

11 What did these scores add to the description of ICU patients?

image The use of the OSF score initially showed that a 100% prediction of death could be made in the most severely afflicted patients after several days. However, the same score was eventually used to demonstrate that care in the ICU had improved over the years, so that published results were no longer valid 10 years later. These investigations documented that such scores are a method to assess ICU performance.

image The mean time of occurrence of each organ failure is not the same. The peak of dysfunction for the neurologic system occurs usually before the second day; for the respiratory, cardiovascular, renal, and coagulation systems, around the third day; and for hepatic dysfunction, around the fifth day.

image The weight of each organ failure in predicting death is not the same: hematologic and hepatic failures have less effect on mortality than respiratory, cardiovascular, renal, and neurologic failures.

image The weight of each organ failure in predicting death is not the same over time. The same increase in respiratory dysfunction has a worse prognosis after 1 week of ICU stay. Hepatic dysfunction has an effect on mortality only after 3 weeks of ICU stay.

Scores at ICU discharge

12 At what time of the ICU stay should either of these scores be used? (See Fig. 89-1.)

Among critically ill patients, many experience clinical deterioration or death shortly after discharge from the ICU. These adverse outcomes might be ascribable to premature ICU discharge. Therefore determining the optimal time for ICU discharge is crucial. Moreover, ICU readmission is associated with a fivefold increase in hospital mortality compared with the initial prediction. Recently, scoring systems at ICU discharge are developed to predict prognosis after ICU discharge and to determine the optimal time of discharge to prevent unplanned ICU readmission or death. The Stability and Workload Index for Transfer (SWIFT) is a score for predicting unplanned ICU readmission that was developed in three ICUs. This score, which is determined at ICU discharge, is based on ICU length of stay, location before ICU admission, and neurologic and respiratory impairment on the discharge day. In two validation cohorts, the SWIFT value predicted ICU readmission but exhibited poor calibration.

image

Figure 89-1 When should we measure the severity scores?

Key Points Scoring Systems for Comparison of Disease Severity in Intensive Care Unit Patientsimage

1. Most common items of the severity scores measured at ICU admission

a. Age
b. Previous health status
c. Acute physiologic imbalance at ICU admission
d. Reason for admission
e. Mode of admission

2. Usual goals of measuring a severity score at ICU admission

a. Prediction of death
b. ICU populations comparison in scientific studies
c. Performance and benchmarking, when linked to observed mortality

3. Common utilization of the organ dysfunction scores

a. Match the patients in case-control or exposed-unexposed studies
b. Adjust the cohorts in risk-factor multivariate analyses

Bibliography

1 Cook R., Cook D., Tilley J., et al. Multiple organ dysfunction: baseline and serial component scores. Crit Care Med. 2001;29:2046–2050.

2 Fagon J.-Y., Chastre J., Novara A., et al. Characterization of intensive care unit patients using a model based on the presence or absence of organ dysfunctions and/or infection: the ODIN model. Intensive Care Med. 1993;19:137–144.

3 Gajic O., Malinchoc M., Comfere T.B., et al. The Stability and Workload Index for Transfer score predicts unplanned intensive care unit patient readmission: initial development and validation. Crit Care Med. 2008;36:676–682.

4 Higgins T.L., Kramer A.A., Nathanson B.H., et al. Prospective validation of the intensive care unit admission Mortality Probability Model (MPM0-III). Crit Care Med. 2009;37:1619–1623.

5 Higgins T.L., Teres D., Copes W.S., et al. Assessing contemporary intensive care unit outcome: an updated Mortality Probability Admission Model (MPM0-III). Crit Care Med. 2007;35:827–835.

6 Knaus W.A., Draper E.A., Wagner D.P., et al. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13:818–829.

7 Knaus W.A., Draper E.A., Wagner D.P., et al. Prognosis in acute organ-system failure. Ann Surg. 1985;202:685.

8 Knaus W.A., Wagner D.P., Draper E.A., et al. The APACHE III prognostic system: risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 1991;100:1619–1636.

9 Kuzniewicz M.W., Vasilevskis E.E., Lane R., et al. Variation in ICU risk-adjusted mortality: impact of methods of assessment and potential confounders. Chest. 2008;133:1319–1327.

10 Le Gall J.R., Klar J., Lemeshow S., et al. The Logistic Organ Dysfunction system: a new way to assess organ dysfunction in the intensive care unit. JAMA. 1996;276:802–810.

11 Le Gall J.R., Lemeshow S., Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA. 1993;270:2957–2963.

12 Lemeshow S., Teres D., Klar J., et al. Mortality Probability Models (MPM II) based on an international cohort of intensive care unit patients. JAMA. 1993;270:2478–2486.

13 Marshall J.C., Cook D.J., Christou N.V., et al. Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care Med. 1995;23:1638–1652.

14 Metnitz P.G., Lang T., Valentin A., et al. Evaluation of the logistic organ dysfunction system for the assessment of organ dysfunction and mortality in critically ill patients. Intensive Care Med. 2001;27:992–998.

15 Moreno R.P., Metnitz P.G., Almeida E., et al. SAPS 3—from evaluation of the patient to evaluation of the intensive care unit. Part 2. Development of a prognostic model for hospital mortality at ICU admission. Intensive Care Med. 2005;31:1345–1355.

16 Moreno R., Vincent J.L., Matos A., et al. The use of maximum SOFA score to quantify organ failure/dysfunction in intensive care: result of a prospective multicenter study. Intensive Care Med. 1999;25:686–696.

17 Vincent J.L., de Mendonca A., Cantraine F., et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Crit Care Med. 1998;26:1793–1800.

18 Zimmerman J.E., Kramer A.A., McNair D.S., et al. Acute Physiology and Chronic Health Evaluation (APACHE) IV: hospital mortality assessment for today’s critically ill patients. Crit Care Med. 2006;34:1297–1310.

Related posts:

Intensive Care Unit Organization, Management, and Value Status Epilepticus Bronchoscopy Skin and Soft Tissue Infections Ultrasound in the Intensive Care Unit Hemodynamic Monitoring