Elsevier

Resuscitation

Volume 80, Issue 5, May 2009, Pages 505-510
Resuscitation

Review
Severity of illness and risk of readmission to intensive care: A meta-analysis*

https://doi.org/10.1016/j.resuscitation.2009.02.015Get rights and content

Abstract

Background

Almost one in every 10 patients who survive intensive care will be readmitted to the intensive care unit (ICU) during the same hospitalisation. The association between increasing severity of illness (widely calculated in ICU patients) with risk of readmission to ICU has not been systematically summarized.

Objective

The meta-analysis was designed to combine information from published studies to assess the relationship between severity of illness in ICU patients and the risk of readmission to ICU during the same hospitalisation.

Data sources

Studies were identified by searching MEDLINE (1966 to August 2008), EMBASE (1980–2008), and CINAHL (1982 to August 2008).

Review methods

Studies included only adult populations, readmissions to ICU during the same hospitalisation and reports of valid severity of illness index.

Results

Eleven studies (totaling 220 000 patients) were included in the meta-analysis. Severity of illness (APACHE II, APACHE III, SAPS and SAPS II) measured at the time of ICU admission or discharge, was higher in patients readmitted to the ICU during the same hospitalisation compared to patients not-readmitted (both p-values < 0.001). The risk of readmission to ICU increased by 43% with each standard deviation increase in severity of illness score (regardless if measured on admission to, or discharge from the ICU) (odds ratio (OR) = 1.43, 95% confidence interval (CI) = 1.3–1.6).

Conclusions

A relationship between increasing intensive care severity of illness and risk of readmission to ICU was found. The effect was the same regardless of the time of measurement of severity of illness (at admission to ICU or the time of discharge from ICU). However, further research is required to develop more comprehensive tools to identify patients at risk of readmission to ICU to allow the targeted interventions, such as ICU-outreach to follow-up these patients to minimize adverse events.

Introduction

Despite more patients surviving intensive care, approximately 10% require readmission to the intensive care unit (ICU).1 It is not clear whether the decision to discharge patients from the ICU or the level of care given to these patients on the general wards or a combination results in readmission to ICU. This event, associated with adverse health events has been highlighted as a potential marker of quality of care.2 Therefore, the ability to identify patients at high risk of readmission to ICU during the same hospitalisation could allow objective decisions to be made by clinicians related to the timing of discharge from ICU, the level of care required by patients on the ward and the need for follow-up by ICU staff.

In an attempt to address this problem several authors have either collected specific data to identify risk factors for readmission to ICU, or have analyzed routinely collected data to identify predictors of readmission.3, 4, 5, 6, 7, 8, 9 Some papers have specifically addressed the ability of predictors to discriminate between patients who are readmitted to ICU and those who are not.4, 10 Importantly, a index of severity of illness (APACHE II)11 routinely collected on admission to the ICU is by itself as accurate as a more complex predictive model that also uses characteristics of the ICU stay, such as length of stay and mechanical ventilation days.4 While severity of illness is routinely calculated in patients admitted to ICU to predict in-hospital mortality in many settings throughout the world,12 the association between severity of illness and the risk of readmission to ICU has not been systematically summarized.

For this reason, a meta-analysis was designed to combine information from published studies to assess the relationship between level of severity of illness in ICU patients and the risk of readmission to ICU during the same hospitalisation.

Section snippets

Methods

This article was prepared in accordance with the Meta-analysis of Observational Studies in Epidemiology (MOOSE) statement.13 Briefly the MOOSE statement outlines a prospective and systematic method for reporting meta-analyses of observational studies including considering the background, search strategy, methods, results, discussion, and conclusion. The rationale for choosing this method is that observational studies potentially reflect the ‘usual care’ of contemporary clinical practice, in

Results

We identified a total of 195 potential studies to be included in the meta-analysis. After excluding review papers, including only adult populations, readmissions to ICU during the same hospitalisation, and studies that reported a valid severity of illness score: 11 studies were retained for analysis. The characteristics of these studies are presented in Table 1. Year of publication ranged from 1993 to 2008. The paper by Chen at al.5 separated results based on teaching or community hospitals.

Discussion

This meta-analysis has shown a relationship between increasing intensive care severity of illness score and risk of readmission to ICU in patients who survive ICU and are discharged to the ward. The effect was consistent, regardless of the timing of measurement of severity of illness (at admission to the ICU or at the time of discharge from ICU). For each standard deviation increase in severity of illness score the risk of readmission to ICU during the same hospitalisation increased by 43%.

Conflict of interest

No potential conflict of interest declared by any of the authors.

References (30)

  • D.A. Chung et al.

    A case–control analysis of readmissions to the cardiac surgical intensive care unit

    Eur J Cardiothorac Surg

    (2002)
  • P.G. Metnitz et al.

    Critically ill patients readmitted to intensive care units—lessons to learn?

    Intensive Care Med

    (2003)
  • A.L. Rosenberg et al.

    Who bounces back? Physiologic and other predictors of intensive care unit readmission

    Crit Care Med

    (2001)
  • O. Gajic 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)
  • J.R. Le Gall

    The use of severity scores in the intensive care unit

    Intensive Care Med

    (2005)
  • Cited by (81)

    • Risk prediction models for intensive care unit readmission: A systematic review of methodology and applicability

      2020, Australian Critical Care
      Citation Excerpt :

      Hence, there has been a high interest in the ICU readmission rate as a quality indicator of critical care.1,8 Nevertheless, current studies have shown that ICU readmission rates are influenced by factors other than quality of care, such as patient characteristics and length of stay.1,9,10 Furthermore, inconsistent results have been reported about the negative impact of ICU readmissions on patient outcomes.

    • Prospective validation and refinement of the APPROACH cardiovascular surgical intensive care unit readmission score

      2019, Journal of Critical Care
      Citation Excerpt :

      The inclusion of respiratory variables, including percent 02 requirements at discharge, tracheostomy, or re-intubation during initial CVICU stay had little impact on model performance. In contrast, the Stability and Workload Index for Transfer (SWIFT) score which is used to assess readmission risk after discharge from general intensive care unit found Pa02/Fi02 ratio was utilized in the final model [20,21]. Similarly, a meta-analysis found that SAPS 2 score, which incorporates the lowest Pa02/Fi02 ratio in a 24-hour period before discharge, was a predictor of general intensive care readmission [20].

    View all citing articles on Scopus
    *

    A Spanish translated version of the summary of this article appears as Appendix in the final online version at doi:10.1016/j.resuscitation.2009.02.015.

    View full text