The association between ICU level of care and mortality in the Netherlands

Intensive Care Med. 2015 Feb;41(2):304-11. doi: 10.1007/s00134-014-3620-1. Epub 2015 Jan 20.


Purpose: The relationship between the number of patients admitted to an intensive care unit (ICU) volume and mortality is currently the subject of debate. After implementation of a national guideline in 2006, all Dutch ICUs have been classified into three levels based on ICU size, patient volume, ventilation days, and staffing. The goal of this study is to investigate the association between ICU level and mortality of ICU patients in the Netherlands.

Methods: We analyzed data from 132,159 patients admitted to 87 ICUs between January 1, 2009 and October 1, 2011. Logistic GEE analyses were performed to assess the influence of ICU level on in-hospital mortality and 90-day mortality in the total ICU population and in different ICU subgroups while adjusting for severity of illness by APACHE IV.

Results: No significant differences were found in the adjusted in-hospital mortality of the total ICU population and in different subgroups admitted to level 1, 2 and 3 ICUs. In-hospital mortality in level 2 and 3 ICUs as opposed to level 1 ICUs was 1.06 (0.93-1.22) and 1.10 (0.94-1.29), respectively, and 90-day mortality was 0.92 (0.80-1.06) and 1.01 (0.88-1.17).

Conclusion: We demonstrated that ICU level was not associated with significant differences in the case-mix adjusted in-hospital and long-term mortality of ICU patients. This finding is in contrast with some earlier studies suggesting a volume-outcome relationship. Our results may be explained by the successful implementation of nationwide mandatory quality requirements and adequate staffing in all three levels of ICUs over the last years.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Cohort Studies
  • Critical Care / methods*
  • Female
  • Hospital Mortality*
  • Humans
  • Intensive Care Units / statistics & numerical data*
  • Length of Stay / statistics & numerical data*
  • Logistic Models
  • Male
  • Middle Aged
  • Netherlands