Correction to “Hospital-Level Variation in the Use of Intensive Care”


This article corrects:

  1. Hospital-Level Variation in the Use of Intensive Care Volume 47, Issue 5, 2060–2080, Article first published online: 30 March 2012

Originally published in Health Services Research, Volume 47, Issue 5 (October 2012)

In Table 2, the reported 95% confidence interval for the hospital-level variance for empty model 1 is incorrect. The correct 95% CI should be (0.28, 0.51). The table is reprinted below:

Table 2. Correlation Coefficients for Hospital-Level Random Effects in Logistic Mixed Regression Models of Intensive Care Use Across 90 Hospitals
 Empty ModelModel with Patient FactorsModel with Patient and Hospital Factors
Model 1Model 2Model 3
  1. *The residual (patient) variance in a mixed logistic regression model is fixed at π(2/3).

  2. **ICC calculated as proportion of hospital variance divided by total variance where total variance is the sum of hospital variance and residual variance π(2/3).

Random effects variance*
Hospital-level variance (95%CI)0.38 (0.28, 0.51)0.81 (0.60, 1.1)0.57 (0.42, 0.76)
Intraclass correlations**, %
Hospital (95% CI)10.3 (7.6, 13)19.7 (15.1, 24.4)14.6 (11, 18.3)