0
Abstract: Poster Presentations |

ACUTE PHYSIOLOGY AND CHRONIC HEALTH EVALUATION (APACHE)IV ICU LENGTH OF STAY BENCHMARKS FOR TODAY’S CRITICALLY ILL PATIENTS FREE TO VIEW

Jack E. Zimmerman, MD*; Andrew A. Kramer, PhD; Douglas S. McNair, MD; Fern M. Malila, MS
Author and Funding Information

George Washington University, Washington, DC


Chest


Chest. 2005;128(4_MeetingAbstracts):297S. doi:10.1378/chest.128.4_MeetingAbstracts.297S
Text Size: A A A
Published online

Abstract

PURPOSE:  To describe the development and validation of APACHE IV benchmarks for ICU length of stay.

METHODS:  The equations used to generate ICU length of stay benchmarks were developed using ICU day 1 data for 116,209 admissions to 104 ICUs at 45 hospitals during 2002 and 2003. Of these admissions 69,692 were used for model development and 46,517 were used to validate the model. A linear regression procedure was used to estimate exact ICU stay in days and fractions of days. Predictor variables were similar to those used for APACHE III estimates, but new variables (mechanical ventilation, thrombolysis, the impact of sedation on Glasgow Coma Score (GCS), and a rescaled GCS and PaO2:FIO2) were added, and different statistical modeling (restricted cubic splines) was used. We assessed the accuracy of APACHE IV ICU stay predictions by examining the degree of correspondence between mean observed and mean predicted ICU stay (paired Student’s t-test), and by calculating a correlation coefficient (R2).

RESULTS:  Based on relative explanatory power, the most important predictor variables were the acute physiology score (50%), ICU admission diagnosis (14%), ventilator status (11%) and the inability to assess a GCS due to sedation (11%). As the acute physiology score rose there was a linear increase in ICU stay until the score exceeded 80, at which point ICU stay decreased. For the validation data set the aggregate mean observed ICU stay was 3.86 days and mean predicted was 3.78 days (p<0.001). Among 116 ICU admission diagnoses there were only two significant differences (p<0.01) between mean observed and mean predicted ICU stay. The model’s R2 was 0.215 indicating that the model accounted for 21.5% of the variation in ICU stay.

CONCLUSION:  APACHE IV predictions of ICU stay are well calibrated and should provide useful benchmarks for evaluating efficiency in U.S. ICUs.

CLINICAL IMPLICATIONS:  Clinicians can use these benchmarks to assess their unit’s throughput efficiency and monitor the impact of protocols aimed at reducing ICU stay for specific patient groups.

DISCLOSURE:  Jack Zimmerman, Consultant fee, speaker bureau, advisory committee, etc. Medical and Research Consultant.

Wednesday, November 2, 2005

12:30 PM - 2:00 PM


Figures

Tables

References

NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543