Physician-attributable Differences in ICU Costs

Abstract & Commentary

By James E. McFeely, MD, Medical Director, Critical Care Units, Alta Bates Summit Medical Center, Berkeley, CA, is Associate Editor for Critical Care Alert.

Dr. McFeely reports no financial relationship to this field of study.

Synopsis: In this single-center study comparing the average daily discretionary costs generated during 1184 ICU admissions, mortality and length of stay were not influenced by which of the 9 staff intensivists was managing the patient, but the identity of the intensivist in charge of the patient had a greater influence on costs than all other variables examined except for severity and type of acute illness.

Source: Garland A, et al. Am J Respir Crit Care Med. 2006;1;174(11):1206-1210.

This study was carried out in a 13-bed closed medical ICU, in a county teaching hospital in which 9 board-certified intensivists rotate at 14-day intervals in covering the ICU. Only patients cared for by a single intensivist were included in this analysis. Multivariable linear logistic models were constructed for daily discretionary ICU costs, length of stay and hospital mortality. Discretionary costs were defined as pharmacy, radiology, laboratory, blood bank and echocardiography. The model included adjustments for patient demographics, co-morbidities, type and severity of acute illness, source of ICU admission, ICU workload, and limitations placed on life support prior to admissions. Severity of illness was measured using Glasgow Coma Scale and APACHE II. Patient characteristics were typical for an urban ICU.

In the modeling of daily discretionary costs, the identity of the intensivists was second only to the severity or type of acute illness in influencing the variance in these costs. All other predictors, including co-morbidities, demographics, source of admission and need for mechanical ventilation, had less effect on the variability of cost than the identity of the managing physician. Discretionary ICU costs per day varied by up to 43% depending on which intensivist was managing the case. This variability included all categories of discretionary costs. Comparison of the highest and lowest terciles showed that the average difference in total discretionary costs per admission (p < 0.0001) was approximately $1,000 per case. The intensivists were shown this data and had accurate perceptions of their ranking on a spectrum of expenditure of discretionary costs.


That there are large differences between ICU physicians in the amount of resources used to manage critically ill patients is not particularly surprising. What is surprising, however, is the magnitude of the variability between physicians in the same institution: The patients included in this study had an average ICU stay of 38 hours, yet the discretionary costs varied up to 43%, with an average difference of $1,000 per case. This variability was found in all categories of spending-pharmacy, radiology, laboratory, blood bank, and echocardiography—and was statistically significant even when the analysis was limited either to the first ICU day or to patients with respiratory failure alone. In fact, although the data analysis appears sound and was adjusted appropriately for variables such as need for mechanical ventilation and co-morbidities, the study may even underestimate the magnitude of the variability, given that house staff and fellows participated in the care of patients and may have worked with more than one attending physician.

It is also interesting that, when presented with this data, physicians were accurately able to identify where they stood on the spectrum of discretionary spending, suggesting that those who spend more know they do so. Presumably the higher-spending physicians do so because they think it improves patient outcomes. However, the study also showed that the vast difference in resource use resulted in absolutely no difference in length of ICU stay or mortality.

There is already data in the medical literature regarding the optimal utilization of discretionary laboratory testing and radiography in ICU patients.1-3 This data has not found its way into clinical practice, and is not easily incorporated into guidelines, order sets or protocols. Physician practice styles will have to change, and this article should help to drive that change by showing that care can be delivered with lower resource utilization with no adverse effect on patient outcomes. To successfully reduce inter-physician variability and cost, each bedside clinician needs to re-evaluate what may be a very comfortable, longstanding practice pattern and adjust on the basis of the best practices identified in the literature.


  1. Wang TJ, et al. A utilization management intervention to reduce unnecessary testing in the coronary care unit. Arch Intern Med. 2002:Sep 9;162(16):1885-1890.
  2. Graat ME, et al. The clinical value of daily routine chest radiographs in a mixed medical-surgical intensive care unit is low. Crit Care. 2006 Feb;10(1):R11.
  3. Barie PS, Hydo, LJ. Lessons learned: durability and progress of a program for ancillary cost reduction in surgical critical care. JTrauma. 1997:Oct;43(4):590-594; discussion 594-696.