Intensivists Versus Hospitalists in the ICU: A Prospective Cohort Study Comparing Mortality and Length of Stay Between Two Staffing Models

Background:

Observational studies conducted prior to the hospitalist movement demonstrate improved patient outcomes in intensive care units (ICUs) staffed by intensivists compared with those staffed by nonintensivists. A large and growing nationwide intensivist shortage, however, precludes ICU staffing exclusively by intensivists. Although many hospitalists care for ICU patients in the United States, no formal investigations have compared outcomes of ICU patients under the care of hospitalists to those of intensivists. Our study's primary aim was to compare inpatient and ICU mortality and length of stay (LOS) between intensivist and hospitalist ICU staffing models.

Methods:

From October 2007 to September 2008, we prospectively enrolled 1445 consecutive medical ICU patients admitted to either an intensivist teaching team or a hospitalist ICU team at an urban, academic, community hospital. Sample size was determined a priori using an expected inpatient mortality of 10%, a power of 80%, and a 2‐sided alpha of 0.05 to demonstrate equivalence in primary outcomes, defined as a mortality difference of less than 5% between the 2 teams. Analysis was blinded. Outcome measures were adjusted using logistic or linear regression, incorporating relevant patient demographics, preexisting comorbidifies, mechanical ventilation status, and disease severity scores using APACHE II.

Results:

There were 1084 ICU patients that met defined inclusion criteria. Patient demographics and preexisting comorbid conditions were similar among the 2 teams, but patients on the hospitalist ICU team had lower average disease severity scores and fewer instances of required mechanical ventilation. Adjusted inpatient mortality of the intensivist team versus the hospitalist ICU team was equivalent [aOR 1.08 (95% Cl: 0.65, 1.93), P = 0.76], as was the adjusted ICU mortality [aOR 1.12 (95% Cl: 0.66, 1.78), P = 0.69]. Adjusted ICU LOS was also equivalent between the 2 teams (P = 0.96), but there was a trend toward a 1.0‐day adjusted hospital LOS reduction (P = 0.07) on the hospitalist ICU team.

Conclusions:

We present the first prospective comparison of patient outcomes between intensivist and hospitalist models of ICU care. Although study limitations include single‐center data, comparison of a teaching intensivist team to a nonteaching hospitalist ICU team, and significant differences in illness severity between the 2 patient groups, the adjusted measures of effect demonstrate equivalence in mortality and LOS between the 2 ICU models. These results suggest hospitalists can help meet current and future critical care needs while maintaining care quality. A randomized trial could help determine if certain ICU patient subgroups are better served under the care of intensivists versus hospitalists.

Author Disclosure:

K. Wise, none; V. Akopov, none; B. Williams Jr., none; D. Dressier, none; K. Leeper, none.