Postoperative Complications and Outcomes
Associated With a Transition to 24/7 Intensivist
Management of Cardiac Surgery Patients
Marc A. Benoit, MD1; Sean M. Bagshaw, MD, MSc2; Colleen M. Norris, PhD1,3,4;
Mohamad Zibdawi, MD2; Wu Dat Chin, MBBCh2; David B. Ross, MD, PhD3;
Sean van Diepen, MD, MSc1,2
Objectives: Nighttime intensivist staffing does not improve patient
outcomes in general ICUs. Few studies have examined the asso-
ciation between dedicated in-house 24/7 intensivist coverage
on outcomes in specialized cardiac surgical ICUs. We sought to
evaluate the association between 24/7 in-house intensivist-only
management of cardiac surgical patients on postoperative com-
plications and health resource utilization.
Design: Before-and-after propensity matched cohort study.
Setting: Tertiary care cardiac surgical ICU.
Patients: Patients greater than 18 years old who underwent car-
diac surgery between January 1, 2006, and April 30, 2013 (night-
time resident model), were propensity-matched (1:1) to patients
from August 1, 2013, to December 31, 2014 (24/7 in-house
intensivist model).
Interventions: Cardiac surgical ICU coverage change from a
nighttime resident physician coverage model to a 24/7 in-house
intensivist staffing model.
Measurements and Main Results: The primary outcome of inter-
est was a composite of postoperative major complications.
1
Division of Cardiology, Department of Medicine, Faculty of Medicine and
Dentistry, University of Alberta, Edmonton, AB, Canada.
2
Department of Critical Care Medicine, Faculty of Medicine and Dentistry,
University of Alberta, Edmonton, AB, Canada.
3
Division of Cardiac Surgery, Department of Surgery, Faculty of Medicine
and Dentistry, University of Alberta, Edmonton, AB, Canada.
4
Faculty of Nursing, University of Alberta, Edmonton, AB, Canada.
This work was performed at the University of Alberta.
Supplemental digital content is available for this article. Direct URL cita-
tions appear in the printed text and are provided in the HTML and PDF
versions of this article on the journal’s website (http://journals.lww.com/
ccmjournal).
Dr. Baghsaw is supported by a Canada Research Chair in Critical Care
Nephrology, and he has received funding from Baxter Healthcare. The
remaining authors have disclosed that they do not have any potential con-
flicts of interest.
For information regarding this article, E-mail: sv9@ualberta.ca
Copyright © 2017 by the Society of Critical Care Medicine and Wolters
Kluwer Health, Inc. All Rights Reserved.
DOI: 10.1097/CCM.0000000000002434
Critical Care Medicine www.ccmjournal.org 1
Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Secondary outcomes included duration of mechanical ventila-
tion, all-cause cardiac surgical ICU readmissions, and surgical
postponements attributed to lack of cardiac surgical ICU bed
availability. A total of 1,509 patients during the nighttime resi-
dent model were matched to 1,509 patients during the intensiv-
ist model. The adjusted risk of major complications (26.3% vs
19.3%; odds ratio, 0.73; 95% CI, 0.36–0.85; p < 0.01), mean
mechanical ventilation time (25.2 vs 19.4 hr; p < 0.01), cardiac
surgical ICU readmissions (5.3% vs 1.6%; odds ratio, 0.31;
95% CI, 0.19–0.48; p < 0.01), and surgical postponements
(3.4 vs 0.3 per mo; p < 0.01) were lower with the intensivist
model.
Conclusions: A transition to a 24/7 in-house intensivist care
model was associated with a reduction in postoperative major
complications, duration of mechanical ventilation, cardiac sur-
gical ICU readmissions, and surgical postponements. These
findings suggest that 24/7 intensivist physician care models
may improve patient outcomes and health resource utilization
in specialized cardiac surgical ICUs. (Crit Care Med 2017;
XX:00–00)
Key Words: critical care; health resources; personnel staffing;
postoperative complications; thoracic surgery
C
ardiac surgery improves patient survival and quality-
of-life for selected coronary, valvular and structural
heart diseases (1–4). The median age, burden of
comorbidities, and perioperative risk of patients undergoing
cardiac surgery have risen over the last 2 decades (5, 6). Impor-
tantly, these factors are associated with an increase in perioper-
ative complications, including hospital-acquired pneumonia,
duration of cardiac surgical ICU (CSICU) stay and hospital-
ization, and mortality (7–10). Although the clinical and man-
agement risk factors for postoperative complications have been
described, little is known about the association between care
delivery models and the occurrence of postoperative complica-
tions and health resource utilization (11).
Benoit et al
A closed ICU operational model is supported by critical care
guidelines; however, nighttime intensivist staffing has not been
shown to improve outcomes in general medical and surgical
ICUs (12–15). Whether these findings are applicable to special-
ized CSICUs is not clear. To date, a single study has reported
that 24/7 in-house intensivist management of CSICU patients
is associated with a reduced CSICU length of stay (LOS) and
lower rates of blood product utilization (16). Importantly, this
study coincided with the transition from a mixed ICU to the
opening of a specialized CSICU, thus introducing the possibil-
ity of a comparator bias. Additionally, the study evaluated a
limited scope of postoperative complications, did not describe
the effect on surgical postponement rates, or adjust for tem-
poral changes in surgical volume and CSICU bed availability.
In 2013, policy changes at our institution removed CSICU
nighttime resident coverage and staff intensivists provided
24/7 in-house coverage. We hypothesized that 24/7 in-house
intensivist coverage would reduce postoperative complica-
tions, resource utilization, and improve clinical outcomes. We
evaluated the association between this operational transition
and outcomes in a cardiac surgical registry.
MATERIAL AND METHODS
Study Design, Setting, and Population
In a propensity-matched before-and-after design, all patients
greater than 18 years old who underwent cardiac surgery
and were admitted to the CSICU at the University of Alberta
Hospital, Mazankowski Alberta Heart Institute, Edmonton,
Alberta, Canada between January 1, 2006, and December 31,
2014, were eligible for inclusion. The hospital is the only car-
diac surgical center for a large geographic region and serves
as a quaternary cardiac surgical referral center for three prov-
inces and three territories in western Canada. It performed an
average of 114 cardiac surgeries per month during the study
period. Postoperative patients are routinely admitted to a spe-
cialized closed-model CSICU with a 1:1 nursing-to-patient
ratio. Medical coverage is provided by a weekly team of three
intensivists (who have completed Royal College of Physicians
and Surgeons of Canada accredited critical fellowships and
were certified in Critical Care Medicine) in collaboration with
the attending surgeon and CSICU nurse practitioners (avail-
able 07:00–19:00). Surgical residents were withdrawn from
CSICU call due to changes in duty hour regulations. Therefore,
the CSICU transitioned to 24/7 in-house intensivist model on
May 1, 2013. The “before” cohort (nighttime resident coverage
model) included patients admitted to CSICU between January
1, 2006, and April 30, 2013. During this time period, intensivists
provided daytime (07:00–18:00) patient care, whereas resident
physicians provided nighttime coverage (18:00–07:00) with
on-call intensivists available by telephone. The “after” cohort,
or 24/7 in-house intensivist staffing model, included patients
admitted to CSICU between August 1, 2013, and December 31,
2014. During this time period, intensivists remained in hospi-
tal and were first responders for all aspects of patient care at
nighttime. Patients admitted during the transition period (May
2 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX
Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
1 to July 31, 2014) were excluded. This study was approved by
the University of Alberta Health Research Ethics Board prior to
commencement (File no. Pro00054119). The requirement for
written informed consent was waived.
Data Sources and Outcomes
The study dataset was created by linking two prospectively col-
lected clinical registries using unique patient identifiers. First,
the Alberta Provincial Project for Outcomes Assessment in
Coronary Heart Disease database prospectively collects cardiac
catheterization and cardiac surgical information on all patients
in the province of Alberta, Canada (17). Trained staff from the
cardiac catheterization laboratory and operating room enter
demographic, medical, and procedural variables at the time of
the procedure. Chart abstracters extract postoperative CSICU
variables and complications using the Society of Thoracic Sur-
geons database definitions and review all previously entered
preoperative and intraoperative variables (18). Mortality
information is obtained through a quarterly data linkage with
the Alberta Bureau of Vital Statistics. Second, the University
of Alberta Cardiac Surgery Administrative Database, wherein
data are recorded by institutional cardiac surgical nurse man-
agers, captures CSICU occupancy, monthly number of oper-
ational CSICU beds, and number of, and reason for, cardiac
surgical postponements. This latter dataset provides deidenti-
fied aggregate data.
The primary outcome was a composite of major postop-
erative complications occurring in the CSICU including myo-
cardial infarction (MI), stroke, acute kidney injury (AKI),
nosocomial infection, reintubation or prolonged mechanical
ventilation (> 24 hr), gastrointestinal hemorrhage or isch-
emia, delirium, resuscitated cardiac arrest, and reoperation.
Secondary clinical outcomes included occurrence of indi-
vidual postoperative complications, duration of mechanical
ventilation, CSICU mortality, and 30-day mortality. Secondary
health service utilization outcomes included all-cause CSICU
readmission, CSICU LOS, and surgical postponement due to
lack of CSICU bed availability.
Given the recognized associations between age and cardio-
pulmonary bypass (CPB) time on postoperative events, we
performed prespecified subgroup analyses on patients greater
than 75 years old and patients receiving CPB for greater than
150 minutes.
Statistical Analysis
Categorical variables were reported as adjusted proportions;
continuous variables were reported as means (sds). Categorical
variables were compared using chi-square test, and continuous
variables were compared using Student t test or the Mann-
Whitney U test, as appropriate. A propensity-matched analysis
was performed due to significant temporal changes in case-mix
during the study period. To obtain a comparable distribution
of demographics, comorbidities, and clinical variables among
postoperative patients in the CSICU who were in the resident
model cohort compared with patients in the intensivists model
cohort, we used the Rosenbaum and Rubin propensity-score

matching technique (19). The propensity score was calculated
as the probability of being in the intensivist cohort conditional
on the observed baseline (measured at the time of angiogra-
phy) characteristics. This technique allows for a high number
of confounding variables and has been used to create strata of
subjects who can be matched on the propensity score whereby
exposure (i.e., 24/7 in-house intensivist or nighttime resident
model) is not confounded with measured baseline covariates.
The propensity score was calculated using logistic regression.
The following variables, selected on the basis of known clinical
associations with postoperartive outcomes and statistical dif-
ferences between the unmatched cohorts, were included in the
model: age, sex, hypertension, dyslipidemia, diabetes, prior MI,
prior percutaneous coronary intervention, prior cardiac sur-
gery, heart failure, chronic pulmonary disease, current smoker,
past smoker, cerebrovascular disease, chronic dialysis, hepatic
impairment or gastrointestinal disease, malignancy, peripheral
vascular disease, fibrinolytic therapy during current hospital-
ization, coronary anatomy, and the ejection fraction. Greedy
matching techniques were applied to match patients in the
intensivist model to patients in the resident model by match-
ing the patients with the nearest propensity score (i.e., within
two decimal places of the propensity score) for each case. Over-
lap of propensity scores between patients in both cohorts were
evaluated using histograms, chi-square values, and probability
values. The distribution of propensity scores between cohorts
is provided in Appendix 1 (Supplemental Digital Content 1,
http://links.lww.com/CCM/C579). Differences in baseline fac-
tors between groups were calculated before and after propen-
sity adjustment to assess balance (20). Covariates were well
balanced following matching (Table 1). To estimate the asso-
ciations between care model and all outcomes, logistic regres-
sion was used and aortic cross-clamp time was included as a
covariate to adjust for potential confounding of differences
between cross-clamp times of the propensity-matched groups.
Associations are presented as adjusted odds ratios and 95% CIs.
Health service utilization outcomes were adjusted for monthly
surgical volume and the number of CSICU beds using logistic
regression. To compare observed versus predicted rates of the
primary outcome in the 2014 during 24/7 in-house coverage
model), linear regression models with year as the only predic-
tor were used to model the rates observed between 2006 and
2013 and predict the rates in 2014. T test was used to compare
observed and predicted rates in 2014. Data analyses were per-
formed using the SPSS (Statistical Package for the Social Sci-
ences; IBM, Armonk, NY) data management system, version 23.
RESULTS
In the final study population, a total of 1,509 patients in the
nighttime resident cohort were propensity matched 1:1 to
1,509 patients in the 24/7 in-house intensivist cohort. The
cohorts were generally well balanced for baseline demograph-
ics, clinical characteristics, and preoperative characteristics
(Table 1). The prevalence of dyslipidemia was higher in the res-
ident cohort, and nonobstructive coronary artery disease was
higher in the intensivist cohort. Operative variables are listed
Critical Care Medicine www.ccmjournal.org 3
Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
in Table 2. The cohorts were well balanced in surgical proce-
dures, incidence, and priority. Patients in the resident cohort
had statistically longer durations of CPB and higher rates of
intraoperative RBC transfusion compared with the intensivist
cohort (Table 2).
Clinical Outcomes
The primary composite endpoint of major postoperative com-
plications was significantly lower during the intensivist model
(26.3% vs 19.3%; OR, 0.73; 95% CI, 0.63–0.85; p < 0.01) com-
pared with the resident model (Table 3 and Fig. 1). Specifically,
the intensivist model was associated with significantly lower
rates of surgical site and other nosocomial infections (e.g.,
pneumonia), and cardiac arrest compared with the resident
model. The intensivist model was associated with a significant
reduction in average duration of postoperative mechanical
ventilation (25.2 vs 19.4 hr; p < 0.01). No significant differences
were observed in CSICU or 30-day mortality. To evaluate if
temporal improvements in care were responsible for observed
improvement, a sensitivity analysis comparing the observed
versus predicted complication rates was performed. The 2014
primary composite outcome was significantly lower than pre-
dicted (17.7% vs 23.3%; p = 0.002). The annual incidence of
CSICU complications are presented in Appendix 2 (Supple-
mental Digital Content 1, http://links.lww.com/CCM/C579).
Health Services Utilization Outcomes
The mean number of operational CSICU beds (19.3 vs 21.8;
p < 0.01) and the monthly number of surgical cases (109 vs
119; p < 0.01) were higher in the intensivist cohort compared
with resident cohort. However, average daily CSICU occu-
pancy rates (95.7% vs 92.7%; p < 0.01) were lower during the
intensivist model period compared with resident cohort. All
health services utilization outcomes adjusted for these afore-
mentioned variables.
The intensivist model was associated with fewer CSICU
readmissions (5.3% vs 1.6%; OR, 0.31; 95% CI, 0.19–0.48;
p < 0.001) and shorter average CSICU LOS (4.4 vs 5.7 d;
p = 0.015) compared with the resident model. No differences
were observed in the hospital LOS (11.5 vs 10.6 d; p = 0.127).
Surgical postponements attributed to a lack of CSICU bed
availability were significantly lower during intensivist model
period (3.4 vs 0.3 postponements/mo; p < 0.01) compared
with resident model.
Subgroup Analyses
In prespecified subgroup analyses of the primary outcome,
the occurrence of major postoperative complications was sig-
nificantly reduced with the intensivist model compared with
the resident model for patients less than 75 years old (17.8%
vs 25.4%; p < 0.01); however, there were no significant differ-
ences for patients greater than or equal to 75 years old (29.7%
vs 25.5%; p = 0.26). There were no significant between-group
differences in patients with CBP time less than 150 minutes
(18.1% vs 15.9%; p = 0.16), CPB time greater than 150 minutes
(44.4% vs 60.2%; p = 0.26).
Benoit et al

TABLE 1. Baseline Demographics and Clinical Characteristics of Patients Stratified by

Nighttime Resident Coverage and 24/7 In-House Intensivist Staffing Models
Nighttime 24/7 In-House
Resident Model Intensivist Model
Variables (n = 1,509) (n = 1,509) p

Demographics
  Age (yr), mean (sd) 62.5 (14.9) 62.4 (14.4) 0.90
 Female, n (%) 347 (23.0) 371 (24.6) 0.31
  Body mass index (kg/m2), mean (sd) 29.2 (4.5) 29.4 (4.6) 0.12
Past medical history, n (%)
  Diabetes mellitus 374 (24.8) 381 (25.3) 0.93
 Hypertension 1,086 (72.0) 1,014 (67.2) 0.01
 Dyslipidemia 1,271 (84.2) 1,111 (73.6) < 0.01
  Current/former smoking 880 (58.3) 880 (58.3) 1.0
  Prior myocardial infarction > 90 d 429 (28.4) 404 (26.8) 0.31
  In-hospital fibrinolytic therapy 31 (2.1) 38 (2.5) 0.39
  Prior percutaneous coronary intervention 218 (14.4) 218 (14.4) 1.00
  Prior coronary artery bypass grafting 43 (2.8) 40 (2.7) 0.74
  Canadian Cardiovascular Society class IV 374 (24.8) 375 (24.9) 0.97
  Heart failure 177 (11.7) 176 (11.7) 0.96
  Cerebrovascular disease 109 (7.2) 108 (7.2) 0.94
  Peripheral vascular disease 38 (2.5) 48 (3.2) 0.27
  Chronic lung disease 412 (27.3) 441 (29.2) 0.24
  Chronic dialysis 22 (1.5) 28 (1.9) 0.39
  Hepatic impairment 20 (1.3) 12 (0.8) 0.16
  Chronic gastrointestinal disease 31 (2.5) 22 (1.7) 0.21
 Malignancy 28 (2.3) 23 (1.8) 0.48
Preoperative investigations
  Extent of coronary stenosis > 70%, n (%)
   Left main or three vessel coronary artery disease 572 (37.9) 551 (36.5) 0.12
  1–2 vessel 260 (17.2) 207 (13.7)
  Not available 470 (31.1) 431 (28.6)
   No obstructive coronary lesion 207 (13.7) 320 (21.2) 0.09
  Left ventricular ejection fraction, n (%) 0.34
  > 50% 349 (23.1) 321 (21.3)
  35–50% 166 (11.0) 149 (9.9)
  20–34% 34 (2.3) 24 (1.6)
  < 20% 9 (0.6) 8 (0.5)
  Not available 951 (63.0) 1,007 (66.7)
   Serum creatinine (µmol/L) (sd) 98.5 (31.9) 99.8 (50.4)

4 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Clinical Investigation

TABLE 2. Operative Variables of Patients Stratified by Nighttime Resident Coverage and

In-House 24/7 Intensivist Staffing Models
Nighttime 24/7 In-House
Resident Model Intensivist Model
Variables (n = 1,509) (n = 1,509) p

Surgical priority, n (%) 0.313

 Emergent 99 (6.5) 96 (6.4)
  Urgent in-hospital 505 (33.5) 532 (35.3)
 Outpatient 905 (59.9) 878 (58.4)
Surgical incidence, n (%) 0.61
  First operation 1,303 (86.3) 1,321 (87.5)
  Second or greater 206 (13.6) 188 (12.4)
Surgery performed, n (%) 0.93
  Isolated CABG 442 (29.3) 460 (30.5)
  CABG and valve repair/replacement 406 (26.9) 384 (25.4)
  Valve surgery only 181 (12.0) 178 (11.8)
  Non-CABG (2+ procedures) 319 (21.1) 317 (21.0)
  Atrial septal defect/ventricular septal defect/aortic surgery 43 (2.8) 43 (2.8)
  Heart or lung transplant 118 (7.8) 127 (8.4)
Intraoperative events
  Cardiopulmonary bypass time (min), mean (sd) 137.8 (64.8) 125.7 (43.5) < 0.01
  Aortic cross clamp time (min), mean (sd) 95.9 (49.8) 91.0 (40.0) 0.01
  Intraoperative RBC transfusion, n (%) 0.95 (2.2) 0.47 (1.5) < 0.01
  Intraoperative fresh frozen plasma transfusion, n (%) 0.58 (1.6) 0.48 (1.7) 0.06
CABG = coronary artery bypass grafting.

DISCUSSION preventable ICU-related complications has not been described

In this large observational cohort study comparing a model of
24/7 in-house intensivist only and nighttime resident cover-
age in a dedicated CSICU, we found several important find-
ings that may have implications for the nighttime coverage,
organizational structure, and CSICU operations. First, the
24/7 in-house intensivist model was associated with a reduc-
tion in postoperative major complications, including surgical
site infection, nosocomial pneumonia, and rates of cardiac
arrest. However, we did not find differences in 30-day mortal-
ity. Second, the transition to a 24/7 in-house intensivist model
was associated with fewer unplanned CSICU readmissions and
surgical postponements due to a lack of CSICU bed capacity.
Demographic changes in the cardiac surgical population
have translated into increased rates of in perioperative compli-
cations (5, 11). In a contemporary cohort of cardiac surgical
patients approximately, 15.9% experience a major postopera-
tive complication (7). Daily rounds by an intensivist have been
associated with a reduction in perioperative complications in
vascular surgery patients; however, the association between
a dedicated 24/7 in-house intensivist model and potentially
in a CSICU population (13). In this analysis, we observed that
24/7 in-house intensivist management was associated with
reduced nosocomial infections, shorter duration of mechanical
ventilation, cardiac arrests, and a trend toward lower rates of
AKI. Although the causal mechanisms leading to shorter dura-
tion of mechanical ventilation are uncertain, our findings are
consistent with previous studies and we hypothesize that on-site
presence of an intensivist with advanced airway and mechanical
ventilation knowledge facilitated earlier patient extubation (16,
21). In addition, recognizing the association between duration
of mechanical ventilation and ventilator-acquired pneumonia,
early postoperative extubation may translate into reduced risk
for pneumonia (11, 22). Reduced surgical site infection rates
were observed in the 24/7 intensivists cohort. We postulate that
improved adherence to, and appropriate dosing of, periopera-
tive prophylactic antibiotics, may in part, mediate the lower
observed infection rates; however, the limited scope of infection
control practices within this registry should preclude causal
inferences. Our findings that 24/7 in-house intensivist manage-
ment also reduced rates cardiac arrest and numerically lower

Critical Care Medicine www.ccmjournal.org 5

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Benoit et al

TABLE 3. Clinical and Health Services Outcomes Stratified by Nighttime Resident

Coverage and In-House 24/7 Intensivist Staffing Models
Nighttime 24/7 In-House
Resident Model Intensivist Model
Outcomes (n = 1,509) (n = 1,509) p ORa (95% CI)

Primary outcome, n (%)

  Any major complication 397 (26.3) 291 (19.3) < 0.01 0.73 (0.63–0.85)
Secondary clinical outcomes, n (%)
  Respiratory complications
   Prolonged mechanical ventilation (> 24 hr) 259 (17.2) 193 (12.8) < 0.01 0.76 (0.62–0.93)
  Reintubation 78 (5.2) 60 (4.0) 0.27 0.82 (0.58–1.16)
  Infectious complications, n (%)
  Pneumonia 133 (8.8) 86 (5.7) < 0.01 0.68 (0.51–0.90)
  Wound complication, n (%)
   Superficial sternal wound infection 54 (3.6) 32 (2.1) 0.049 0.64 (0.41–0.99)
   Deep sternal wound infection 14 (0.9) 5 (0.3) 0.06 0.37 (0.13–1.05)
   Saphenous vein harvest site infection 35 (2.3) 15 (1.0) 0.01 0.46 (0.25–0.84)
  Cardiovascular complications, n (%)
  Cardiac arrest 41 (2.7) 22 (1.5) 0.03 0.56 (0.33–0.95)
  Myocardial infarction 15 (0.8) 10 (0.7) 0.67 0.84 (0.36–1.94)
  Neurologic complication, n (%)
   Stroke (ischemic or hemorrhagic) 0 (0) 5 (0.3) 0.03 Not applicable
  Delirium 74 (4.9) 84 (5.6) 0.41 1.20 (0.87–1.66)
  Renal complication, n (%)
   Acute kidney injury 117 (7.8) 82 (5.4) 0.06 0.68 (0.56–1.01)
  RRT 48 (3.2) 38 (2.5) 0.27 0.85 (0.55–1.31)
  Intermittent dialysis 12 (0.08) 10 (0.07) 0.77 0.88 (0.38–2.06)
  Continuous RRT 48 (3.2) 48 (3.2) 0.65 0.90 (0.58–1.41)
  Gastrointestinal complication, n (%)
  Gastrointestinal hemorrhage 31 (2.1) 34 (2.3) 0.57 1.15 (0.70–1.90)
 Mortality, n (%)
   Cardiac surgical ICU mortality 29 (1.9) 38 (2.5) 0.07 1.59 (0.96–2.64)
   Death within 30 d of surgery 47 (3.1) 54 (3.6) 0.18 1.13 (0.88–1.97)
OR = odds ratio, RRT = renal replacement therapy.
a 
Odds ratios are reported as 24/7 in-house intensivist versus resident coverage models. Controlling for aortic cross clamp time.

rates of AKI are novel. It is plausible that improved hemody-
namic management, including the early recognition, or timely
treatment of antecedent events such as volume depletion or
bleeding by an experienced intensivist may have reduced these
complications (11, 21). This hypothesis is supported by prior
work showing that the majority of delays in recognition and
treatment of perioperative complications are preventable (23).
The 24/7 in-house intensivist model was associated with
a reduction in the rate of CSICU readmissions and surgical
postponements due to CSICU bed availability. Previous studies
have reported a strong association between postoperative com-
plications, nosocomial infections, and all-cause ICU readmis-
sions (24, 25). Furthermore, patients with unplanned CSICU
readmissions have longer CSICU stay and higher mortality (26,
27). The aforementioned reductions in postoperative compli-
cations with the 24/7 in-house intensivist model may be the
putative mechanism underpinning the improved operational
efficiency, observed reduction in readmissions and, thereby,

6 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Clinical Investigation

association between 24/7 inten-

sivist staffing models and blood
transfusion rates has been pre-
viously described (16). Second,
changes in surgical techniques,
advances in intensive care
treatment, and other potential
unmeasured variables may not
be accounted for in the propen-
sity-matching. Additionally, the
observational design using data
from a single center, includ-
ing different durations of the
before and after cohorts, pre-
cludes causal inferences and the
results should be considered
hypothesis generating. Finally,
this study was not powered to
detect differences in mortal-
ity and did not aim to evaluate
the economic impact of either
model, thus we cannot com-
ment on the cost-effectiveness
of such a change in nighttime
coverage models.

Figure 1. Odds ratio (OR) plot of postoperative complications in nighttime resident and 24/7 in-house
intensivist staffing models. CSICU = cardiac surgical ICU.
bed utilization. This outcome should be interpreted with cau-
tion given neither the antecedent events nor the cause of read-
mission were specifically adjudicated. Notwithstanding, we
observed no differences in CSICU LOS, which is in contrast to
a previous study (16). Whether this finding is due to variation
in institutional practice patterns, surgical case-mix, or patient
flow management remains unclear. In an era of rising preva-
lence and costs of cardiovascular disease and strained critical
care capacity, these findings present opportunities to improve
health services delivery and resource utilization (23, 28).
CONCLUSIONS
In a contemporary special-
ized CSICU patient population, we observed that a transition
to a 24/7 in-house intensivist coverage was associated with a
reduction in major postoperative complications, duration of
mechanical ventilation, CSICU readmissions, and surgical
postponements due to lack of CSICU bed availability when
compared with nighttime resident coverage. These findings
suggest in specialized CSICUs, 24/7 in-house intensivist cov-
erage by experienced professionals reduces patient morbidity
and improves resource utilization and operational efficiency.

Implications for Policy and Practice ACKNOWLEDGMENTS

Nighttime intensivist staffing has not been shown to improve
mortality in medical and surgical ICUs (14, 29). The results
from these aforementioned multidisciplinary ICUs may not
be generalizable to specialized CSICUs that care for uniquely
complex populations of patients undergoing cardiothoracic
surgery, heart and/or lung transplantation, extracorporeal
membrane oxygenation, and mechanical circulatory support.
Our results suggest that a 24/7 in-house intensivist model may
improve clinical outcomes and resource efficiency in a special-
ized CSICU setting. Future studies should be directed at con-
firming our findings across multiple sites and evaluating the
potential cost implications associated with this model (30–32).
Limitations
There are several limitations to this study. First, data on blood
product utilization in CSICU were not available; however, the
We thank Ms. L. Soulard for copyediting the article.
REFERENCES
1. Farkouh ME, Domanski M, Sleeper LA, et al; FREEDOM Trial Inves-
tigators: Strategies for multivessel revascularization in patients with
diabetes. N Engl J Med 2012; 367:2375–2384
2. Mohr FW, Morice MC, Kappetein AP, et al: Coronary artery bypass
graft surgery versus percutaneous coronary intervention in patients
with three-vessel disease and left main coronary disease: 5-year
follow-up of the randomised, clinical SYNTAX trial. Lancet 2013;
381:629–638
3. Hillis LD, Smith PK, Anderson JL, et al: 2011 ACCF/AHA Guide-
line for Coronary Artery Bypass Graft Surgery: Executive summary:
A report of the American College of Cardiology Foundation/Ameri-
can Heart Association Task Force on Practice Guidelines. Circulation
2011; 124:2610–2642
4. Nishimura RA, Otto CM, Bonow RO, et al; American College of
Cardiology/American Heart Association Task Force on Practice
Guidelines: 2014 AHA/ACC guideline for the management of

Critical Care Medicine www.ccmjournal.org 7

Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Benoit et al
patients with valvular heart disease: Executive summary: A report
of the American College of Cardiology/American Heart Association
Task Force on Practice Guidelines. J Am Coll Cardiol 2014; 63:
2438–2488
5. Ferguson TB, Hammill BG, Peterson ED, et al: A decade of change—
risk profiles and outcomes for isolated coronary artery bypass grafting
procedures, 1990–1999: A report from the STS National Database
Committee and the Duke Clinical Research Institute. Ann Thorac
Surg 2002; 73:480–490
6. ElBardissi AW, Aranki SF, Sheng S, et al: Trends in isolated coro-
nary artery bypass grafting: An analysis of the Society of Thoracic
Surgeons adult cardiac surgery database. J Thorac Cardiovasc Surg
2012; 143:273–281
7. Rahmanian PB, Kröner A, Langebartels G, et al: Impact of major non-
cardiac complications on outcome following cardiac surgery proce-
dures: Logistic regression analysis in a very recent patient cohort.
Interact Cardiovasc Thorac Surg 2013; 17:319–326
8. Atoui R, Ma F, Langlois Y, et al: Risk factors for prolonged stay in the
intensive care unit and on the ward after cardiac surgery. J Card Surg
2008; 23:99–106
9. Shahian DM, O’Brien SM, Filardo G, et al; Society of Thoracic Sur-
geons Quality Measurement Task Force: The Society of Thoracic
Surgeons 2008 cardiac surgery risk models: Part 1–coronary artery
bypass grafting surgery. Ann Thorac Surg 2009; 88:S2–22
10. O’Brien SM, Shahian DM, Filardo G, et al; Society of Thoracic Sur-
geons Quality Measurement Task Force: The Society of Thoracic
Surgeons 2008 cardiac surgery risk models: Part 2–isolated valve
surgery. Ann Thorac Surg 2009; 88:S23–S42
11. Gelijns AC, Moskowitz AJ, Acker MA, et al; Cardiothoracic Surgical
Trials Network (CTSN): Management practices and major infections
after cardiac surgery. J Am Coll Cardiol 2014; 64:372–381
12. Haupt MT, Bekes CE, Brilli RJ, et al; Task Force of the American Col-
lege of Critical Care Medicine, Society of Critical Care Medicine:
Guidelines on critical care services and personnel: Recommenda-
tions based on a system of categorization of three levels of care. Crit
Care Med 2003; 31:2677–2683
13. Pronovost PJ, Jenckes MW, Dorman T, et al: Organizational character-
istics of intensive care units related to outcomes of abdominal aortic
surgery. JAMA 1999; 281:1310–1317
14. Kerlin MP, Small DS, Cooney E, et al: A randomized trial of night-
time physician staffing in an intensive care unit. N Engl J Med 2013;
368:2201–2209
15. Wallace DJ, Angus DC, Barnato AE: Nighttime intensivist staffing and
mortality among critically ill patients. N Eng J Med 2012; 367:881
16. Kumar K, Zarychanski R, Bell DD, et al; Cardiovascular Health
Research in Manitoba Investigator Group: Impact of 24-hour in-house
intensivists on a dedicated cardiac surgery intensive care unit. Ann
Thorac Surg 2009; 88:1153–1161
17. Ghali WA, Knudtson ML: Overview of the Alberta Provincial
Project for outcome assessment in coronary heart disease. On
behalf of the APPROACH investigators. Can J Cardiol 2000; 16:
1225–1230
8 www.ccmjournal.org XXX 2017 • Volume XX • Number XXX
Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
18. The Society of Thoacic Surgeons: DataCollection. Available at: http://
www.sts.org/sts-national-database/database-managers/adult-cardiac-
surgery-database/data-collection. Accessed December 16, 2016
19. Rosenbaum PR, Rubin DB: Constructing a control group using mul-
tivariate matched sampling methods that incorporate the propensity
score. Am Stat 1985; 39:33–38
20. Guo S, Fraser MW: Propensity Score Analysis Statistical Methods
and Applications. New York, Sage Publications, 2015
21. Netzer G, Liu X, Shanholtz C, et al: Decreased mortality resulting from
a multicomponent intervention in a tertiary care medical intensive care
unit. Crit Care Med 2011; 39:284–293
22. The American Thoracic Society and Infectious Diseases Society of
America: Guidelines for the management of adults with hospital-
acquired, ventilator-associated, and healthcare-associated pneumo-
nia. Am J Respir Crit Care Med 2005; 171:388–416
23. Shannon FL, Fazzalari FL, Theurer PF, et al; Michigan Society of Thoracic
and Cardiovascular Surgeons: A method to evaluate cardiac surgery mor-
tality: Phase of care mortality analysis. Ann Thorac Surg 2012; 93:36–43
24. Cannon MA, Beattie C, Speroff T, et al: The economic benefit of
organizational restructuring of the cardiothoracic intensive care unit.
J Cardiothorac Vasc Anesth 2003; 17:565–570
25. Cheng DC, Karski J, Peniston C, et al: Early tracheal extubation after
coronary artery bypass graft surgery reduces costs and improves
resource use. A prospective, randomized, controlled trial. Anesthesi-
ology 1996; 85:1300–1310
26. Greco G, Shi W, Michler RE, et al: Costs associated with health care-
associated infections in cardiac surgery. J Am Coll Cardiol 2015;
65:15–23
27. Reason J: Human error: Models and management. BMJ 2000;
320:768–770
28. Joskowiak D, Wilbring M, Szlapka M, et al: Readmission to the inten-
sive care unit after cardiac surgery: A single-center experience with
7105 patients. J Cardiovasc Surg (Torino) 2012; 53:671–676
29. Litwinowicz R, Bartus K, Drwila R, et al: In-hospital mortality in cardiac
surgery patients after readmission to the intensive care unit: A single-
center experience with 10,992 patients. J Cardiothorac Vasc Anesth
2015; 29:570–575
30. Halpern SD: ICU capacity strain and the quality and allocation of criti-
cal care. Curr Opin Crit Care 2011; 17:648–657
31. Heidenreich PA, Trogdon JG, Khavjou OA, et al; American Heart Asso-
ciation Advocacy Coordinating Committee; Stroke Council; Council on
Cardiovascular Radiology and Intervention; Council on Clinical Cardiol-
ogy; Council on Epidemiology and Prevention; Council on Arterioscle-
rosis; Thrombosis and Vascular Biology; Council on Cardiopulmonary;
Critical Care; Perioperative and Resuscitation; Council on Cardiovascu-
lar Nursing; Council on the Kidney in Cardiovascular Disease; Council
on Cardiovascular Surgery and Anesthesia, and Interdisciplinary Coun-
cil on Quality of Care and Outcomes Research: Forecasting the future
of cardiovascular disease in the United States: A policy statement from
the American Heart Association. Circulation 2011; 123:933–944
32. Gooch RA, Kahn JM: ICU bed supply, utilization, and health care spend-
ing: An example of demand elasticity. JAMA 2014; 311:567–568