152 | www.critpathcardio.com Critical Pathways in Cardiology • Volume 13, Number 4, December 2014
ORIGINAL ARTICLE
Objective: Physicians’ gender may impact test utilization in the diagnosis of
acute cardiovascular disease. We sought to determine if physician gender affected
stress test utilization by patient gender in a low-risk chest pain observation unit.
Methods: This was a retrospective consecutive cohort study of patients
admitted to a chest pain unit in a large volume academic urban emergency
department (ED). Inclusion criteria were age >18, American Heart Association
low-to-intermediate risk, electrocardiogram nondiagnostic for acute coronary
syndrome, and negative initial troponin I. Exclusion criteria were age >75
with a history of coronary artery disease, active comorbid medical problems,
or inability to obtain stress testing in the ED for any reason. T-tests were used
for univariate comparisons and logistic regression was used to estimate odds
ratios (ORs) for receiving testing based on physician gender, controlling for
race, insurance, and Thrombolysis In Myocardial Infarction (TIMI) score.
Results: Three thousand eight hundred and seventy-three index visits were
enrolled during a 2.5-year period. Mean age was 53 ± 20, 55% (95% CI,
53–56%) were female. There was no difference in overall stress utilization
based upon physician gender (P = 0.28). However, after controlling for other
variables, male physicians had significantly lower odds of stress testing
female patients (OR
M
, 0.82; 95% CI, 0.68–0.99), whereas no difference was
found in female physicians (OR
F
, 0.80; 95% CI, 0.57–1.14).
Conclusions: Male physicians appear less likely to utilize stress testing
in female patients even after controlling for objective clinical variables,
including TIMI score. Although adverse outcomes are uncommon in this
patient cohort, further investigation into provider-specific practice patterns
based on patient gender is necessary.
Key Words: low-risk chest pain, gender, physician gender, stress test
utilization
(Crit Pathways in Cardiol 2014;13: 152–155)
I
n 2006, cardiovascular disease represented greater than 25% of all-
cause mortality, representing the leading cause of death in both men
and women in the United States.
1
Despite advances in therapy and
improvements in mortality, treatment disparities between men and
women persist in patients with acute coronary syndrome (ACS).
2–7
Studies looking for differences in resource utilization in patients with
acute chest pain being evaluated for potential ACS have had mixed
results.
4,8
Studies examining provider level differences have also had
mixed results. Recent work has indicated that provider gender does
not seem to be associated with referral for catheterization by attend-
ing physicians
9
or in the workup of patients with potential coronary
heart disease by medical students and residents.
10
However, other
studies have found an association between female gender of physi-
cians and better evidence-based guideline adherence for all of their
patients compared with male physicians.
11,12
Chest pain observations units are now a common component
of emergency care of the low-risk chest pain patient. Such units
potentially offer more objective risk stratification and utilization of
stress testing. In 2 single-center studies, we demonstrated that stress
test utilization based on physician discretion was safe, efficient, and
lowered resource utilization
8
without any evidence of gender bias.
13
The potential effect of provider level bias on the workup of nonspe-
cific chest pain, particularly in a chest pain unit (CPU), has not been
studied. As a result, we sought to evaluate whether the gender of the
physician was associated with disparities in stress utilization of male
and female patients in emergency department (ED) CPU patients.
MATERIALS AND METHODS
Study Design and Setting
This was a retrospective consecutive cohort study of patients
admitted to a CPU in a large volume academic urban ED with an
annual census of 105,000. All clinical, historical, laboratory, stress
test, and outcomes information were prospectively entered into a clin-
ical registry. The CPU is a 7-bed unit that lies within the ED and is
open 7 days a week, 24 hours a day. It is overseen by the Emergency
Physician on duty and operated by ED staff. All patients undergo a
6-hour observational protocol including serial electrocardiograms
(ECGs), serial biomarkers, and continuous telemetry. After comple-
tion of this 6-hour observation period, all patients receive consultation
by an attending cardiologist who makes the recommendation as to
the necessity and type of stress testing. Exercise stress testing, stress
nuclear imaging, stress echocardiography, and computed tomographic
(CT) coronary angiography are all available directly from the CPU.
The Institutional Review Board of the participating hospital
approved this study.
Inclusion/Exclusion Criteria
All index patients admitted to the CPU were included for analysis.
Study inclusion criteria, therefore, were equivalent to criteria for admis-
sion to the CPU. These consist of age >18, American Heart Association
(AHA) low-to-intermediate risk, ECG nondiagnostic for ACS, and a
negative initial troponin I (Bayer ACS: Centaur Troponin I, Tarrytown,
NY).
14
Patients >75 years of age with a history of coronary artery dis-
ease (CAD), those with active comorbid medical problems, or those
unable to obtain stress testing in the ED for any reason were excluded.
Study Protocol
Demographic, historical, and physical examination features
were collected on each patient using a standardized data collection
form designed before the study. We defined the primary hypothesis,
inclusion and exclusion criteria, and all desired data variables before
data abstraction. A trained chart abstractor with extensive experience
using the electronic medical record and clinical registry, who was
blinded to the hypothesis of the study, extracted the data based on the
predefined inclusion and exclusion criteria. We did not assess interra-
ter reliability because pertinent data extracted were objective and did
not require interpretation. One of the authors of the study (not blinded
to the hypothesis) confirmed that the correct filter had been applied
before data analysis. The original and serial ECGs were coded as nor-
mal, nondiagnostic, dynamic, or ischemic at the time of cardiology
Gender Disparities in Stress Test Utilization in Chest Pain Unit
Patients Based Upon the Ordering Physician's Gender
Anthony M Napoli, MD, Esther K Choo, MD, Alyson McGregor, MD
Copyright © 2014 by Lippincott Williams & Wilkins
ISSN: 1003-0117/14/1304-0152
DOI: 10.1097/HPC.0000000000000026
Received for publication June 3, 2014; accepted August 28, 2014.
From the Department of Emergency Medicine, Warren Alpert Medical School of
Brown University, Providence, RI.
Reprints: Anthony M Napoli, MD, Department of Emergency Medicine, Warren
Alpert Medical School of Brown University, 593 Eddy Street, Providence, RI.
E-mail: anapoli@lifespan.org.