Sex Preferences in Cardiovascular Testing: The Contribution of the Patient-Physician Discussion

Abstract

Objectives

Prior studies suggest that a decreased cardiac catheterization rate for women with acute coronary syndrome (ACS) is partially attributed to gender differences in patient preferences for testing. The hypothesis was that these preferences are influenced by physician recommendations for cardiovascular testing.

Methods

This was a cohort study of patients who presented to an emergency department (ED) with chest pain. At the time of ED disposition, patients were surveyed to assess whether or not the doctor discussed the possibility of coronary artery disease (CAD), the physician’s recommendations for diagnostic testing, and the patient’s preferences for further testing. The survey was repeated at the time of discharge from the hospital if the patient was admitted, and again at 30 days for all patients. The main outcome was patient-reported physician recommendation for testing, and the secondary outcome was patient preference for cardiovascular testing.

Results

There were 206 patients enrolled (118 women, 88 men). Women were less likely than men to receive cardiac catheterization as inpatients (9.8% vs. 20.0%, p = 0.04). In the ED, women were less likely to be recommended for stress testing (8.5% vs. 19.3%, p =0.02), cardiac catheterization (4.2% vs. 13.6%, p = 0.02), or to see a cardiologist (8.5% vs. 22.7%, p < 0.01). As inpatients, women were more likely to be told that no further testing was needed (70.5% vs. 50.0%, p = 0.03). While there was higher incidence of prior CAD and MI among the men in the study, analysis after removal of these patients did not alter results. Physicians were less likely to counsel women about diagnostic testing options in the ED (10.1% vs 22.7%, p = 0.03), as inpatients (11.5% vs. 40.0%, p < 0.01), and as outpatients (26.1% vs. 48.6%, p = 0.04). No patients in the study refused their doctors’ recommendations. Women were less likely to prefer catheterization in the ED (5.1% vs 15.9%, p = 0.01), and were more likely to prefer no further testing as inpatients (60.7% vs. 40.0%, p = 0.02).

Conclusions

Women who presented to the ED with symptoms concerning for ACS reported lower rates of physician recommendation for cardiovascular testing, as well as lower rates of counseling regarding cardiac etiologies of their chest pain. These findings suggest sex differences in preference for cardiovascular testing may be partially explained by the discussions between women and their doctors.

INTRODUCTION

Cardiovascular disease has affected more women than men every year for almost two decades, claiming the life of a woman every minute in the United States.¹ Despite this increased risk of mortality, studies have continually showed delayed treatment in women with coronary artery disease (CAD), particularly with invasive procedures such as diagnostic catheterization.^2–9 Women are also less likely to receive front-line intervention such as electrocardiogram (ECG), oxygen saturation measurement, and evidence-based pharmacotherapies.^10–12 Research has identified a number of explanations for these discrepancies, which can be categorized into three general themes: quantifiable clinical differences, physician bias, and patient preferences.

Sex disparities in intervention for CAD were first attributed to a number of differences in the clinical presentation and disease course.^13–24 Subsequent investigations, however, have shown that these differences in clinical presentation and non-invasive testing results do not fully account for the disparities in referral for percutaneous coronary intervention (PCI). ^3,24–27 Chang et al. examined 3,514 patients with potential acute coronary syndrome (ACS), and found that men were more likely to be referred for cardiac catheterization (OR 1.72, 95% CI = 1.40 to 2.11) after controlling for differences in presenting symptoms, history, risk factors, electrocardiogram (ECG) reading, and diagnosis.³ Schulman et al. studied primary care physicians (PCPs) who watched videos of standardized patients with chest pain, and found that after controlling for history and symptoms, physicians were less likely to refer women for cardiac catheterization.²⁵ We are unaware of any study to date that has attempted to gather data on physician recommendations from real patient encounters.

In today’s health care climate, physicians may not be the only population driving discrepancies in referral for PCI. Investigation has suggested women may be less likely to undergo cardiac catheterization because they prefer less invasive testing.²⁸ Mumma et al. found that women who presented to the ED with chest pain said they would be less likely to accept their physicians’ recommendations for cardiac catheterization (10% difference);²⁹ however, this study did not look at actual recommendations made by physicians or if any testing was actually refused. To our knowledge, this study is the first to investigate if women are refusing physician referrals for cardiac catheterization, and how discussions between the physician and the patient influence patient preferences.

We examined patient preferences for cardiovascular testing after evaluation for chest pain in the emergency department (ED). Patients were asked to report the recommendations made by physicians regarding further cardiovascular testing. We hypothesized that women were less likely than men to prefer further testing because of differences in their physicians’ recommendations.

METHODS

Study Design and Population

We conducted an observational cohort study of patients presenting to the emergency department (ED) with chest pain to determine how doctor - patient discussions influence patient preferences for cardiovascular testing. The study was approved by the institutional review board. All subjects provided written informed consent. The study was conducted from October 2011 to March 2012 at the Hospital at the University of Pennsylvania, which includes a 24-hour, 17-bed observational unit and cardiac catheterization facilities.

All patients with chief complaints of chest pain were screened by trained research assistants (RAs) in the ED. Patients who presented to the ED with primary complaints of chest pain were invited to enroll if ACS was on the differential diagnosis, per discussion with the emergency physician (EP). Patients were excluded if they were under 35 years of age, displayed ECG changes consistent with acute myocardial infarction, were medically unstable, or were unable to answer survey questions due to altered mental status or English proficiency. Pregnant patients were excluded because testing would be modified in this cohort. Inclusion criteria were kept intentionally broad so that results would be generalizable to patients presenting to the ED with CAD-related symptoms.

Survey Content and Administration

The survey was designed to capture conversations between patient and physician, and was developed through input from EPs, hospitalists, and experienced RAs (see Data Supplement S1). The patient was asked if his or her physician was considering CAD as a cause of the symptoms, and what recommendations were made for further testing. The survey evaluated the patient’s perception of the discussion by measuring his or her degree of satisfaction with these recommendations (on a scale of 0 to 10), if he or she refused any of these recommendations, and his or her own preference for testing. We also assessed quality measures of the encounter by asking the patient, on a scale of 0 to 10, how well the doctor understood the symptoms and preferences for testing, answered any questions, and counseled on testing options and risks of heart disease. The survey was developed specifically for this study, as we were unaware of any validated instruments designed for this purpose. The survey was administered right before disposition from the ED to ensure discussions were completed between the patient and the EP.

We collected demographic information, cardiac history and risk factors, and presenting symptoms before administering the survey. A patient was considered to have prior CAD if he or she self-reported the diagnosis, had positive imaging prior to enrollment, or had a history of catheterization in which stents were placed. Patients who were admitted to the hospital were given the survey again before hospital discharge to evaluate any discussions with their inpatient physicians. All patients were called 30 days after enrollment to obtain interval history, and the survey was administered again if the patient saw a PCP or outpatient cardiologist in that time (see Data Supplement S2). Survey questions were asked by an RA rather than given to patients to fill out independently in order to eliminate inaccurate responses secondary to poor health education or comprehension. A single RA administered all enrollment and follow-up surveys in an effort to improve internal validity.

Testing obtained from the ED, in the hospital, and as an outpatient was verified when possible by electronic medical records in the health system, which includes three area hospitals. Diagnostic testing included stress testing, cardiac computed tomographic angiography (CCTA), and cardiac catheterization.

The main outcomes were sex differences in physician recommendation for testing, and CAD-related counseling to patients. The secondary outcome was sex differences in patient preference for cardiovascular testing while in the ED, and as inpatients.

Data Analysis

Data were imported and stored on Microsoft Access 2010 (Redmond, WA) and analyzed using Stata 12 (Stata Corp, College Station, TX). Baseline comparisons between men and women were performed with t-tests for continuous variables, chi-square tests for categorical variables, and Wilcoxon rank sum tests for scaled variables. The Mantel-Haenzsel method was used to detect confounding effects from population characteristics.

RESULTS

We screened 327 patients for enrollment; 81 patients were ineligible and 40 patients refused participation (Figure 1). We enrolled 206 patients in the ED (118 women and 88 men). One hundred and twenty-one patients were admitted to the hospital (61 women and 60 men), and we were able to contact 165 patients (81%) for 30-day follow-up (95 women and 70 men).

Figure 1.

Women and men were well matched in age, racial and ethnic background, education, and income (Table 1). Both women and men were well matched in most cardiac risk factors; however, fewer women had prior histories of CAD (19.5% in women vs. 44.3% in men), myocardial infarction (12.7% vs. 28.4%), prior catheterization (18.6% vs. 43.2%), and stenting (7.6% vs. 25%). There were no sex differences in presenting symptoms, and both women and men were most likely to have left anterior chest pain (36.4% vs. 43.2%), sharp in quality (40.7% vs. 46.7%), radiating down the left arm (22.9% vs. 22.7%), with associated shortness of breath (57.6% vs. 51.1%).

Table 1.

Characteristics of Study Population

Characteristic	Women (n = 118)	Men (n = 88)	p-value
Demographics
Age, yr (95% CI)	56.6 (54.2–59.0)	56.3 (53.7–59.0)
Race			0.1
White	40 (33.9)	38 (43.2)
African American	77 (65.3)	47 (53.4)
Other	1 (0.9)	3 (3.4)
Ethnicity			0.3
Hispanic	4 (3.4)	1 (1.1)
Non-Hispanic	114 (96.9)	67 (98.9)
Yearly Income			0.1
Unemployed	73 (61.9)	46 (52.3)
Under 20K	7 (5.9)	3 (3.4)
20K to 40K	18 (15.25)	11 (12.5)
40K to 80K	14 (11.9)	14 (15.9)
80K to 120K	3 (2.5)	8 (9.1)
Over 120K	3 (2.5)	6 (6.8)
Medical History
Risk factors for ACS
Smoking			0.2
Former	35 (29.7)	23 (26.1)
Current	25 (21.2)	28 (31.8)
Diabetes	35 (29.7)	24 (27.3)	0.7
Hypertension	69 (58.5)	54 (61.4)	0.7
High cholesterol	48 (40.1)	50 (36.8)	0.02
Coronary disease	23 (19.5)	39 (44.3)	<0.01
Prior MI	15 (12.7)	25 (28.4)	<0.01
Family Hx of CAD	62 (52.5)	31 (35.23)	0.01
Prior testing
Stress test	74 (62.7)	57 (64.7)	0.7
Within 1 year	27 (22.9)	24 (27.3)	0.8
CCTA	36 (30.8)	37 (42.1)	0.1
Within 1 year	27 (22.9)	27 (30.7)	0.1
Catheterization	22 (18.6)	38 (43.2)	<0.01
Within 1 year	6 (5.1)	20 (22.8)	<0.01

K = thousand; ACS = acute coronary syndrome; MI = myocardial infarction; CAD = coronary artery disease; Hx = history; CCTA = cardiac computer tomography angiography

Data reported as n (%) unless otherwise noted

Our primary outcome was physician recommendation for testing and CAD-related counseling (Table 2). When asked if their doctors were concerned about CAD as a cause of their chest pain, 45.8% of women and 59.1% of men responded ‘yes’ in the ED (p = 0.06), and the difference was significant at the inpatient follow-up (42.6% vs. 61.7%, p = 0.04), and at 30 days (26.1% vs. 56.8%, p = 0.01).

Table 2.

Patient-reported Clinical Recommendations and Associated Patient Satisfaction

Variable	Emergency Department			Inpatient			30-day Follow-Up
	Women (n=118)	Men (n=88)	p-value	Women (n=61)	Men (n=60)	p-value	Women (n=46)	Men (n=37)	p-value
Physician told patient he or she was considering CAD as a possible cause of patient’s chest pain	54 (45.8)	52 (59.1)	0.06	26 (42.6)	37 (61.7)	0.04*	12 (26.1)	21 (56.8)	<0.01*
Patient response to ‘What did your doctor tell you is the most likely diagnosis causing your symptoms?’
CAD	7 (5.9)	16 (18.1)	0.01*	7 (11.5)	19 (31.7)	0.01*	6 (13.0)	15 (40.5)	<0.01*
Reflux disease	9 (7.6)	1 (1.1)	0.03*	6 (9.8)	2 (3.3)	0.2	6 (13.0)	2 (5.4)	0.2
Unknown	74 (62.7)	47 (53.4)	0.2	30 (49.2)	21 (35.0)	0.1	16 (34.8)	11 (29.7)	0.6
Other	28 (23.73)	24 (27.3)	0.6	18 (29.5)	18 (30.0)	>0.9	18 (39.1)	9 (24.3)	0.2
Physician recommendations for further testing
None (observation)	76 (64.4)	45 (51.1)	0.06	43 (70.5)	30 (50.0)	0.03*	39 (84.8)	29 (78.4)	0.4
Cardiologist	10 (8.5)	20 (22.7)	<0.01*	8 (13.1)	18 (30.5)	0.02*	6 (24.3)	9 (24.3)	0.2
Stress test	10 (8.5)	17 (19.3)	0.02*	13 (21.3)	14 (23.7)	0.8	1 (2.2)	4 (10.8)	0.1
CT angiography	27 (22.9)	26 (29.6)	0.3	0 (0)	5 (8.5)	0.02*	0 (0)	0 (0)	--
Cardiac catheterization	5 (4.2)	12 (13.6)	0.02*	7 (11.5)	13 (22.4)	0.1	1 (2.2)	3 (8.11)	0.2
Mean (±SD) rating of patient satisfaction with recommendation†
None (observation)	8.5 (2.7)	8.6 (2.8)	0.8	9.0 (2.4)	9.1 (1.8)	0.7	8.5 (2.8)	8.6 (2.7)	0.9
Stress Test	9.1 (1.5)	9.4 (1.1)	0.8	9.2 (2.8)	10.0 (0)	0.1	9.1 (1.5)	9.4 (1.1)	0.6
CT Angiography	9.1 (1.6)	9.8 (0.5)	0.09	--	10.0 (0)	--	9.1 (1.6)	9.8 (0.5)	0.7
Cardiac Catheterization	7.2 (2.6)	9.2 (2.3)	0.1	9.9 (0.4)	9.9 (0.3)	0.7	7.2 (2.6)	9.2 (2.3)	0.6

^†0 = dissatisfied, to 10 = satisfied;

^*statistically significant; Data reported as n (%) unless otherwise noted

Women were less likely to be recommended to see cardiologists (8.5% vs. 22.7%, p < 0.01), receive stress testing (8.5% vs. 19.3%, p =0.02), or receive cardiac catheterization (4.2% vs. 13.6%, p = 0.02) in the ED. They were less likely to be recommended to see cardiologists (13.1% vs. 30.5%, p = 0.02) or receive CCTA (0% vs. 8.5%, p = 0.02) as inpatients, and were more likely to be told they did not need any further testing (70.5% vs. 50.0%, p = 0.03). There was no difference between men and women in their satisfaction ratings of their doctors’ recommendations, even when we accounted for sex discordance between patient and physician.

Given the increased incidence of prior CAD, myocardial infarction, and stenting among men, a sub-analysis of patients with prior cardiac history (24 women, 39 men) was performed to evaluate for any confounding effect on physicians’ recommendations. The actual crude odd ratios (ORs) were not different than the stratum-specific ones for both sexes, and tests of homogeneity were non-significant at all three time points. For example, among study subjects with previous cardiac history, the OR for a cardiologist referral from the EP for women vs. men was 0.5 (95% CI = 0.1 to 1.6); among those without CAD history, it was 0.4 (95% CI = 0.1 to 1.1). The Mantel-Haenszel combined OR (MHOR) was 0.4 (95% CI = 0.2 to 0.9), with a test for homogeneity of p = 0.8, compared to the crude OR across both strata of 0.3 (95% CI = 0.1 to 0.7). These results were also similar in regards to ED referrals for CTA (with cardiac history 1.8, 95% CI = 0.5 to 7.1; without cardiac history 0.4, 95% CI = 0.2 to 0.9; MHOR 0.6, 95% CI = 0.3 to 1.1; test for homogeneity p = 0.06, crude OR 0.7, 95% CI = 0.4 to 1.3 ), stress testing (with cardiac history 0.2, 95% CI = 0.0 to 1.8; without cardiac history 0.4, 95 % CI = 0.2 to 1.1; MHOR 0.3, 95% CI = 0.1 to 0.8; test for homogeneity p = 0.5, crude OR 0.4, 95% CI = 0.2 to 0.9), and catheterization (with cardiac history 0.5, 95% CI = 0.1 to 2.4; without cardiac history 0.2, 95% CI = 0.0 to 1.4; MHOR 0.4, 95% CI = 0.1 to 1.2; test for homogeneity p = 0.5, crude OR 0.3, 95% CI = 0.1 to 0.8).

A small proportion of patients across both sexes received any counseling or education regarding CAD in the ED; the difference between sexes was significant as inpatients (11.5% in women vs. 28.1% in men, p < 0.01) and at 30 days (8.7% vs. 51.3%, p < 0.01) (Table 3). When patients were asked specifically about discussion regarding diagnostic testing, women were less likely to report their doctors discussed cardiovascular testing options with them in the ED (10.1% vs. 22.7%, p = 0.03), as inpatients (11.5% vs. 40.0%, p < 0.01), and as outpatients (26.1% vs. 48.6%, p = 0.04).

Table 3.

Patient-Reported Qualitative Assessment of Doctor-Patient Discussion

Variable	Emergency Dept			Inpatient			30-day Follow-Up
	Women (n=118)	Men (n=88)	p-value	Women (n=61)	Men (n=60)	p-value	Women (n=46)	Men (n=37)	p-value
Physician provided counseling regarding the symptoms, risk factors, or testing used for CAD	16 (13.6)	11 (12.5)	0.8	7 (11.5)	17 (28.33)	0.02*	4 (8.7)	19 (51.3)	<0.01*
Physician understood the symptoms that brought them to the ED ‘very well’	85 (72.0)	72 (81.8)	0.07	48 (78.7)	54 (90)	0.1	35 (77.1)	31 (83.8)	0.4
Physician answered questions about CAD ‘very well’	83 (70.3)	61 (69.3)	0.9	40 (66.7)	52 (86.7)	0.01*	37 (80.4)	32 (86.5)	0.5
Physician adequately discussed diagnostic testing options for CAD	12 (10.1)	20 (22.7)	0.03*	7 (11.5)	24 (40.0)	<0.01*	12 (26.1)	18 (48.6)	0.04*
Physician asked the patient about their testing preferences	8 (6.7)	23 (26.1)	0.02*	10 (16.4)	25 (41.7)	0.01*	11 (23.9)	19 (51.4)	0.02*

Data reported as n (%)

^*Statistically significant

Our secondary outcome was patient preference for diagnostic testing (Table 4). No subjects reported they refused recommendations made by their doctors. When asked about testing preferences in the ED, women were less likely to prefer catheterization (5.1% vs, 15.9%, p = 0.01). As inpatients, women were more likely to prefer no further testing, or observation (60.7% vs. 40%, p = 0.02). When we considered whether or not these preferences matched the physician’s recommendation, there was no difference detected between sexes.

Table 4.

Patient Testing Preferences

Variable	Emergency Department			Inpatient			30-day Follow-Up
	Women (n=118)	Men (n=88)	p-value	Women (n=61)	Men (n=60)	p-value	Women (n=46)	Men (n=37)	p-value
Testing preferred by patient
None (observation)	49 (41.5)	27 (30.7)	0.1	37 (60.7)	24 (40.0)	0.02*	36 (78.3)	27 (73.0)	0.6
Stress test	21 (17.8)	15 (17.1)	0.9	14 (23.0)	16 (26.7)	0.6	6 (13.0)	3 (8.1)	0.5
CT angiography	39 (33.1)	32 (36.4)	0.6	1 (1.6)	7 (11.7)	0.03*	2 (4.4)	3 (8.1)	0.5
Catheterization	6 (5.1)	14 (15.9)	0.01*	7 (11.5)	12 (20.0)	0.2	1 (2.2)	3 (8.1)	0.2
Undecided	3 (2.5)	0	0.1	2 (3.3)	1 (1.7)	0.6	1 (2.2)	1 (2.7)	0.9
Testing preferred by patient compared to recommendations			0.8			0.3			0.5
Preferred same testing as recommendation	82 (70.1)	58 (65.9)		56 (91.8)	50 (83.3)		39 (84.8)	34 (91.9)
Preferred testing when none was recommended	31 (26.5)	27 (30.7)		5 (8.2)	9 (15.0)		6 (13.0)	3 (8.1)
Preferred no testing even when it was recommended	4 (3.4)	3 (3.4)		0 (0)	1 (1.7)		1 (2.1)	0 (0)

Data reported as n (%)

^*statistically significant

Figure 1 illustrates testing outcomes for both cohorts. In the ED, women and men received CCTA, the only test available in the department, at similar rates (15.3% in women vs 17.0% in men, p = 0.8). One patient received a stress test directly from the ED because she was ineligible for a CCTA and there were no available inpatient beds. As inpatients, women were less likely to receive any form of testing (19.7% vs. 26.7%, p = 0.01). When we analyzed each type of test separately, we found the sex difference was most pronounced for catheterizations (9.8% vs. 20.0%, p = 0.04). Only five patients (three women, two men) received catheterization as inpatients after positive stress tests or CCTA. There were no sex differences detected in testing or outpatient follow-up at 30 days, and no patients reported seeking care outside the health system when asked specifically about follow-up appointments and testing.

DISCUSSION

Our study found that women were less likely than men to report physician recommendations for cardiovascular testing including stress test, CCTA, and catheterization in the ED. Even though ACS was considered in all patients included in the study, women were less likely to be told their symptoms could be a result of heart disease, and they were more often given a likely diagnosis of reflux disease or ‘unknown’ etiology. In the inpatient setting, women were more likely than men to report that their physicians recommended no further testing by the end of their hospitalizations (and women were less likely to receive testing in the inpatient setting). As this study was not designed to look at testing results, final diagnosis, or patient outcomes, the extent to which these recommendations were appropriate cannot be evaluated. These recommendations become meaningful, however, in the context of patient preferences.

The results of our study highlight the effect that these recommendations may have on patient perceptions about cardiovascular testing. First, no patients in the study refused their doctors’ recommendations; this refutes prior research that theorized women might be receiving fewer catheterizations because they are refusing their doctors’ suggestions for PCI.²⁹ Second, both men and women most often preferred the same testing that was recommended by their physicians, and more importantly, the sex differences in patient preferences coincide with the differences we found in physician recommendations. In the ED, women were less likely to prefer catheterization, and as inpatients, women were more likely to prefer no further testing. As the first study to look at clinical discussions between patients and physicians regarding cardiovascular testing, our findings suggest these informal discussions may have a powerful effect on patient preferences for diagnostic testing. This is particularly relevant given recent research that has shown how patient preferences and shared decision making are contributing to clinical decisions more than ever before in the United States.^30–33

It is important to note that a higher proportion of men in our study population presented with prior diagnoses of heart disease. While analysis did not reveal any confounding effect of prior CAD or MI, the sample was small and it is thus difficult to assess how this may have affected results. In patients with prior CAD, we would expect physicians to be more likely to consider a cardiac cause of chest pain, and thus recommend further diagnostic testing. If we continue to stratify patients without prior testing to lower-risk categories, however, we will continue to miss the under-treated female population. Research has shown that cardiac risk factors are not predictive of ACS in the ED population, and furthermore, that physicians cannot predict the presence of ACS in ED chest pain patients based on the character of their symptoms.^34,35 Given the elusive nature of ACS, patients presenting with chest pain and no prior cardiac workup are in particularly need of counseling on risk factor modification and parameters for seeking medical attention in the future.

The lack of disease-related counseling provides another possible explanation for the sex differences in patient preference. Our results show that women were less likely to be counseled about diagnostic testing options at all three time points, and notably low proportions of both sexes were educated about CAD risk factors and symptoms. This suggests that patients, particularly women, may have been making decisions about further intervention with little knowledge about CAD or available testing modalities for the disease. Even if we assume all recommendations regarding further testing were appropriately tailored to the patients’ risk for ACS, education and increased awareness are relatively efficient, inexpensive modalities to improve CAD-related outcomes. While public awareness campaigns have shown to be an ineffective means of improving the delay to treatment in ACS, women have been substantially under-represented in these studies,^36–39 and no study to our knowledge has looked specifically at one-on-one, physician-directed education as a public awareness intervention.

Any lack of counseling is particularly concerning given prior research that has suggested women are more likely to underestimate their risk for developing heart disease,^40–42 and that their innate ‘risk aversion’ may affect testing preferences.^43–46 If these theories are true, it is even more important for physicians to provide adequate information so that education, rather than patient bias, is informing patient preferences.

In our testing outcomes, we found women were less likely to receive any form of testing. Women who were admitted to the hospital were less likely than men to receive catheterizations (Figure 1). This may be because women were less likely to receive any form of testing as inpatients; however, only five patients received catheterization after positive index tests (and three of those patients were women). Given most of these catheterizations were the only test performed, the difference in catheterization could also be attributed to patient preference. It is also interesting that there was no bias with CCTA in the ED, and prior research has shown there is no sex difference in further testing within 30 days after this test.⁴⁷

LIMITATIONS

We recognize that the ED chest pain patient population does not fully encompass all patients at risk for heart disease, and so our study is not generalizable to all patients who may be in need of further cardiovascular testing. The urban chest pain population of our ED, furthermore, is predominantly African American, and so our results may be less relevant to other racial groups. In an effort to make our study more generalizable, we designed our inclusion criteria to be broad.

The combination of high- and low-risk patients included in this study may have affected patient perceptions of the discussions with their physicians. Prior encounters with health care professionals regarding cardiovascular disease may have affected patient preferences for testing and subjective ratings of physician performance. Additionally, patients interact with many different health care providers over the course of their care who may have indirectly affected these responses. We attempted to decrease this potential bias by asking patients, when answering survey questions, to focus on the one physician they felt was making the decisions for their care.

We did not survey physicians regarding their recommendations. While this would have helped to address possible recall and miscommunication bias, we were more interested in the patient’s report of these recommendations, as it is what ultimately shapes their preferences. Our study was also not powered to address the effect of these preferences on patient outcomes.

CONCLUSIONS

Our study found that women were less likely than men to be recommended for further cardiac workup for their chest pain, suggesting that sex differences in preference for cardiovascular testing may be partially explained by conversations between women and their doctors. With the growing emphasis on patient-centered care and shared decision-making, the effect of informal recommendations and patient education is playing an increasingly important role in clinical decision-making. More research is needed to elucidate the effect of clinical discussions on patient outcomes, and to further understand how to approach these discussions in an effort to eliminate unintended sex discrepancies in rates of cardiovascular testing.