Author + information
- Received July 24, 2015
- Revision received October 23, 2015
- Accepted October 27, 2015
- Published online February 9, 2016.
- Reshma Jagsi, MD, DPhila,∗ (, )
- Cathie Biga, RNb,
- Athena Poppas, MDc,
- George P. Rodgers, MDd,
- Mary N. Walsh, MDe,
- Patrick J. White, MPHf,
- Colleen McKendry, MStatg,
- Joseph Sasson, PhDf,
- Phillip J. Schulte, PhDg and
- Pamela S. Douglas, MDg
- aDepartment of Radiation Oncology, University of Michigan, Ann Arbor, Michigan
- bCardiovascular Management of Illinois, Woodridge, Illinois
- cRhode Island Hospital/Brown University, Providence, Rhode Island
- dSeton Heart Institute, University of Texas at Austin Dell Medical School, Austin, Texas
- eSt. Vincent Heart Center of Indiana, Indianapolis, Indiana
- fMedAxiom, Neptune Beach, Florida
- gDuke Clinical Research Institute, Durham, North Carolina
- ↵∗Reprint requests and correspondence:
Dr. Reshma Jagsi, Department of Radiation Oncology, University of Michigan, UHB2C490, SPC 5010, 1500 East Medical Center Drive, Ann Arbor, Michigan 48109-5010.
Background Much remains unknown about experiences, including working activities and pay, of women in cardiology, which is a predominantly male specialty.
Objectives The goal of this study was to describe the working activities and pay of female cardiologists compared with their male colleagues and to determine whether sex differences in compensation exist after accounting for differences in work activities and other characteristics.
Methods The personal, job, and practice characteristics of a national sample of practicing cardiologists were described according to sex. We applied the Peters-Belson technique and multivariate regression analysis to evaluate whether gender differences in compensation existed after accounting for differences in other measured characteristics. The study used 2013 data reported by practice administrators to MedAxiom, a subscription-based service provider to cardiology practices. Data regarding cardiologists from 161 U.S. practices were included, and the study sample included 2,679 subjects (229 women and 2,450 men).
Results Women were more likely to be specialized in general/noninvasive cardiology (53.1% vs. 28.2%), and a lower proportion (11.4% vs. 39.3%) reported an interventional subspecialty compared with men. Job characteristics that differed according to sex included the proportion working full-time (79.9% vs. 90.9%; p < 0.001), the mean number of half-days worked (387 vs. 406 days; p = 0.001), and mean work relative value units generated (7,404 vs. 9,497; p < 0.001) for women and men, respectively. Peters-Belson analysis revealed that based on measured job and productivity characteristics, the women in this sample would have been expected to have a mean salary that was $31,749 (95% confidence interval: $16,303 to $48,028) higher than that actually observed. Multivariate analysis confirmed the direction and magnitude of the independent association between sex and salary.
Conclusions Men and women practicing cardiology in this national sample had different job activities and salaries. Substantial sex-based salary differences existed even after adjusting for measures of personal, job, and practice characteristics.
Women have constituted nearly one-half of the medical student body in the United States for more than a decade. Nevertheless, women continue to be dramatically underrepresented in the specialty of cardiology, accounting for only 21% of first-year cardiology fellows training in 2012 to 2013 (1). This situation has fostered concerns in the United States (2) and abroad (3–6) that cardiology may no longer be accessing the full pool of talent in the educational pipeline, potentially jeopardizing the ability of the field to continue generating the highest quality clinical care, teaching, and research.
The lack of sex diversity in the cardiology workforce is striking, with recent workforce estimates suggesting that women constitute only approximately 12% of general cardiologists and even smaller proportions of specialties, such as interventional cardiology and clinical cardiac electrophysiology (7). Both the American College of Cardiology (ACC) and the British Cardiac Society have convened working groups charged with addressing women’s persistent underrepresentation in the field of cardiology (4,8). These groups have identified several consistent concerns, including the challenges of work–life balance in a field in which on-call duties can be frequent and demanding, and occupational radiation exposure, sex bias, and overt discrimination exist. They have called for increased mentorship and exposure to female role models, along with efforts to address the “image issues for the prior traditionally male domain of cardiology” (2).
Despite these and other laudable efforts to increase the representation of women in the specialty, much remains unknown about the experiences of those women who have slowly begun to join this previously male-dominated specialty. Professional life surveys conducted in 1996 and 2006 by the ACC have suggested that women in cardiology are less likely to be married or have children than their male colleagues (9,10), less likely to practice interventional cardiology, less likely to perceive career advancement and/or salaries to be higher than their peers, and less satisfied with their level of financial compensation. Although illuminating, these surveys collected little information about the detailed work activities or actual pay of practicing cardiologists.
Thus, little is currently known about the distribution of working activities or pay of either male or female practicing cardiologists. To provide this information, we used a large national dataset to evaluate job descriptions and to compare compensation in men and women after controlling for differences in work activities.
An original dataset was obtained from MedAxiom, a membership network and service provider for cardiology practices, hospitals, and academic centers. This dataset, which represented the entire calendar year of 2013, contained data voluntarily reported by member executives regarding the personal, job, and practice characteristics of 3,187 cardiologists from 161 practices. Subjects were excluded if they were missing values for either sex (n = 37) or salary (n = 287) because those were the 2 primary variables of interest in this analysis. In addition, subjects were excluded if they worked <40 half-days (n = 12), reported <500 work relative value units (wRVUs), or had >25,000 wRVUs (n = 172). In the United States, wRVUs are allocated based on billing claims codes submitted for reimbursement and are therefore a productivity measure linked to clinical reimbursement. The final analytic sample included data for 2,679 cardiologists. A diagram depicting these exclusions can be found in Figure 1.
Key measures included personal, job, and practice characteristics reported in the surveys. Specifically, age was categorized into 5 groups (24 to 38, 39 to 48, 49 to 58, 59 to 68, and ≥69 years). Race/ethnicity was grouped into 8 categories as listed in Table 1. Subspecialty was grouped as electrophysiology, general/noninvasive, interventional, invasive, or other.
Job characteristics included whether the subject was working full-time. Two definitions of full-time work were included: “self-reported,” which corresponded to the designation reported by the administrator completing the survey, and “investigator-defined,” which corresponded to a response of working >400 half-days in the last year. We also considered the number of half-days worked and whether the respondent participated in on-call duties (full call, partial call, or no call). The number of wRVUs and new patient office visits were also measured. In addition, we evaluated whether the respondent’s job included certain activities: office consultations, return office visits, hospital consultations, initial hospital care, hospital observation cases, hospital visits, outpatient pacemaker checks, permanent pacemaker implantations, catheterizations, angioplasty, echocardiograms, electrocardiograms, positron emission tomography (PET) scans, single-photon emission computed tomography (SPECT) scans, stress echocardiograms, and treadmill tests. For these items, we first evaluated whether the subject performed any, 1 to 4, or ≥5 procedures in each category in the last year and then, among those performing ≥5 procedures, the absolute number of such procedures.
Information was collected on practice characteristics, including geographic region (grouped as Midwest, Northeast, South, and West), practice composition (cardiology only, cardiology and cardiothoracic surgery, cardiology and multiple surgeries, or cardiology and vascular surgery), whether the practice had a female administrative director, practice size, the practice compensation model (blended, equal share, productivity, or salary plus bonus), whether the practice owned imaging facilities (none, SPECT only, or SPECT and PET), provider-based billing details (diagnostic only, diagnostic plus evaluation and management services, evaluation and management only, or no provider-based billing), and whether the practice owned various clinical laboratories (catheterization, computed tomography [CT] scans, magnetic resonance, or sleep laboratories).
Finally, information was collected on salary for each physician in the practice, as reported to MedAxiom by the financial executives at each practice. MedAxiom asked the practice to report the total W2 compensation for each physician, and it further clarified the request by specifying that the practice not include pension or benefits.
The personal, job, and practice characteristics of the analytic sample were first described according to sex. The procedure volumes were then compared after restricting the sample to those who performed each procedure at least 5 times in the previous year. Continuous variables were compared by using a Wilcoxon rank-sum test; categorical variables were compared by using a Pearson chi-square or Fisher exact test. Salary according to sex was then described and regression models developed to evaluate the association between sex and salary, independent of the other measured characteristics. We first evaluated the unadjusted difference in compensation according to sex, the association between compensation and wRVUs within each sex separately, and the difference in compensation according to sex after adjustment for wRVUs alone. To adjust for additional potential confounders and mediators of the sex–salary relationship, a full model was then constructed that included the following independent variables: age, sex, subspecialty, full-time status (both as reported in surveys and as defined by the investigators), half-days worked, days off, new patient office visits, wRVUs, office consultations, return office visits, hospital consultations, initial hospital care, hospital observation care, hospital visits, outpatient pacemaker checks, permanent pacemaker implantations, catheterizations, echocardiograms, electrocardiograms, PET scans, SPECT scans, stress echocardiograms, treadmill tests, practice geographical area, practice compensation model, and practice ownership model. For the continuous variables representing procedure volumes, the continuous variable was included as well as indicator variables capturing any use of procedure and separately at least 5 reported procedures. The adjusted association between sex and salary from this full model was reported, and a sensitivity analysis was then conducted in the subset of cases for whom call schedule was known; the goal was to evaluate the impact of including this variable in addition to all others in the full model on the magnitude of any association between salary and sex. In addition, a best regression model for salary was developed that was fit via stepwise selection, minimizing the Schwarz-derived Bayesian information criterion to describe a parsimonious multivariable model of all characteristics associated with compensation (procedure volume variables were used exclusively rather than indicator variables).
For further evaluation of salary differences according to sex, the Peters-Belson technique was also applied. This approach is the preferred statistical method to evaluate compensation and has been described in detail elsewhere (11,12). In brief, it involves the development of a linear regression model of salary in the subgroup of male subjects alone, followed by the application of that model to predict expected salaries in the female subjects (given their specific personal, job, and practice characteristics). This approach allows for an evaluation of salary observed versus expected, had the women in the sample been men with all other measured characteristics held constant. Confidence intervals (CIs) for the difference in predicted versus observed were obtained by using the percentile method of nonparametric bootstrap with 1,000 bootstrap replications.
Statistical analyses were conducted by using SAS version 9.4 (SAS Institute, Inc., Cary, North Carolina). Values of p < 0.05 were considered significant in all analyses.
Overall, 229 subjects in the analytic sample were female, and the remaining 2,450 were male. Table 1 describes the personal, job, and practice characteristics of the analytic sample grouped according to sex. The majority of the sample was white, and nearly one-half practiced in the South.
Comparison of personal characteristics according to sex reveal that in this sample, the distributions of age, race, and subspecialty differed among women and men (p < 0.001, p = 0.02, and p < 0.001, respectively). Women tended to be younger, with 56% of women being <49 years of age compared with 38.8% of men, and less likely to be white (67.3% vs. 73.2%). Women were also far more likely to have specialized in general/noninvasive cardiology (53.1% vs. 28.2%), whereas a lower proportion (11.4% vs. 39.3%) reported an interventional subspecialty compared with men.
Comparison of job and practice characteristics according to sex
Numerous job characteristics also differed according to sex, including the proportion of subjects who were working full-time (79.9% of women vs. 90.9% of men based on survey report, and 62.4% vs. 72.7% based on the investigator definition of at least 400 half-days worked; both p < 0.001). Women worked a mean of 387 half-days and men a mean of 406 half-days (p = 0.001). Call responsibilities also varied significantly according to sex (p < 0.001); although most subjects had full on-call responsibilities (82.5% overall), higher proportions of women compared with men took no call (17.5% vs. 7.2%) or partial call (15.0% vs. 8.9%) rather than full call (67.5% vs. 83.9%; p < 0.001). Women saw fewer new patient office visits (mean 167 vs. 180; p = 0.04) and generated fewer wRVUs (mean 7,404 vs. 9,497; p < 0.001).
Women were less likely to perform many of the activities we evaluated and were particularly less likely to perform procedures such as pacemaker implantations (12.7% vs. 22.9% performing ≥5 procedures; p < 0.001) and catheterizations (34.5% vs. 60.1% performing ≥5 procedures; p < 0.001). Women were more likely to perform SPECT scans (66.8% vs. 59.3% performing ≥5 procedures; p = 0.04) and stress echocardiograms (71.2% vs. 61.2% performing ≥5 procedures; p = 0.01), and no less likely to perform echocardiograms generally (90.4% vs. 87.4% performing ≥5 procedures; p = 0.26) or treadmill testing (88.2% vs. 87.3% performing ≥5 procedures; p = 0.78).
Most practice characteristics did not differ significantly according to sex. However, practice composition was significantly different (p = 0.01), with women slightly less likely to be part of groups that included cardiology only (70.0% vs. 74.8%).
Table 2 displays the mean values of services and demonstrates that, for most of the services evaluated, procedure volumes were similar for men and women who had performed at least 5 of those services in the preceding year. However, per-physician annual volumes differed for certain activities, including most dramatically the number of hospital visits (mean 566 for women performing ≥5 procedures vs. 676 for men; p < 0.001), outpatient pacemaker checks (mean 274 among women performing ≥5 procedures vs. 403 for men; p < 0.001), and catheterizations (mean 133 performed by women who performed ≥5 procedures vs. 217 by men; p < 0.001).
Comparison of compensation according to sex
Men generally had higher salaries than women. Mean salary was $400,162 ± $192,124 (median $394,586; interquartile range $256,064 to $518,277) among the women and $510,996 ± $216,337 (median $502,251; interquartile range $381,417 to $621,306) among the men. Figure 2 provides waterfall plots of unadjusted compensation according to sex in this sample. Table 3 presents compensation according to sex in several meaningful subgroups.
Figure 3 displays the plots of compensation versus wRVUs for men and for women. The slope of the association between compensation and wRVUs among women was $34.36 per wRVU (95% CI: 28.64 to 40.08); the slope among men was $32.47 per wRVU (95% CI: 30.88 to 34.06). There was no significant difference between slopes (p = 0.53), suggesting that the relationship between wRVUs and salary was similar for both sexes. However, a significant difference in the intercepts (p = 0.03) suggests a difference in compensation by sex that was not fully explained by wRVUs. The fit with a wRVU-only model was modest: R2 were 0.39 for men and 0.45 for women.
The Central Illustration depicts the difference according to sex in compensation before any adjustment, after adjustment for clinical productivity (as measured solely by wRVUs), and after adjustment for multiple measured characteristics in the dataset. As shown, although adjustment for >40 personal, job, practice, and productivity characteristics as described in the Methods substantially improved the model fit (R2 rose to 0.59), sex remained independently associated with salary in this sample (p = 0.001), with an effect size of $37,717 (95% CI: 15,056 to 60,378) for male versus female physicians in the fully controlled model (R2 = 0.59); the full model is presented in Online Table 1. In a sensitivity analysis of the subset of 1,301 respondents whose call schedule was known in addition to the other characteristics included in the full model, sex was again found to be independently associated with salary (p = 0.006), with a similar effect size ($43,217 higher predicted pay for a man vs. a woman with all other characteristics held constant).
The best parsimonious model of salary in this sample (R2 = 0.55) revealed associations between salary and sex along with 6 job characteristics (subspecialty, full-time status, wRVUs, hospital observation care, hospital initial care, and stress echocardiograms) and 2 practice characteristics (geographical area and ownership model) (Table 4). The sex difference observed in the parsimonious model was again similarly sized to that observed in the full model ($38,485) and significant (p < 0.001).
Peters-Belson analysis revealed that, on the basis of productivity and other characteristics, the women in this sample would be expected to have a mean salary of $432,631 if they had been male (model R2 = 0.56). The actual observed mean salary among women was $400,882. Thus, there was a $31,749 difference in salary (95% CI: 16,303 to 48,028) according to sex that was not explained by the detailed measures of personal, job, and practice characteristics evaluated in this study.
This analysis of practicing cardiologists’ job activities and compensation provides novel information about the ways cardiologists in general, and men and women in particular, allocate their professional effort. Our observation of a significant sex difference in compensation, even after adjusting for differences in multiple measures of productivity and job activity in this sample of practicing cardiologists, is also—to our knowledge—unique. The present study suggests that although some of the overall variation in salary between male and female cardiologists can be explained by differences in work performed, a substantial difference remains even after detailed measurement of procedural volumes, on-call coverage, and other work activities. Furthermore, the magnitude of this difference is remarkably constant when analyzed in various ways.
Although sex differences in physician pay have been documented in numerous previous studies (13–20), questions have remained about the extent to which they relate to differences in specialty, productivity, and other job characteristics, especially for physicians practicing outside of academic medical centers (21,22). Moreover, there continues to be hope that differences documented in older studies might have dissipated over time. To our knowledge, the only data regarding salaries in the field of cardiology have been limited by lack of adjustment for meaningful differences in productivity and other job characteristics (23).
The present study’s findings regarding salary are interesting when considered in the context of the 2006 ACC survey (10). In that survey, a bigger difference in dissatisfaction with financial compensation was observed between men and women in academic positions, with 29% of women working in such settings expressing moderate or severe dissatisfaction, compared with 18% of men. By contrast, similar proportions of cardiologists of either sex (17% of women and 18% of men) working in other settings, presumably largely in community practice, expressed dissatisfaction with financial compensation. Previous studies have clearly established sex differences in salary within homogeneous groups of academic physicians, even after accounting for numerous measures of academic and clinical productivity and other characteristics (21,22). The present study suggests that these unexplained differences are not restricted to academic settings but also exist in the compensation of a single specialty of physicians practicing in the community.
Our findings regarding working patterns also offer an interesting complement to the data previously reported in the ACC surveys (10). The ACC surveys documented a decrease over time in the proportion of both men and women who described themselves as working full-time, from 86% and 95% of women and men, respectively, in 1996, to 80% and 82% in 2006. In the present sample, men were more likely to report working full-time than women (91% vs. 80%), although both men and women were likely to overestimate working full-time when this was recalculated as at least 400 half-day sessions per year. Such differences may relate to differences in the sampling frame of the 2 studies, and further research is necessary to obtain a more comprehensive understanding of the ways in which cardiologists may deviate from full-time practice, which cardiologists do so, and how satisfied they may be.
Other findings of this study, such as women’s lower likelihood of taking full call, are consistent with previous studies. Numerous studies have documented our society’s recalcitrance toward changing the traditional norms of a gendered division of labor in which women continue to be more likely to shoulder the greater burden of domestic responsibilities, as well as its impact on women’s experiences in the medical profession (24–28). Even among highly career-oriented young physicians, studies have reported large differences in the hours per week devoted to parenting and domestic activities by women versus men (29). In that context, it is hardly surprising that women are less likely to participate in full call. An important contribution of this study, however, is to document that such arrangements do exist in a number of cardiology practices, indicating that young women should not be unduly pessimistic about the possibility of combining a career in cardiology with other desired personal roles. Others have previously noted the importance of disseminating this information to medical students and internal medicine residents (8).
It is important to note that sex differences in participation in traditional full-time, full-call roles may magnify the possible impact of sex differences in negotiation behaviors on compensation. Contracts for nontraditional positions, including part-time jobs or ones that involve lesser participation in on-call duties, are likely to be less standardized than contracts for more traditional positions. The small percentage of compensation models (approximately 25%) that are solely and strictly based on productivity or on an equal share among all physicians in a practice would be unlikely to reveal differences in sex; however, the majority include other components and may be more likely to result in some nonobjective variability in pay. This may amplify documented differences in negotiation behaviors between men and women (30–32) in ways that could explain some of the differences observed in the present study. Although we did control the models for full-time status and performed a sensitivity analysis that included on-call duties, if women in non–full-time and/or non–full-call positions required nonstandard contracts and also negotiated less aggressively than men pursuing similar positions, that could explain some of the residual salary difference observed. Unfortunately, women’s failure to negotiate as aggressively as men is, to some extent, related to observations that women who behave in ways discordant to sex stereotypes may be perceived differently and experience other adverse consequences.
Women’s lesser likelihood of specializing in interventional cardiology is a well-documented phenomenon (7,10). To the extent that women are discouraged from pursuing interventional careers due to sex bias, discrimination, or unfounded concerns about occupational radiation exposure, this topic merits ongoing attention. To the extent that workforce projections forecast a shortage of general cardiologists, and thought leaders seek to encourage trainees to pursue that path (33), the greater tendency for women to pursue careers in general or noninvasive cardiology suggests that efforts to recruit women to cardiology training are even more important.
The observation that women generated fewer wRVUs in this study must be taken in the context of the observations regarding differences in specialization and not simply a difference in productivity. The tendency of the current U.S. system of clinical reimbursement to reward intervention over cognitive activity can magnify sex inequity in a society in which young men are socialized to exhibit “agentic” behaviors and young women are expected to demonstrate “communal” behaviors. This tendency may in turn lead to some of the differences in specialization and reward observed. It is particularly striking that sex differences in pay exist even after accounting for differences in specialization and wRVU assessment.
This study has numerous strengths, including access to data on detailed work activities, productivity metrics, and pay of practicing cardiologists, that have not been available through other sources. Nevertheless, it also has limitations. First, data in this study were reported by practice administrators; although these individuals are likely to have accurate information about productivity and pay metrics, the potential for misreporting exists. However, we believe that systematic misreporting according to sex is unlikely, minimizing the impact of this factor on the main study findings. Second, this sample was limited to MedAxiom member organizations that responded to the 2014 annual survey, which covers all data from calendar year 2013; the generalizability of these findings to practices dissimilar to those who reported data (particularly academic or smaller practices that may not utilize its services) may be limited. Most of the sample came from the Midwest and South, which could also affect the generalizability of the findings across the United States, although controlling for practice location did not eliminate the sex difference observed. We had access to a large number of productivity metrics, but there may be relevant unmeasured factors that could have explained and possibly justified the sex differences in pay observed. In particular, information on call responsibilities was not available for many individuals, although we observed no modifying effect of including that variable within the subset of respondents for whom these data were available. Other potential sources of compensation such as medical directorships, which may vary meaningfully according to sex, were not captured and could account for some of the unexplained differences in overall compensation observed. However, if the process of determining medical directorship or other leadership positions itself reflects sex-biased processes of evaluation, this would not necessarily be a justifiable source of differences in pay. Similarly, information regarding differences in access to support (such as mid-level providers) was not available to us, but if not uniform may affect both productivity measures and income. To protect the anonymity of providers and individuals, we did not have practice identifiers that would allow us to control for correlation of individuals from the same practice. If certain practices that provided the highest levels of physician compensation were predominantly male (and the ability to provide this compensation was not itself due to productivity measures already controlled for in the analysis), this could account for some of the differences observed, rather than inconsistencies in remuneration by sex within a given practice. Finally, we did not have information on other aspects of compensation such as retirement contributions, insurance, or practice equity from partnership status that may differ by sex, nor did we have access to information on differences in patient insurance mix.
This study provides valuable information about the diverse jobs held by men and women practicing cardiology in the United States today. In a landmark working group report, Warnes et al. (8) state, “It is imperative that female medical students and internal medicine residents become better informed about the broad spectrum of career opportunities in cardiology, several of which are compatible with a desire to achieve better work-life balance.” The findings of the present study suggest that there is indeed a broad range of working patterns and job characteristics available to men and women practicing cardiology in the community. These findings may therefore be particularly useful to cite when seeking to recruit promising young men and women to the discipline. The findings should also motivate efforts to ensure equity in pay, including the promotion of greater transparency and consistency in salary structures at the practice level. One of the first steps in creating meaningful change is to recognize that sex equity is not solely a “women’s issue” but rather an issue that affects the entire profession. Professional societies and cardiovascular practices must carefully examine their cultures as well as objective practices, with special attention to the insights provided in the growing literature on how best to mitigate barriers to the careers of women physicians.
COMPETENCY IN PROFESSIONALISM: Men and women practicing cardiology in the United States have diverse job activities. A broad spectrum of career opportunities within the specialty may be attractive to young men and women alike.
TRANSLATIONAL OUTLOOK: Future research should focus on the developing interventions that foster greater equality in compensation for both male and female cardiologists.
Mr. White is a shareholder in MedAxiom and serves as its president. Dr. Sasson is employed by MedAxiom. Ms. Biga is a partner in Cardiovascular Management of Illinois. Dr. Poppas is a stockholder in GE. Ms. McKendry was supported by a National Institutes of Health grant (T32 HL079896). All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- American College of Cardiology
- confidence interval
- positron emission tomography
- single-photon emission computed tomography
- work relative value units
- Received July 24, 2015.
- Revision received October 23, 2015.
- Accepted October 27, 2015.
- American College of Cardiology Foundation
- ↵American Board of Internal Medicine (ABIM). Percentage of first-year fellows by gender and type of medical school attended. Available at: http://www.abim.org/about/examInfo/data-fellow/chart-05.aspx. Accessed December 15, 2015.
- Wenger N.K.
- Andreotti F.,
- Crea F.
- Timmis A.D.,
- Baker C.,
- Banerjee S.,
- et al.
- Timmis A.D.,
- English K.M.
- Rodgers G.P.,
- Conti J.B.,
- Feinstein J.A.,
- et al.
- Warnes C.A.,
- Fedson S.E.,
- Foster E.,
- et al.
- Limacher M.C.,
- Zaher C.A.,
- Walsh M.N.,
- et al.
- Poppas A.,
- Cummings J.,
- Dorbala S.,
- Douglas P.S.,
- Foster E.,
- Limacher M.C.
- Weeks W.B.,
- Wallace T.A.,
- Wallace A.E.
- ↵Medscape. Medscape Physician Compensation Report 2015. Available at: http://www.medscape.com/features/slideshow/compensation/2015/cardiology?src=wnl_edit_specol&uac=66019PZ&impid=683796. Accessed April 21, 2015.
- Fox M.F.
- Babcock L.,
- Laschever S.
- Fye W.B.