Work Environment, Burnout, and Intent to Leave Current Job Among Cardiologists and Cardiology Health Care Workers: Results From the National Coping With COVID Survey

Abstract

Background

Little is known about factors contributing to burnout and intent to leave in cardiologists and other cardiology health care workers.

Methods and Results

The Coping With COVID survey assessed work conditions, burnout, and intent to leave among physicians, nurses, advanced practice providers, and other clinical staff (OCS) from April 2020 to December 2020. Single‐item measures assessed work conditions, burnout (emotional exhaustion), and intent to leave. Multilevel logistic regression examined work life variables' relationships to burnout among role types and feeling valued as a mediator. Open‐ended comments analyzed via grounded theory contributed to a conceptual model. Coping With COVID was completed by 1199 US cardiology health care workers (354 physician/520 nurses/198 advanced practice providers/127 OCS). Nurses were most likely to report burnout (59% nurses, 57% OCS, 46% advanced practice providers, 40% physicians, P<0.0001). Workload correlated with burnout in all groups (adjusted odds ratios [aORs], 4.1–17.4; Ps<0.005), whereas anxiety/depression related to burnout in all except OCS (aORs, 3.9–8.3; Ps≤0.001). Feeling valued was related to lower burnout in most groups. Intent to leave was common (23%–45%) and was lower in physicians and advanced practice providers who felt valued (aORs, 0.26 and 0.22, respectively; Ps<0.05). Burnout was highest for nurses in practice 16 to 20 years, and intent to leave was highest for OCS in practice 16 to 20 years. Themes contributing to burnout included personal and patient safety, leadership, and financial issues.

Conclusions

Burnout was prevalent among cardiology health care workers and highest in nurses and OCS. Addressing factors associated with burnout in different role types may improve work life sustainability for all cardiology health care workers.

Nonstandard Abbreviations and Acronyms

APP

advanced practice provider

CwC

Coping With COVID

HCW

health care worker

ITL

intent to leave

OCS

other clinical staff

Clinical Perspective.

What Is New?

• In 1199 US cardiology health care workers, 354 physician/520 nurses/198 advanced practice providers, and 127 other clinical staff, nurses were the most likely group to report burnout as defined by a measure most closely correlated with emotional exhaustion (59% nurses, 57% other clinical staff, 46% advanced practice providers, 40% physicians; P<0.0001).

• Workload correlated with burnout in all groups, whereas feeling valued was related to lower burnout in most groups. Intent to leave was common (23%–45%) and lower in physicians and advanced practice providers who felt valued (adjusted odds ratios, 0.26 and 0.22, respectively, Ps<0.05).

What Are the Clinical Implications?

• Addressing factors associated with burnout in different role types may improve work life sustainability for all cardiology health care workers and impact burn out and retention.

Although a recent survey released by the American College of Cardiology revealed that rates of burnout among cardiology physicians had doubled since 2015, ¹ there is less information about how work life in cardiology and work conditions during the pandemic have affected other team members, including nurses, advanced practicing providers (APPs), and other clinical staff (OCS). Furthermore, there is little knowledge of work conditions associated with burnout and intent to leave (ITL) a current job in cardiology, including the impact of variables such as work overload, childcare stress, anxiety or depressive symptoms due to COVID, sense of mission and purpose, and feeling valued by one's organization. Most current studies ² have focused solely on physicians, but cardiology, like many other disciplines in health care, is a team effort and requires the well‐being of all health care team members. Finally, with recent studies emphasizing the need for work life balance and flexibility in cardiology, ³ medical residents' perceptions of cardiology as a challenging career, ⁴ stress among cardiology fellowship program directors, ⁵ and high distress among allied health personnel and nurses in cardiovascular health care workers (HCWs) in Canada, ⁶ a broader understanding of work life components influencing cardiology team members could be helpful in terms of improving work environments, and increasing recruitment and retention.

We reviewed data from the national Coping With COVID (CwC) survey to determine work life factors related to burnout, feeling valued, and ITL among cardiology physicians (cardiologists), nurses, APPs, and OCS. Our study objectives related to (1) how burnout, ITL, and contributing factors vary among physicians, nurses, APPs, and OCS and (2) if burnout is mediated by the same variables in each role type. We also focused attention on a key variable, a sense of feeling valued by one's organization, and assessed how important this was to different cardiology HCWs.

Methods

Sample

The American Medical Association CwC survey canvassed >58 000 HCWs, spanning >100 US health care organizations, with the current sample obtained between April 2020 and December 2020. Those sampled included 1199 cardiology HCWs (physicians, nurses, APPs, and OCS). Other clinical staff comprised medical assistants, radiology technicians, physical therapists, laboratory technicians, pharmacists, and social workers (Table 1). Cardiologists are defined as physicians who specialize in the cardiovascular system. Data and study materials will not be available for other researchers.

Table 1.

Demographics in Cardiology Health Care Workers in the Coping With COVID Survey (N=1199, Organizations=111)

Demographic	Physician		APP		Nurse		OCS
	N=354		N=198		N=520		N=127
	%=29.5		%=16.5		%=43.3		%=10.5
	N	%	N	%	N	%	N	%
Race and ethnicity
Missing	31	8.76	18	9.09	32	6.15	2	1.57
White	194	54.8	146	73.74	373	71.73	92	72.44
Hispanic or Latino	7	1.98	3	1.52	10	1.92	12	9.45
Black	1	0.28	2	1.01	12	2.31	4	3.15
American Indian	1	0.28	1	0.51	1	0.19	2	1.57
Asian/Pacific Islander	69	19.49	13	6.57	33	6.35	1	0.79
Prefer not to answer	46	12.99	13	6.57	57	10.96	14	11.02
Other (please specify)	5	1.41	2	1.01	2	0.38	2	1.57
Sex
Missing	0	0	0	0	0	0	1	0.79
Prefer not to answer	18	5.08	8	4.04	41	7.88	8	6.3
Men	279	78.81	28	14.14	45	8.65	29	22.83
Women	57	16.1	162	81.82	433	83.27	89	70.08
Nonbinary/third gender	0	0	0	0	1	0.19	0	0
Time in practice
Missing	3	0.85	1	0.51	8	1.54	5	3.94
1–5 y	44	12.43	41	20.71	116	22.31	39	30.71
6–10 y	50	14.12	40	20.2	90	17.31	24	18.9
11–15 y	52	14.69	34	17.17	63	12.12	16	12.6
16–20 y	55	15.54	27	13.64	53	10.19	12	9.45
>20 y	150	42.37	55	27.78	190	36.54	31	24.41
Setting
Inpatient	252	71.2	123	62.1	413	79.4	84	66.0
Outpatient	102	28.8	75	37.8	107	20.5	43	34.0

APP indicates advanced practice provider; and OCS, other clinical staff.

Study Design

Surveys were administered locally, electronically, and anonymously, with data sent to the American Medical Association‐associated data laboratory in Madison, Wisconsin (Forward Health Group). Due to local administration of surveys, an overall median response rate for all respondents could be calculated, but individual specialty and role response rates could not.

Theory

Items for inclusion in the survey were based on a conceptual model linking work conditions and clinician reactions such as stress, burnout, and ITL from the MEMO (Minimizing Error Maximizing Outcome) study, ⁷ a model that has since been applied to other HCWs. In brief, this model links challenging work conditions (such as workload [work control], chaotic environments [pace of work], time pressure, and a supportive or nonsupportive organizational culture to stress, burnout, satisfaction, and ITL). The model was initially drawn from the demand‐control theory of job stress by Karasek et al, ⁸ who demonstrated that work demands are balanced by work control, and that the lack of such balance leads to not only higher stress but also adverse cardiovascular outcomes in a random sample of Swedish male workers, and confirmed in a study involving both male and female Swedish workers by Johnson and Hall. ⁹ Using both quantitative and qualitative data analyzed in the current study, we have revised the demand‐control theory‐based MEMO model to address relationships between work conditions and HCW reactions in cardiology.

Measure

The national CwC survey was administered during the early phase of the pandemic and was thus rolled out quickly. Approximately a dozen items were included, such as previously identified correlates of burnout in other specialties (eg, work overload, fear of exposure, anxiety and depression due to COVID care, and childcare stress) and likely mitigators (feeling valued, and sense of mission and purpose). Most items were scored on 4‐point Likert scales, with the top 2 choices representing presence of an item (eg, work overload) and the other 2 responses being absence (see Data S1 for relevant survey items). Outcomes included a single item burnout measure validated against the emotional exhaustion subscale of the Maslach Burnout Inventory, ¹⁰ and a standard question on ITL within 2 years (top 3 choices [moderate, likely, or definite] = ITL). ¹¹ The burnout and ITL items were added later in 2020 and thus had reduced responses within different organizations.

The Hennepin Healthcare Institutional Review Board deemed the CwC project exempt from human subjects review, and therefore consent was not required.

Statistical Analysis

Quantitative Analyses

Descriptive data portrayed percent burned out (options 3, 4, and 5 on a 5‐choice item, with symptoms of burnout [3], severe burnout [4], and extreme burnout [5] categorized as burned out) and likelihood of ITL practice (moderate, likely, or definite categorized as ITL) in each of the 4 role groups (physicians, nurses, APPs, and OCS), along with prevalence of potentially associated work life variables (workload, childcare stress, mission/purpose, fear of exposure/safety, and perceived anxiety/depression due to their work). We also analyzed the strength of relationships between feeling valued and burnout in the 4 role groups. Work life factors, such as feeling valued and work overload (full list specified above), were evaluated as correlates of burnout and ITL via multiple multilevel, multivariable logistic regressions that determined adjusted odds ratios (aORs), adjusted relative risks, and absolute risk differences. Clusters were the organizations with intercepts as random effects. We computed the adjusted relative risks and absolute risk differences in Stata SE version 18.0 and computed adjusted relative risks and absolute risk differences (via the Stata Journal) based on our logit models. The significance level used for hypothesis testing was P<0.05. Models for ITL were unstable due to smaller samples of those responding (eg, 13 other staff with ITL) but did converge when work life variables alone were included. To examine relationships of work life factors with burnout and ITL in the 4 role groups through the lens of a potential mediator, feeling valued, ¹² we used a structural probit model to assess the relationship between binary response variables and ≥1 explanatory variable. In such a model, the binary response variable is assumed to follow a latent variable model, where the observed binary outcome is determined by whether the underlying latent variable exceeds a certain threshold. The latent variable is modeled as a linear combination of the explanatory variables and an error term, which is assumed to be normally distributed with mean 0 and constant variance. ¹³ Because structural models included a mediator variable, the associations between predictor variables and burnout are not precisely the same as in the regressions. We then looked at how burnout, ITL, and associated work life variables varied in each of role group by years in practice. Together, these findings allowed us to devise a theory of wellness and retention in cardiology, based upon the original MEMO model and recent theoretical work in nursing stress. ¹⁴

Qualitative Analyses

Open‐ended comments added context and texture by illuminating team member perceptions of workplaces and work life. Using a grounded theory approach ¹⁵ to discover theory from the qualitative data by looking for emerging patterns, comments were reviewed and stratified by role group (physician/nurse/APP/OCS). Two reviewers (S.M. and E.S.) identified themes and coded data independently. ¹⁶ Coding disagreements were resolved in conference with other coauthors. The writing team met in a large group to confirm findings, updating the developing model of work life in cardiology. Strategies such as training and creating a codebook were used to ensure interrater reliability.

Results

Sampling and Response Rate Calculations

Data were collected between April 2020 and December 2020, with surveys completed by 1199 cardiology HCWs from 111 institutions in 23 states (54% Western, 24% Northwest, 13% South, and 9% Midwest). Although response rates among cardiology HCWs were not available, median response rate among the 111 organizations with cardiology HCWs was 34% for all respondents.

Sample Constituents and Prevalence of Burnout and ITL

Cardiology respondents were split into 4 groups (Table 1): 354 physicians, 520 nurses, 198 APPs, and 127 OCS. Nurses were the most likely to report burnout (see Table 2, with 59% burnout in nurses, 57% in OCS, 46% in APPs, and 40% in physicians; P<0.0001 for differences among roles). Nurses were also most likely to report ITL (45% in nurses, 35% in APPs, 32% in OCS, and 23% in physicians; P<0.001).

Table 2.

Prevalence of Burnout, Intent to Leave, and Work Life Variables by 4 Cardiology Role Groups (not Including Missing Data)

Variable	Physicians, % (n)	Nurses, % (n)	Advanced practice providers, % (n)	Other clinical staff, % (n)
Burnout*	40% (129)	59% (286)	46% (83)	57% (71)
Intent to leave	23% (57)	45% (54)	35% (38)	32% (13)
Fear	61% (216)	69% (359)	62% (122)	68% (87)
Anxiety/depression	21% (76)	45% (233)	34% (67)	40% (51)
Work overload	34% (122)	52% (270)	37% (74)	48% (61)
Childcare stress	22% (77)	21% (107)	25% (49)	22% (28)
Mission/purpose	38% (136)	49% (256)	38% (75)	46% (59)
Feeling valued	48% (170)	44% (230)	46% (92)	33% (42)

^*Burnout is defined primarily as emotional exhaustion.

Burnout Correlates in Different Role Types

Work overload was an independent correlate of burnout in all 4 role types (aOR, 6.69 [95% CI, 3.39–13.2]) in physicians; P<0.001; aOR, 7.03 [95% CI, 4.19–11.78] in nurses; P<0.001; aOR, 4.13 [95% CI, 1.71–9.99] in APPs; P=0.002; and aOR, 17.4 [95% CI, 4.49–67.89] in OCS; P<0.001). Anxiety/depression (emotional health) in relation to COVID care was meaningful in relation to burnout in 3 groups (aOR, 8.27 [95% CI, 3.31–20.6] in physicians; P<0.001; aOR, 3.90 [95% CI, 2.27–6.70] in nurses; P<0.001; and aOR, 4.53 [95% CI, 1.84–11.15] in APPs; P=0.001; Table 3 and Table S1). Feeling valued was associated with statistically significant lower burnout in 3 groups: aOR, 0.48 [95% CI, 0.25–0.91] in physicians (or 52% lower odds of burnout if feeling valued); P=0.026; aOR, 0.28 [95% CI, 0.16–0.46] in nurses (72% lower burnout odds); P<0.001; and aOR, 0.29 [95% CI, 0.12–0.65) in APPs (71% lower odds); P=0.003. ITL was substantively lower in physicians who felt valued (aOR, 0.26; P=0.001) and in APPs feeling valued (aOR, 0.22; P=0.02).

Table 3.

Multivariable Regression Analysis Evaluating Work Life Variables to Identify Independent Predictors of Burnout and Intent to Leave Among Cardiology Team Members

Variable	Physicians		Nurses		Advanced practice providers		Other clinical staff
	aOR	P value	aOR	P value	aOR	P value	aOR	P value
Burnout prevalence model
Fear	1.51	0.216	0.93	0.801	1.73	0.208	0.45	0.287
Anxiety	8.27	<0.001	3.90	<0.001	4.53	0.001	3.12	0.132
Workload	6.69	<0.001	7.03	<0.001	4.13	0.002	17.4	<0.001
Childcare stress	0.65	0.357	1.10	0.742	1.33	0.564	0.72	0.668
Purpose	0.54	0.073	0.88	0.648	1.03	0.943	0.56	0.333
Feeling valued	0.48	0.026	0.28	<0.001	0.29	0.003	0.33	0.088
McKelvey‐Zavoina pseudo R 2	0.53		0.45		0.55		0.50
Intent to leave current job prevalence model*
Fear	1.83	0.119	0.888	0.832	1.07	0.906	2.58	0.557
Anxiety	1.25	0.623	1.99	0.180	1.97	0.224	0.156	0.300
Workload	1.11	0.777	2.53	0.083	2.39	0.144	7.51	0.141
Childcare stress	0.742	0.558	0.332	0.139	1.19	0.775	0.734	0.837
Purpose	0.912	0.813	0.545	0.243	1.05	0.934	0.423	0.566
Feeling valued	0.262	0.001	0.441	0.101	0.222	0.019	0.325	0.525
McKelvey‐Zavoina pseudo R 2	0.28		0.32		0.41		0.73

aOR was adjusted for race and ethnicity, sex, years in practice, and inpatient vs outpatient setting. Burnout is defined primarily as emotional exhaustion. aOR indicates adjusted odds ratio.

^*Due to the small samples, effects are overestimated. All variables are binary; 1 = higher level of the variable, and 0 = lower level of the variable.

Mediation of Outcomes by Feeling Valued

Structural models examined a mediating role for feeling valued; these models assessed if feeling valued may have mediated, or explained in part, the relationship between work environment variables and burnout (Figures 1, 2, 3, and 4). Mplus software modeled a simple multilevel (individuals nested within organizations) single mediator value structure model for physician and nurse burnout separately. Bootstrapping with 1000 draws was used to estimate standard errors for indirect effects. In these models, a strong association of feeling valued on (reducing) burnout is seen in all role types. In addition, a partial mediating effect was seen in nurses, with feeling valued mediating the impact of anxiety/depression and workload on burnout (Table S1), and a full mediation effect for feeling valued and purpose. To illustrate partial mediation, we see that workload can decrease feeling valued, which in turn reduces burnout. If workload is associated with a direct increase in burnout, and feeling valued leads to reduced burnout, then we can say feeling valued partially mediates the effect of workload on burnout, with both a direct effect of workload on burnout and an indirect effect through feeling valued. Feeling valued explained 27% of the total effect of workload on burnout and explained 90% of the total effect of purpose on burnout.

Figure 1. Mediation model for burnout in physicians.

Potentially associated variables are on the left. Paths are statistically significant logits.

Figure 2. Mediation model for burnout in nurses.

Potentially associated variables are on the left. Paths are statistically significant logits.

Figure 3. Mediation model for burnout in advanced practice providers.

Potentially associated variables are on the left. Paths are statistically significant logits.

Figure 4. Mediation model for burnout in other clinical staff.

Potentially associated variables are on the left. Paths are statistically significant logits.

Variation in Outcomes by Years in Practice

Trends in rates of burnout and ITL by number of years in practice (Figures S1 and S1) show differences between role types, with particular vulnerability seen at 16 to 20 years in nurses and OCS.

Qualitative Analyses

Open‐ended comments were provided by 516 (43%) respondents; the majority of comments (218) were from nurses (comments written by 48% of respondents), with 137 from physicians (comments inby 45%), 93 from OCS (comments in 89%), and 68 from APPs (comments by 39%). Main themes identified, including concerns for personal and patient safety, inadequate support from leaders and administrators, financial stresses, feeling undervalued, and workload and related work environment factors, are summarized below.

Personal and Patient Safety

Cardiology HCWs in all roles felt safety was a work‐related challenge. Physicians, nurses, and APPs expressed frustrations about lack of available personal protective equipment. All roles mentioned worries about bringing COVID home to loved ones. Physicians felt maintaining productivity presented challenges to their safety. One physician felt “there was no consistent plan to keep caregivers safe.” Many responses, mainly by nurses, also voiced concerns on patient safety. One nurse stated, “[T]he frequent redeployment of nursing staff to new units increased the risk of transmitting COVID to new and vulnerable patients.”

Inadequate Support From Leaders and Administrators

All role types commented about what they found lacking in administration and leadership. Physicians felt administrators were “out of touch,” felt that they were under “constant scrutiny,” and desired more physicians in decision‐making roles. Nurses felt leaders did a poor job with communication, commenting, “I wish there was more communication and transparency on the redeployment of staff to COVID units.” Other clinical staff felt a lack of support from administration, a sentiment ubiquitous in all role types. An APP commented, “[I] do not feel like administration truly cares about those on the front lines as individuals.”

Financial Stress

Financial stress was a prominent factor affecting wellness. Physicians felt inadequately compensated for the amount of work they were doing. One physician commented on the “vicious stressors induced by organizations with threats to fire/furlough health care providers or reduce their salaries.” Nurses felt a lack of financial incentives to work more. Comments on inadequate compensation were not as prevalent among APPs and OCS.

Feeling Undervalued

Feeling undervalued was a major theme in all roles, especially nurses. Nurses felt they received little recognition and felt “expendable.” Many nurses mentioned “putting themselves at risk while not feeling appreciated by leadership.” APPs and OCS felt similarly. One APP commented, “I feel like I was not a valued part of the organization, but just a cog in the machine, and a disposable one at that.” For physicians, feeling undervalued included sentiments such as “[T]hey do not care about us.”

Workload and Environmental Factors

Physician comments mentioned increased pressures of productivity being particularly stressful during the pandemic, as in, “Physicians cannot be held to the [same] productivity standards [as] pre‐COVID, because we cannot safely see that volume of patients in the ambulatory setting…. Having longer workdays or working many weekends to accomplish this will lead to burnout.” The core theme of increased workload was predominant in nurse comments. Nurses were distressed by lack of control over schedule, breaks, and time off. Many respondents, especially nurses and OCS, felt the consequences of being short‐staffed. Numerous nurses experienced distress when reassigned to different hospital units due to staffing challenges. One nurse commented, “I was mandated to ‘redeploy’ to the intensive care unit during the peak of the COVID‐19 crisis and was very stressed out and worried. I was placed in an unfamiliar work location and felt very out of my comfort zone.”

Conceptual Model Constructed From Quantitative and Qualitative Data

Figure 5 shows an exploratory conceptual model linking demographic and work life variables to outcomes (satisfaction, burnout, and ITL), through the mediator of feeling valued. In addition to reflecting theory in the demand‐control and MEMO models, the theoretical model is supplemented by the work of Cox, where emotional exhaustion (measured by the burnout item in CwC) caused by enduring work stressors is the initial dimension of the burnout syndrome. ¹⁷ The relative importance of work life variables in describing burnout varied depending upon role type, and not all variables were fully tested in the current study. Sex, for example, could not be assessed due to sex imbalances within role types and small samples of individuals of female or male sex within role types. The percent of variance explained by the quantitative variables (McKelvey‐Zavonia pseudo R ²) is seen in the full regression models in Table S1, and ranges from 45% to 55% for percent of burnout explained in the 4 role groups (ie, approximately half of burnout is explained by these variables). A preliminary model derived from the regressions and mediation models was then supplemented by data from the qualitative analysis, which added leadership, safety, and finance to the final proposed model in Figure 5.

Figure 5. General conceptual model of work life and outcomes in cardiology.

Model linking demographic and work life variables to outcomes through a mediator, feeling valued. Note the relationships of these variables vary in their intensity and importance by role type (physician, nurse, advanced practice provider, and other clinical staff). Sex was not addressed due to sex imbalance within role types and small samples of specific sexes within each group. Percent of variance explained by quantitative variables within the model in multivariable regressions was 45% to 55% for burnout (Table S1). Leadership, safety and financial concerns were added in qualitative analyses.

Discussion

In this US national study of 1199 HCWs in cardiology, we assessed the prevalence of burnout and ITL, and multiple workplace‐associated variables as potential contributors among cardiology physicians, nurses, APPs, and clinical staff. Our findings close gaps in the literature by examining all cardiology HCWs and demonstrating: (1) burnout, feeling undervalued, and ITL are common among all cardiovascular HCWs; (2) differences among roles, including nurses reporting the highest prevalence of burnout and ITL; (3) factors including work overload, psychosocial concerns, safety, and finance were prevalent concerns; whereas (4) other factors including feeling valued and supportive leadership appeared to be protective. Structural models produced findings suggesting that interventions aimed at increasing purpose and sense of value could reduce burnout among most cardiology HCWs. These work life factors, defined during the unique times of the pandemic, may be salient for future stressful times that occur postpandemic.

The literature on work life has focused mainly on cardiologists ² and has only rarely addressed other members of the cardiology health care team, ⁶ despite the known relationships between team and physician well‐being. The relational coordination model ¹⁸ emphasizes that shared goals, shared knowledge, and mutual respect lead to favorable patient, team, and clinician outcomes. Thus, one of the unique aspects of our work is the focus on individuals in these understudied but essential roles in cardiology. We found that nursing had the highest burnout and ITL and authored many comments to express work life challenges. Recent articles ¹⁹ , ²⁰ address burnout among nurses, which are now both extended and focused by our data specific to cardiology nurses. Likewise, work conditions, and reactions to those conditions, among cardiology clinical staff have not been well studied (we found only 1 such recent study in a small number of ancillary staff ⁶ ). Given recent staffing shortages and the profound impact this has on clinician well‐being and team efficiency, these findings have particular salience. Our work contributes new insights into challenges faced by each group, as well as the vulnerable career time period of 16 to 20 years in the role for burnout and ITL.

There were certain similarities among role types. Approximately half of all groups (40%–60%) reported burnout, whereas 25% to 50% reported ITL. Workload was related to burnout in all 4 groups. A sense of purpose was not a statistically significant preventative factor, although in the mediation models, this variable facilitated reduced burnout in certain groups as it related to a sense of feeling valued.

Physicians

Although physicians had the lowest burnout rate (40%), this is substantively higher than what was found (26.8%) before the pandemic. ¹ , ² Physicians' feeling valued was strongly associated to lower burnout (aOR, 0.48; P=0.026), confirming the importance of feeling valued in physicians. Meanwhile, feelings of anxiety/depression due to working with COVID were directly associated with higher burnout (aOR, 8.27; P<0.001) in physicians. Although this suggests that psychosocial support strategies (buddy programs, ²¹ peer support, ²² coaching, ²³ or on‐ or offsite counseling) could be valuable to address burnout in cardiologists, it also opens the question as to how system‐level interventions could address these outcomes upstream before symptom onset. Workload was related to burnout among cardiologists (aOR, 6.69), and we speculate this may have been even more of a factor had clinics not been closed during the early pandemic.

Nurses

Nurses reported the highest prevalence of burnout (59%) and the highest prevalence of ITL their jobs (45%). Like physicians, anxiety/depression related to COVID work was directly associated with higher burnout (aOR, 3.9; P<0.001). Nurses' feeling valued mediated against burnout in relation to workload and psychosocial health, and also showed a direct effect on lower burnout (aOR, 0.28 or 72% lower odds of burnout if feeling valued; P<0.001). Feeling valued in nursing also related to mission and purpose. Thus, addressing workload and psychosocial health, as well as promoting purpose and feeling valued, may be beneficial to nurses. To our knowledge, the importance of these factors among nurses in cardiology is new.

Advanced Practice Providers

Burnout in APPs was associated with higher workload (aOR, 4.13; P=0.002) and adverse emotional health (aOR, 4.53; P=0.001). Feeling valued had favorable direct effects on burnout (aOR, 0.29; P=0.003). Mediation models suggested a sense of meaning and purpose contributed to feeling valued, which in turn was associated with lower burnout. These data offer insights into how to address burnout and improve work life among APPs in cardiology.

Other Clinical Staff

Higher workload was strongly associated with burnout in OCS (aOR, 17.4; P<0.001). Mediation models confirmed feeling valued was associated with a direct effect on reduced burnout, and that a sense of purpose facilitated feeling valued. Thus, OCS work life and work intentions may be favorably affected by attention to promoting mission/purpose and feeling valued, and by addressing work overload.

Our study has limitations, and also some strengths. Roll out by institution precluded calculation of response rates by specialty (in this case cardiology) and role types. Details on institutions, including absolute size and academic versus nonacademic status, are not known. Although the rapid roll out and pandemic‐related demands on HCWs may have reduced overall response rate, 34% is greater than most national physician surveys and surveys in cardiologists at the time, ranging from 6% to 20%. ²⁴ It also mandated a briefer survey with fewer work conditions and outcomes than might be optimal for testing. The cross‐sectional analysis does not allow direct assessment of causation. With only 350 cardiologists in the study, and >100 000 cardiologists in the United States, we acknowledge this requires viewing many findings as exploratory. The survey was performed at the height of the pandemic and therefore represents a snapshot of a unique time. Although some variables (eg, fear of exposure) are unique, many others (workload, safety, childcare stress, mission and purpose, feeling valued, and supportive leadership) and their relationships to each other and to burnout are, we believe, generalizable to other pre‐ and postpandemic times. Finally, the survey was administered in 2020, when different states were going through different stages of the COVID pandemic, which might have influenced survey responses, and although we compared variables among different team members, we do not have information of results compared with prior years in most role types. For strengths, there is wide geographic diversity in dozens of health care systems, and attention to 4 role types within cardiology, something not often reported, and the measures have for the most part been used for decades, with good to excellent performance characteristics.

Clinical Implications

Our findings have implications to the well‐being of cardiology HCWs, and, by extension, their ability to provide optimal care as codified by the quadruple aim of health care. ²⁵ Although burnout and ITL were common in all groups, our data identify protective and exacerbating factors, specific to each role, that can directly inform efforts to improve work life. To promote well‐being in cardiology, teams may implement methods to promote wellness and feeling valued. Future interventions should not solely rely on superficial approaches, but also prioritize fundamental improvements in work design and structure to address burnout effectively. These methods may be tailored for each role group. Implementing interventions based upon these data will hopefully lead to higher retention, improved physical/mental health, and better morale/performance among cardiology team members.

Conclusions

Burnout is prevalent among all cardiology HCWs and is highest among nurses and OCS. The pandemic may have altered prevalence of certain variables such as burnout or its contributors. However, we anticipate the relationships between variables may be more stable and comparable in the years to come, because many are reflected in studies performed before the pandemic. Addressing unique factors associated with burnout in different cardiology HCWs could lead to a rewarding and sustainable work life for all team members, including cardiologists. Future studies should address the impact of interventions based upon the remediable work life factors identified, including work overload, anxiety/depression due to stressful work, sense of mission and purpose, safety, leadership, and feeling valued.

Sources of Funding

This work was supported by the American Medical Association.

Disclosures

Dr Sinsky, N. Nankivil, and P. Shah are employed by the American Medical Association. The opinions expressed in this article are those of the authors and should not be interpreted as American Medical Association policy. Dr Linzer performs burnout reduction studies for the American Medical Association, Institute for Healthcare Improvement, Optum Office for Provider Advancement, and other large health care systems. The remaining authors have no disclosures to report.

Supporting information

Data S1