Occupational stress and diminished levels of overall well-being among health care workers were issues of concern even before the coronavirus disease 2019 (COVID-19) pandemic, which exacerbated existing stressors and introduced new challenges to this workforce.1,2 Mental health concerns that have been the focus of attention in this workforce include depression, anxiety, substance use disorders, posttraumatic stress disorder (PTSD), burnout, compassion fatigue, and suicide.

Health care workers have long faced a convergence of stressors that are less common in other types of work. These stressors include the emotional burden of dealing with individuals who are seriously ill or dying; witnessing traumatic events, which is associated with PTSD, particularly among first responders3,4; secondary traumatic stress after exposure to traumatized patients, particularly in emergency departments5,6; witnessing or being the target of workplace violence,7 which can result in adverse physical, psychological, social, and emotional effects7–10; and workplace bullying.10–12 The prevalence of these problems has been reported to be particularly pronounced in emergency and psychiatry departments within hospitals, in the home health setting, and in nursing care facilities.7,9,13 Among the additional stressors for many health care workers are poor job design, management challenges, suboptimal safety climate/safety culture, high caseloads, the effects of shift work and long work hours, and exposure to pathogens.

Most research on mental health in the health care workforce has focused on physicians (including physicians in training) and nurses. Problems noted among physicians include depression, anxiety, substance use disorders, burnout, and suicide.13–17 As with the general public, estimated prevalences of depression among health care workers vary depending on the case definition and characteristics of the measurement instrument (eg, criteria met for depressive disorder versus presence of subclinical depressive symptoms, self-reported symptoms versus provider-diagnosed depression, current depressive symptomatology versus incidence in the last 12 months), as well as demographic characteristics of the respondents (eg, variation by age).14–16,18,19 Depression has also been reported among physicians in training; a systematic review noted estimates of the prevalence of depression or depressive symptoms among residents ranging from 21% to 43%, depending on the case ascertainment variables listed previously,16 whereas a prospective study found an increase in depression based on Patient Health Questionnaire-9 (PHQ-9) scores from 3.9% before the internship year to more than 20% during each quarter of the internship.20 As with depression, suicide risk appears to accelerate during the physician training period.19 Beyond the training period, female physicians have been found to have higher rates of completed suicide, at 1.4 to 2.3 times the rate in the general population,18 although a recently published analysis suggests that the overall suicide risk among physicians is not significantly different from that of the general population.21 Identified risk factors for mental health issues among physicians are both individual and occupational, with the latter including the stress of patient interactions and expectations, easy access to medication, heavy workload, adverse work schedules, and problematic or limited social interaction in the workplace.13

A high prevalence of depression has been noted among registered nurses (RNs).22 A survey of nurses employed by hospitals reported rates of depressive symptomatology of 18%, approximately twice that of the US population, with job satisfaction, body mass index, number of health problems, mental well-being, and health-related productivity significantly associated with depression scores.23 The prevalence of depression among RNs is reported to be highest among those who are young, female, or working in intensive care or psychiatric units.22 Nurses appear to be at higher risk for suicide than both physicians and the general public.24,25

Mental health concerns in the health care workforce are not restricted to physicians and nurses. However, information about the prevalence of mental health problems among other health care occupations is more limited. Rates of depression, stress, and PTSD have been reported for emergency services personnel, but estimates vary widely.26 The scant literature on health care support workers (eg, patient care aides; occupational, physical, and dental aides; phlebotomists) has found that care and support workers have worse mental health than the general working population27 and that patient care aides are more likely to report depression than nurses.28 Female health care support workers have also been found to have elevated rates of suicide.25 Janitors across all industries have a higher prevalence of depression than other workers.29 Although ancillary health care workers such as housekeeping staff do not have direct patient care responsibilities, they frequently work in patient care areas.

Mental health is also a concern for social workers, counselors, psychologists, and others who are tasked with directly addressing the mental health needs of others.30,31 Male human service workers have been found to have higher levels of antidepressant use than other workers at the same skill level.32 Elevated suicide rates have been reported among male welfare support workers, social workers, and female welfare support workers.33

Although the pandemic has led to new mental health challenges for workers globally, health care workers have been particularly at risk because of increased emotional, physical, and organizational demands, as well as increased risk of infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus that causes COVID-19. However, fully addressing the long-term mental health needs of this workforce also requires understanding the levels of baseline, prepandemic mental health issues. To assess baseline mental health and well-being among health care workers in different occupations, we examined 2017 to 2019 data from the Behavioral Risk Factor Surveillance System (BRFSS). Although BRFSS does not include prevalence data on the full range of mental health conditions and is limited to self-reported conditions and diagnoses, it does sample from a wide range of health care (and other) industries and occupations.

To our knowledge, this is the first study to evaluate prepandemic mental health and well-being of the health care workforce using a broad definition of this workforce (health care industry workers who have patient care responsibilities or who work in patient care areas) and including low-wage health care workers. The purpose of this study was to identify segments of the health care workforce that had the highest prepandemic prevalences of selected adverse conditions related to mental health and well-being, as they might require additional services during and after the pandemic.

METHODS Study Population

The BRFSS is a national survey of the noninstitutionalized US adult population (18 years or older) administered by state and jurisdictional health departments.34 Respondents are selected for the survey using random digit dialing techniques on both cellular phones and landlines. Overall response rates for this survey for landlines and cellphones, respectively, by year were 45.3% and 44.5% (2017); 53.3% and 43.4% (2018); and 53.5% and 45.9% (2019). Response rates overall and by state can be found at https://www.cdc.gov/brfss/data_documentation/index.htm.

In addition to a core survey, the BRFSS includes modules that states can opt to include. One of these modules is sponsored by the National Institute for Occupational Safety and Health and collects the industry and occupation of respondents who are employed for wages, out of work for less than 1 year, or self-employed. Occupation and industry are collected through open-ended questions: “What kind of work do you do?” followed by “What kind of business or industry do you work in?” This module is not implemented by the same states each year. A total of 33 states included this module during at least 1 year between 2017 and 2019 (22 states in 2017, 30 in 2018, and 25 in 2019, with 17 states participating all 3 years). We used the 3 most recent years of prepandemic data to enhance reportability for smaller health care occupations.

During BRFSS survey years 2017 to 2019, 314,078 respondents reported that they were employed or self-employed. A total of 51,895 respondents (16.5%) were excluded because of missing or uncodable industry or occupation, active-duty military status, or conflicting employment status information (respondents who reported being employed but whose responses to the industry or occupation question indicated they were unpaid workers, disabled, or retired). Industry and occupation free-text responses were autocoded to 2010 US Census Bureau industry and occupation codes by the National Institute for Occupational Safety and Health Industry and Occupation Computerized Coding System or, for items that could not be coded automatically, by human coders using computer-assisted coding.35

Although we provide results for all health care industry workers combined (all organizationally and self-employed workers with census industry codes 7970 to 8270), the focus of this study was on workers who interact directly with patients, as well as those who work in patient care areas as part of their duties, such as janitors and maids. Non–health care workers (employed outside both health care industries and health care occupations) comprised the comparison group for this work. A third, smaller set of workers are employed in health care occupations but outside the health care industries (eg school nurses, dieticians employed in the sports industry); we excluded them from reporting. Within the health care industry, we present results for occupational groups that had reportable results (denominator size ≥50 and relative standard error for prevalence estimates ≤30%) for at least 3 of the 6 conditions of interest.

Measures

We calculated distributions of demographic characteristics for each health care occupation. We also calculated unadjusted and adjusted prevalences of 6 health conditions elicited in the survey (Table 1). Because well-being and physical and mental health are not independent,36 in addition to conditions that explicitly concern mental health, we assessed prevalences of self-rated overall health, frequent physical distress, and insufficient sleep. Conditions evaluated were self-rated health (fair or poor general health), frequent physical distress (physical health not good at least 14 of past 30 days), frequent mental distress (mental health not good at least 14 of past 30 days), activity limitations (poor physical or mental health preventing usual activities for at least 14 of past 30 days), diagnosed depression, and insufficient sleep (<7 hours average sleep per 24-hour period; elicited only in 2018 BRFSS survey). Because responses for the items reported as number of days cluster at 0 and at multiples of 5 and 7, we did not treat them as continuous variables, instead dichotomizing them.

TABLE 1 - BRFSS Survey Questions Related to Mental Health and Well-being, 2017–2019 Metric Title BRFSS Question Cut Point Poor self-rated health Would you say that in general your health is: excellent, very good, good, fair or poor? Fair or poor = poor self-rated health Frequent physical distress Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good? ≥14 d = yes Frequent mental distress Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good? ≥14 d = yes Activity limitations During the past 30 days, for about how manydays did poor physical or mental health keep
you from doing your usual activities, such as self-care, work, or recreation? ≥14 d = yes Diagnosed depression Has a doctor, nurse, or other health professional ever told you that you had a depressive
disorder (including depression, major depression, dysthymia, or minor depression)? N/A Insufficient sleep* On average, how many hours of sleep do you get in a 24-hour period? < 7 h average per 24-h period
Analysis

To account for the complex survey design and incorporate respondent sampling weights in BRFSS, we used SAS version 9.4 (SAS Institute Inc, Cary, NC) and SAS-callable SUDAAN version 11.0 (RTI International, Research Triangle Park, NC). To estimate population counts and weighted unadjusted prevalences for all variables, we used the SURVEYFREQ procedure. We identified differences in health conditions by health care worker occupations or industries using the RLOGISTIC procedure. We compared health care workers to non–health care workers by performing logistic regressions and estimating adjusted prevalence ratios (aPRs) and their 95% confidence intervals (CIs). Non–health care workers served as the comparison group for the full group of health care workers, as well as for specific subgroups of health care workers. We considered CIs for aPRs that do not span the null to be statistically significant.

Adjustment variables in the primary regression models were sex, race/ethnicity combined (classified as white non-Hispanic, black non-Hispanic, other non-Hispanic, Hispanic), age in years (18–34, 35–54, ≥55), and marital status (collapsed to married or part of an unmarried couple [as a proxy for level of social support] vs all other). All estimates in this report were weighted. Because of the complex relations between income and demographics, occupation/industry, and health outcomes,37,38 we did not adjust for household income.

RESULTS

The 37,685 BRFSS respondents who worked in health care industries were the focus of the study (Table 2). Another 4627 health care workers were employed in non–health care industries; results for this group are not further reported. The 219,871 non–health care workers comprised the comparison population. The largest subset of the 37,685 health care respondents were from the hospital industry (47%), followed by ambulatory care (29%), nursing care facilities (10%), home health (8%), dental offices (4%), and other health care industries (2%).

TABLE 2 - Distribution of Workers From Health Care Occupations Across Health Care Industries, BRFSS 2017–2019 2010 Census Occupation Codes Sample
Size Weighted n (times 1000) Health Care Industry
(U.S. Census Industry Codes) Other
(Census 8180)
(%*) Dental Office
(Census 7980)
(%*) Home Health
(Census 8170)
(%*) Ambulatory Care
(Census 7970, 7990, 8070–8090)
(%*) Hospitals
(Census 8190)
(%*) Nursing Care Facilities
(Census 8270)
(%*) Non–health care workers† 219,871 74,862 Health care workers in non–health care industries‡ 4627 1330 Health care industry workers§ 37,685 12,051 3.7 8.4 29.4 46.5 10.1 1.9   Health care occupation grouping∥ Community and social service occupations 2000–2060 1288 317 NR 2.8 55.6 34.5 5.9 NR  Counselors 2000 516 142 0.0 NR 77.9 17.7 NR NR  Social workers 2010 554 125 0.0 4.1 46.3 40.5 8.9 NR Health care practitioners and technical occupations 3000–3540 18,598 5646 3.5 4.5 26.6 55.8 6.9 2.8 Health diagnosing practitioners ¶ 4100 1235 6.9 1.1 52.2 37.9 1.5 NR  Physicians and surgeons 3060 2445 758 NR NR 58.1 40.6 NR NR Health treating practitioners # 10,737 3126 NR 6.5 20.4 63.7 9.2 NR  Physical therapists 3160 529 152 0.0 10.5 52.7 26.8 10.0 0.0  Registered nurses 3255 8959 2626 NR 6.8 17.1 66.6 9.4 NR Miscellaneous health technologists and technicians 3300–3535 3679 1257 8.6 2.9 16.2 53.7 6.7 11.9  Clinical laboratory technologists and technicians 3300 600 192 0.0 NR 8.4 68.0 NR 18.7  Health practitioner support technologists and technicians 3420 325 119 NR NR 11.1 86.7 NR NR  Licensed practical and licensed vocational nurses 3500 788 251 0.0 13.6 23.0 36.0 27.5 0.0  Miscellaneous health technologists and technicians, other 3535 403 132 NR NR 36.2 59.7 NR NR Health care support occupations 3600–3655 5313 1931 7.2 20.7 22.2 24.2 25.0 0.7  Nursing, psychiatric, and home health aides 3600 3436 1223 0.0 31.6 7.5 24.3 36.6 NR  Dental assistants 3640 385 144 95.8 0.0 NR NR 0.0 0.0  Medical assistants 3645 823 356 NR NR 74.8 23.2 NR NR  Phlebotomists 3649 154 54 0.0 0.0 NR 63.4 0.0 19.5 Food preparation and serving 4000–4160 470 117 0.0 NR 6.6 50.7 41.4 0.0 Building and grounds cleaning and maintenance occupations 4200–4250 777 262 NR NR 16.3 60.3 21.1 NR  Janitors and building cleaners 4220 393 143 NR NR 26.1 58.8 NR NR  Maids and housekeeping cleaners 4230 338 109 0.0 NR NR 62.6 31.3 0.0 Personal care and service occupations 4300–4650 1336 476 NR 58.5 12.0 14.4 14.9 NR  Personal care aides 4610 1156 432 NR 63.6 11.0 13.7 11.5 NR Office and administrative support occupations 5000–5940 3731 1280 5.5 2.1 43.4 44.1 3.7 1.2 Trades** 6200–9750 679 225 NR NR 24.2 50.5 10.1 7.3

Heavier shading indicates broader occupational grouping.

NR, not reported because relative standard error of estimates is >30%.

*Weighted.

†Respondents with census industry codes (0170-7890 or 8290-9500) and census occupation not in (3000–3655).

‡Respondents with census industry codes (0170-7890 or 8290-9550) and census occupation in (3000–3655).

§Respondents with census industry codes 7970 to 8270.

∥Within health care industry, includes health care occupational groups with at least 3 reportable mental health related outcomes (see Table 3).

¶3000, 3010, 3040, 3050, 3060, 3110, 3120, 3140, 3230, 3250, and 3258.

#3030, 3150, 3160, 3200, 3210, 3220, 3235, 3245, 3255, 3256, 3257, and 3260.

**Construction, extraction, maintenance, production, and transportation and materials moving industries.

Demographics of Health Care Workers by Occupation

Demographic characteristics of respondents differed markedly by health care occupation. Although approximately 65% of health care diagnosing and treating practitioners were White, most health care support workers (55%) were non-White, primarily non-Hispanic African American or Hispanic (Table 3). Age distributions also varied by occupation. Educational attainment generally tracked with educational requirements for the occupation: 90% of health diagnosing practitioners and 64% of health treating practitioners had completed college, but only 28% of health technicians and technologists and 14% of health care support workers had done so. Income distribution and home ownership levels followed patterns similar to those observed for education.

TABLE 3 - Demographic Characteristics (Percentages) of Health Care Workers by Occupation, BRFSS 2017–2019 Age, %* Sex, %* Race/Ethnicity, %* Educational Attainment, %* Household Income, %* 18–34 y 35–54 y ≥55 y Male Female Non-Hispanic White Non-Hispanic African American Non-Hispanic Other Hispanic High School or Less Some College/Technical School College Graduate or More <$35 k $35–<$50 k $50–<$75 k ≥$75 k Non–health care industries† 32.3 43.3 24.4 59.5 40.5 60.8 10.8 8.4 19.9 38.6 29.8 1431.6 26.6 12.4 15.8 45.2 Health care industry workers‡ 29.0 45.8 25.2 25.0 75.0 60.3 16.2 11.6 11.9 19.3 34.9 45.8 20.1 11.0 16.8 52.1   Health care occupation grouping§ Community and social service occupations 26.2 50.0 23.8 24.8 75.2 63.4 25.8 4.0 6.8 5.9 17.6 76.5 14.0 12.9 19.7 53.4  Counselors 24.8 53.1 22.1 31.3 68.7 58.8 29.9 NR NR 5.1 19.9 75.0 18.5 14.5 21.6 45.4  Social workers 29.7 45.6 24.6 14.7 85.3 69.5 21.1 NR 4.7 NR 10.2 83.8 8.5 11.7 15.5 64.3 Health care practitioners and technical occupations 26.2 49.1 24.7 24.5 75.5 66.2 12.4 12.9 8.4 6.5 32.0 61.6 7.4 8.2 16.8 67.6 Health diagnosing practitioners 17.6 52.8 29.6 47.3 52.7 64.1 6.8 21.6 7.5 3.4 6.3 90.3 3.2 3.3 5.2 88.3  Physicians and surgeons 14.6 52.9 32.4 56.2 43.8 62.4 6.0 25.1