Coping Mechanisms and Their Associations with Depression and Anxiety Among Healthcare Workers in the Aftermath of COVID-19 Pandemic

Abstract

Objective:

This study investigated the associations between anxiety, depression, and coping in healthcare workers (HCWs) during COVID-19.

Methods:

A survey used the Generalized Anxiety Disorder-2 to assess anxiety, the Patient Health Questionnaire-2 to assess depression, and the Brief COPE scale to evaluate coping mechanisms.

Results:

Of the 1,172 participants who provided data, 24.7% screened positive for anxiety and 15.7% for depression. Logistic regression indicated that avoidance coping (OR=3.92 (95% CI: 2.91–5.29), p < .001) and substance use coping (OR = 1.39 (95% CI: 1.00–1.90), p = .049) were associated with higher odds of depression. Avoidance coping (OR = 2.81 (95% CI: 2.15–3.67), p < .001), and social support coping (OR = 1.59 (95% CI: 1.22–2.08), p = .001) were associated with higher odds of anxiety.

Conclusions:

Interventions are needed to improve HCWs’ coping and wellbeing.

Introduction:

Working in the health professions is inherently stressful, making it essential for healthcare workers (HCWs) to effectively manage the daily demands of patient care¹. Recent research demonstrated an added burden of the COVID-19 pandemic on HCWs’ psychological wellbeing due to several factors including increased workload and engagement hours, emotional exhaustion, staff shortages, and poor physical health^2,3,4. Studies on the impact of the pandemic on mental health have revealed elevated rates of depression (18.7%−29%) and anxiety symptoms (10.1%−32%) among HCWs, with both conditions persisting beyond the peak of COVID-19^1,3. Coping mechanisms—strategies used to manage stress and emotional challenges—play a crucial role in reducing depression and anxiety symptoms^4,5. HCWs employ a variety of coping strategies to mitigate the impact of work-related stressors and manage the additional burden of patient care during the COVID-19 period on their emotional well-being and other aspects of their lives, such as family commitments and caregiving responsibilities⁴.

According to the Stress and Coping Theory, coping reflects strategies, tactics, responses, cognitions, and behaviors that can help individuals adapt to stress and life adversities⁵. Coping may include adaptaive or maladaptive coping mechanisms^3–7. Adaptive coping mechansims are constructive strategies that can help reduce anxiety and improve psychological well-being (e.g., problem-solving, active coping, positive reframeing, and social support coping)^6,7. Maladaptive coping mechansims are counterproductive strategies that can exacerbate stress and distress^6,7. The majority of the studies examining coping mechanism have used the Brief COPE scale and attest to its ability to measure unique dimensions of adaptive and maladaptice coping mechanism among different populations ^3,4. However, these studies also showed greater inconsistencies in the emerging subscales due to several factors, including the research population, study method, magnitude and type of stressful events, and various adaptations to the original items or instructions by the researchers^3,4.

Research assessing coping strategies used by HCWs during the pandemic showed that adaptive coping mechanisms were essential in reducing the psychological burden of the COVID-19 pandemic⁴. Maladaptive coping mechansims such as substance abuse and avoidance coping strategies have been shown to increase both depression and anxiety among HCWs^4,8. A recent systematic review of 89 studies published between 1997 and 2021 including studies on HCWs showed great variability in the coping strategies resulting from the use of the Brief COPE measure ⁹ Factors identified in this review ranged from 2 to 15, with dichotomous coping factors (e.g., adaptiave and maladaptive coping) most frequently identified (25%; n = 21). Studies included in this review were conducted prior to COVID-19 and included only 2 studies on HCWs in addition to other different populations (e.g., cancer patients, family caregivers, community volunteer) making it difficult to confirm the factor structure of the Brief COPE in HCW populations. Therefore, to ensure appropriate examination of the associations between symptoms of depression, anxiety, and coping mechanisms among HCWs, it is important to understand which coping strategies are typically used by HCWs during the pandemic informed by a factor analysis of the Brief COPE items and their unique associations with depression and anxiety.

Thus, the purposes of this study are: a) to examine the incidence of depression and anxiety symptoms in a diverse population of HCWs following the peaks of the COVID-19 pandemic, and b) to investigate the relationship between coping mechanisms and the incidence of depression and anxiety symptoms among HCWs, while controlling for potential socio-demographic and other potential covariates.

Methods

Recruitment

Study participants were recruited via email to complete an anonymous, cross-sectional survey at a large New York City urban academic health system from September to November 2022. Institutional email addresses for a random sample of 10% (4998) of non-physician HCWs (nurses, administrative staff, research staff, scientists, food service staff, security, etc) were gathered through health system leadership. We chose to use a random sample to minimize the large survey burden on our entire workforce. Upon completion, participants were eligible to enter a raffle for prizes such as Apple Watches, Apple AirPods, and Apple iPads. This project was approved by the Institutional Review Board at the study institution. The design and reporting of this study followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines. The STROBE checklist was used to ensure comprehensive and transparent reporting of the study’s methodology and findings (see Supplementary File 1).

Measures

The Generalized Anxiety Disorder-2 (GAD-2) was used to screen for anxiety symptoms ^10. The GAD-2 consists of two items derived from the longer Generalized Anxiety Disorder-7 (GAD-7) scale. These two items assess the frequency of core anxiety symptoms over the past two weeks. Each item is scored on a 4-point scale (not at all (0)- nearly everyday (3). A positive screen for significant GAD-2 symptoms was defined by a score ≥ 3. Cronbach’s alpha coefficient of this scale is 0.89 based on the study data analyses.

The Patient Health Questionnaire-2 (PHQ-2) was used to screen for depressive symptoms ¹¹. The PHQ-2 consists of two items derived from the longer Patient Health Questionnaire-9 (PHQ-9) scale. These items assess the frequency of core depressive symptoms over the past two weeks. Response options were on a 4-point scale (not at all (0)- nearly everyday (3). A positive screen for significant PHQ-2 symptoms was defined by a score ≥ 3. Cronbach’s alpha coefficient is 0.87 based on the study data analyses.

Brief COPE. Following recent reviews of coping strategies and our prior work in this specific population⁴, 18 items from the Brief COPE were included that assess problem solving, seeking social support, and positive reframing in addition to other related items assessing substance abuse and avoidance⁷. Score ranges from one (“I haven’t been doing this at all”) to four (“I’ve been doing this a lot”) with higher scores indicating higher levels of coping strategies used. The result section provides additional information on the internal stability of the 18 items of the Brief COPE used in this study.

Data analysis

First, descriptive analyses were used to examine the charactersitics of the study sample and the rates of anxiety and depression symptoms reported by the study participants. Second, comparative analyses were conducted to examine the differences between the study sample and the study targeted population of interest based on available instituional data. Third, exploratory factor analysis (EFA) was performed on the 18 Brief COPE items. Minimum sample size recommendation for factor analysis includes 20 times the number of variables ¹² or at least 300 ¹³. Our sample size of 1172 was sufficient for EFA. The Kaiser-Meyer-Olkin Test of Sampling (KMO) and Bartlett’s Sphericity Test were computed to assess the adequacy of the variables for factor analysis ¹⁴. We used principal component analysis for the extraction method, as our main goal was to reduce the number of items to representative components. We used oblique rotation as we assume there are interrelationships among the coping factors. To determine the number of factors to extract, we examined the scree plot, eigenvalue value (> 1), and factor loadings (> 0.4). Subscale scores of the Brief COPE were then computed as the mean of the item scores in each factor based on the EFA.

Two multivariables logistic regression models were conducted to determine the association of screening positive for depression and anxiety symptoms with the four identified coping strategies while adjusting for basic demographic variables including age, gender, race, and profession. All the ORs are adjusted ORs which control for all of the other variables in the models.

Missing data were imputed ¹⁵ using the ‘mice’ package from R. Missing data were <15% in each variable except for race/ethnicity (18%). This variable had higher missing data as the question about race/ethnicity has an option for “preferences for not answering this question. We have imputed missing data related to race/ethnicity as well because of the importance for controlling the influence of race/ethnicity in the regression analyses given the focus of the study. Because data on gender, depression, and anxiety were not available for for 56 participants, we limited the study regression analyses to participants who have available data (n=1110). The threshold used to define Alpha is 0.05. All analyses were conducted in SPSS version 28 (IBM, 2021) and R version 4.2.1 (R Core Team, 2022).

Results

Of the 4998 hospital system staff who were sent the survey, 1398 (28.0%) responded. Of those, 1,172 had data on the Brief COPE, GAD-2, and PHQ-2, and were included in the analysis. Comparison of gender, profession, and work setting in the study sample with available human resource data on the 4998 invitees indicated that the study sample had a lower proportions of men, nurses, and administrative personnel (p < .05).

Of the 1,172 participants, 15.7% of participants screened positive for depression symptoms, and 24.7% for anxiety symptoms. The Barlett’s Test of Sphericity, X²(153) = 9401.47, p < .001 and the KMO (0.871) indicated that the 18 items had enough shared variance to make them appropriate for factor analysis. Results of the EFA revealed a four-factor solution that explained 61.9% of the total variance. Factor loadings are presented in Table 2.

Table 2.

Factor loadings from exploratory factor analysis of the 18 items of the Brief COPE

	Pattern Matrix				Structure Matrix
	Factor				Factor
Item	1	2	3	4	1	2	3	4
I’ve been getting help and advice from other people.	0.795	-0.016	-0.033	0.097	0.838	0.134	0.072	0.494
I’ve been getting emotional support from others.	0.773	-0.119	0.084	0.043	0.816	0.048	0.036	0.504
I’ve been getting comfort and understanding from someone.	0.764	-0.107	-0.036	0.146	0.806	0.191	0.100	0.507
I’ve been trying to get advice or help from other people about what to do.	0.740	0.041	-0.023	0.122	0.785	0.056	0.140	0.416
I’ve been expressing my negative feelings.	0.706	0.196	0.044	-0.082	0.704	0.313	0.196	0.334
I’ve been giving up trying to deal with it.	0.016	0.802	-0.012	-0.046	0.129	0.788	0.286	0.168
I’ve been giving up the attempt to cope.	0.088	0.735	0.103	-0.103	0.175	0.763	0.379	0.143
I’ve been saying to myself “this isn’t real.”	-0.086	0.712	-0.023	0.194	0.132	0.739	0.257	0.331
I’ve been refusing to believe that it has happened.	-0.055	0.701	-0.009	0.084	0.106	0.709	0.257	0.235
I’ve been saying things to let my unpleasant feelings escape.	0.409	0.469	0.071	-0.001	0.498	0.566	0.299	0.336
I’ve been using alcohol or other drugs to make myself feel better.	0.025	-0.040	0.966	0.013	0.124	0.356	0.958	0.127
I’ve been using alcohol or other drugs to help me get through it.	-0.003	-0.008	0.959	0.018	0.145	0.331	0.955	0.130
I’ve been trying to see it in a different light, to make it seem more positive.	0.078	-0.045	0.000	0.751	0.520	0.114	0.015	0.792
I’ve been taking action to try to make the situation better.	0.166	-0.080	-0.062	0.736	0.453	0.162	0.082	0.779
I’ve been concentrating my efforts on doing something about the situation I’m in.	-0.030	0.091	0.023	0.729	0.489	0.111	-0.010	0.747
I’ve been looking for something good in what is happening.	0.158	-0.061	-0.089	0.693	0.360	0.282	0.140	0.740
I’ve been turning to work or other activities to take my mind off things.	-0.130	0.122	0.125	0.629	0.227	0.308	0.229	0.609
I’ve been doing something to think about it less, such as going to movies, watching TV, reading, daydreaming, sleeping, or shopping.	0.267	0.083	0.044	0.444	0.513	0.259	0.161	0.607

Factor loadings > 0.4 are in bold.

Factor 1 had high loadings of items related to the use of social supports to cope with stress; we named this factor “Social Support Coping.” Factor 2 had high loadings of items related to giving up or denial; we named this factor “Avoidance Coping.” Factor 3 had high loadings of the two items related to drug and alcohol use; we named this factor “Substance Use Coping.” Finally, factor 4 had high loadings of items related to actively thinking or using strategies to solve the problem; we name this factor “Active Coping.” Consistent with the original factor analysis of the Brief COPE (Carver, 1997)⁷, each of the two items that comprised each factor from the original Brief COPE scale had loadings of > 0.4 on the same factor of this EFA and there was no cross-loading. The one exception was for the two items of the original “Venting” subscale. While the item “I’ve been expressing my negative feelings” had a loading of > 0.4 in the Social Support Coping factor with no cross-loading, the item “I’ve been saying things to let my unpleasant feelings escape” cross-loaded on the Social Support Coping (0.409) and the Avoidance Coping (0.469) Factors. Because theoretically, both of these venting items presumably are associated with communicating feelings to others rather than avoiding the stressor altogether, we decided to keep both items in the Social Support Coping factor (see Table 3). Cronbach’s alpha of the 18 items of the Brief COPE scale used in our study is 0.88. The Cronbach’s alpha for each emerging subscales is as follows: 0.85 for the social support coping subscale, 0.76 for the avoidance coping subscale, 0.92 for the substance use coping subscale, 0.81 for the active coping subscale.

Table 3.

Descriptives for the coping factors, depression, and anxiety in the n=1172

	Minimum	Maximum	Mean	Std. Deviation
Social Support Coping	1.00	4.00	2.14	0.72
Avoidant Coping	1.00	4.00	1.43	0.58
Substance Use Coping	1.00	4.00	1.17	0.48
Active Coping	1.00	4.00	2.49	0.72
The Generalized Anxiety Disorder-2 (GAD-2)	0	6	1.74	1.805
The Patient Health Questionnaire-2 (PHQ-2)	0	6	1.19	1.601

Logistic Regressions

Results from the binary logistic regressions are detailed in Table 4. Controlling for demographic and professional role variables, higher likelihoods of screening positive for depression symptoms were significantly associated with avoidance coping strategies, with an odds ratio [OR] of 3.92 (95% confidence interval [CI]: 2.91–5.29), p < .001, and substance use coping strategies, OR = 1.39 (95% CI: 1.00–1.90), p = .049. Similarly, greater odds of screening positive for anxiety symptoms were associated with avoidance coping strategies, OR = 2.81 (95% CI: 2.15–3.67), p < .001, and social support coping strategies, OR = 1.59 (95% CI: 1.22–2.08), p = .001, after controlling for demographic and professional role variables.

Table 4.

Results of binary logistic regression examining associations between coping factors and depression and anxiety.

	Depression				Anxiety
	95% Confidence Interval				95% Confidence Interval
	Adjusted Odds Ratio	Lower Bound	Upper Bound	p-value	Adjusted Odds Ratio	Lower Bound	Upper Bound	p-value
Demographic variables
Age < 40	0.82	0.56	1.19	0.295	0.64	0.47	0.88	0.005
Female gender	0.89	0.56	1.41	0.612	1.46	0.96	2.24	0.079
Race (ref: White)				0.496				0.185
Asian	0.72	0.42	1.23	0.235	0.70	0.46	1.08	0.110
Latinx	0.93	0.53	1.63	0.794	0.65	0.41	1.04	0.073
Black	0.90	0.52	1.55	0.692	0.66	0.42	1.04	0.072
Other or Multiracial	1.31	0.69	2.46	0.408	1.00	0.59	1.70	0.998
Professional role (ref: nurse)				0.899				0.831
Administrative	1.03	0.63	1.71	0.896	1.26	0.82	1.94	0.298
Research staff	1.13	0.61	2.10	0.689	1.37	0.82	2.27	0.230
Medical assistant	0.95	0.45	2.02	0.892	1.19	0.64	2.22	0.585
Nurse practitioner/physician assistant	0.92	0.44	1.94	0.836	1.31	0.73	2.34	0.360
Other role	0.76	0.43	1.34	0.349	1.15	0.72	1.82	0.564
Coping variables
Social Support Coping	1.27	0.92	1.76	0.146	1.59	1.22	2.08	0.001
Avoidant Coping	3.92	2.91	5.29	<0.001	2.81	2.15	3.67	<0.001
Substance Use Coping	1.38	1.00	1.90	0.049	1.24	0.92	1.68	0.160
Active Coping	0.86	0.61	1.22	0.404	1.12	0.84	1.48	0.453

Discussion

The objectives of this paper were to assess the prevalence of probable depression and anxiety symptoms, coping strategies, and their interrelationships among HCWs during the pandemic. In our study sample of HCWs, mental health screening outcomes revealed elevated levels of probable anxiety (24.7%) and depression symptoms (15.7%) by self-report screening tools (GAD-2 and PHQ-2). These levels remain similar to levels reported during the height of COVID-19 pandemic in other studies ^{2, 16–19}. Research has demonstrated heightened levels of anticipatory anxiety among HCWs when prompted to consider the evolving work environment and patient care during the pandemic. This underscores the importance of enhanced support from leadership and the availability of resources needed to improve mental health among HCWs ¹⁸.Those who provided care to COVID-19 patients reported greater levels of anxiety anddepression symptoms, and stress with nurses and trainee physicians reporting the highest levels ¹⁹ .

Several studies have used the Brief COPE to examine coping strategies in HCWs^3–4, however, our study is novel in that we included healthcare support staff (e.g., administration, food service, and security services), thus providing the perspetive of an understudied population who may be at risk for higher levels of work-related stress. The unique associations between probable depression and anxiety symptoms, and coping domains studied suggest potential targets for psychological and behavioral interventions. Of the four unique coping domains identified in our factor analysis of the 18 items selected from the Brief COPE instrument, avoidance coping and substance use coping were associated with greater odds of screening positive for depression symptoms while avoidance coping and social support coping was associated with greater odds of screening positive for anxiety symptoms. Our findings are in line with previous research that used the Brief COPE scale to examine coping strategies typically used by HCWs. A recent review²⁰ showed that of the 6 studies that used factor analyses to examine the Brief COPE domains (i.e., based on a prior selection of items), the analyses identified four-factor solutions including task-focused active coping, seeking social support, maladaptive management of emotion coping, and acceptance²¹ Variation in the emerging factors could be attributed to sample heterogeneity, types of stressors, and cultural norms that influence both the individual understanding of the meaning of the Brief COPE items and type of coping strategies used.

Our findings of the association between avoidance coping or strategies that include denial or ‘giving up’ and greater odds of screening positive for depression and anxiety symptoms reflect similar findings in other reports studying HCWs^22–23. Indeed, this association was found in previous epidemics, such as the Severe Acute Respiratory Syndrome (SARS), where the effects lasted even after the epidemic ended²³. Avoidance coping, which has generally been considered a maladaptive coping mechanism, has been found to be associated with psychopathology⁸ and therefore, is often the target of psychological interventions, such as cognitive behavioral therapies.

While previous studies have focused primarily on frontline HCWs, such as nurses or physicians, during the early phase of the pandemic, our study results highlights potentially maladaptive coping strategies (i.e., substance use and avoidant coping) across a wider sample of HCWs after the 2020–2021 waves and in mid-2022 as well. These results have important implications for psychological interventions that can focus on replacing HCWs’ avoidant coping behaviors with more active or problem-solving coping strategies.

Additionally, our results also showed that substance use coping strategies used by HCWs were associated with greater odds of screening positive for depression symptoms. Substance use disorders and mental health conditions have often been found to co-occur ^24–25. This was especially true during the pandemic, which saw a large increase in substance use and correlated mental health disorders^26. Increasing evidence showed that health care environments place HCWs at risk for substance use disorders ²⁷ . Over recent years, the incidence of burnout, workplace violence, stigma within diverse HCWs populations, and mental health issues (including suicide risk) has escalated during the COVID-19 pandemic ²⁸ . These challenges have exacerbated difficulties faced by HCWs, prompting the adoption of maladaptive coping strategies such as substance use to manage work-related stress and emotional strain. Consequently, this cycle contributes to substance use disorders and exacerbates mental health issues, including suicide risk among HCWs.

Research in coping and adjustment among HCWs that used the Brief COPE scale showed traumatic stress and burnout decrease during coping interventions but have a slight increase in maladaptive coping mechanisms.⁸ In addition, it was found that adaptive coping strategies such as active coping and help-seeking, were associated with higher wellbeing and better quality of working life whereas maladaptive coping strategies, such as avoidance, were risk factors for lower wellbeing and worse quality of working life⁸.

Lastly, our results showed that the social support coping factor, including two items (“I’ve been expressing my negative feelings” and “I’ve been saying things to let my unpleasant feelings escape”), was associated with greater odds of screening positive for anxiety symptoms. Indeed, previous research has shown a significant association between social support and mental health outcomes ^29–31. However, our finding differs from other research findings supporting a buffering effects of social support coping on mental health (i.e., social support coping is associated with lower risk of anxiety). ^32,33 The negative association between social support coping and anxiety symptoms revealed by our study findings could possibly be due to the cross-sectional nature of our study design as it is likely that HCWs with higher levels of anxiety tend to seek support from their social network, as indicated by the two items above comprising this domain, to reduce their anxiety and stress levels. More longitudinal studies will be needed to explore this possibility.

Limitations

There were several limitations to this study. First, the low response rate of 28.0% reduced the representativeness of the study participants. Some professional groups have less access to email and had lower response rates; therefore, the results may not be generalizable to the entire the population of our organization. Second, due to the cross-sectional survey design, we cannot make causal inferences regarding the relationship between coping, and probable depression and anxiety. Third, we utilized self-report measures to evaluate anxiety and depressive symptoms, rather than clinical ratings, so we cannot establish clinical diagnoses. Lastly, only 18 out of 28 Brief COPE items were included in our survey, limiting generalizability to some degree and potentially missing other key findings. Despite these shortcomings, significant conclusions on a HCW population that is uncommonly studied are drawn from this study and analysis.

Conclusion

In the crucible of the COVID-19 pandemic, our investigation into the interplay of probable depression and anxiety, and coping strategies among HCWs has yielded valuable insights with profound implications for both clinical practice and mental health interventions. The associations between substance use, avoidant coping, and social support seeking coping mechanisms reveals a complex tapestry of factors influencing the psychological well-being of this critical workforce. Our study has illuminated a concerning association between avoidant and substance use coping strategies and heightened levels of depression and anxiety among HCWs. The inclination towards avoidance behavior including substance use, perhaps a mechanism to navigate the overwhelming stressors of the pandemic, appears to exacerbate the burden on mental health. Conversely, the use of social support seeking coping, a potenetially stress mitigating factor (i.e., HCWs with higher levels or anxiety and depression are likely to use social support seeking behavior to mitigate stress levels), suggest a target for future interventions. As such, interventions promoting the cultivation and utilization of social support resources, including educational workshops or peer-support programs (David et al, 2021) ³⁴, may serve as a pivotal element in the mental health toolkit for healthcare workers. Future research should employ longitudinal methodologies and objective assessments to further refine our understanding of these intricate associations.

Table 1:

Descriptive characteristics of the study sample

	n (%)
Total n	1172
Age<40	618 (52.7)
Gender, female	966 (83.0)1
Race
White	285 (24.3)
Asian	258 (22.0)
Latinx	261 (11.3)
Black	254 (21.7)
Other or Multiracial	114 (9.7)
Profession
Nurse	351 (29.9)
Administrative	272 (23.2)
Research staff	135 (11.5)
Medical assistant	99 (8.4)
NP/PA	93 (7.9)
Other	222 (18.9)
Positive depression screen (PHQ-2)	175 (15.7)2
Positive anxiety screen (GAD-2)	276 (24.7)2

PHQ-2=Patient Health Questionnaire-2-item; GAD-2=General Anxiety Disorder-2-item;

¹Gender was missing and could not be imputed for n=8;

²PHQ-2 and GAD-2 were missing and could not be imputed for n=56

Learning Outcomes (LO):

1. Assess the prevalence of depression and anxiety symptoms among healthcare workers.

2. Examine the types of coping strategies employed by healthcare workers.

3. Analyze the relationship between coping mechanisms and the incidence of depression and anxiety symptoms among healthcare workers.