1 INTRODUCTION

As of January 10, the COVID-19 pandemic has led to 1,932,948 deaths worldwide and 30,049 in Peru (Johns Hopkins University of Medicine, n.d.). The rates of professionals with COVID-19 vary among countries, but it is worth noting that the number is substantial. In Latin America, the deaths of 2741 physicians have been reported dead, including 257 in Peru (Colegio Médico del Peru, n.d.).

In all countries, health systems, including hospitals and health centers, had to be reorganized to reflect new care priorities for patients with COVID-19. These adaptations within health systems forced changes and disrupted the usual work dynamics of health professionals; in many cases, these adaptations caused work changes, such as the transfer of professionals from their usual practice to first-line care, and, in many cases, forced a reorganization of all care flow patterns (Moreno et al., 2020). In addition, the lack of adequate resources and protective gear to guard against possible infections generated high rates of COVID-19 infections among professionals (Erquicia et al., 2020). Changes in routine practice related to COVID-19, the high risk of COVID-19 infection, and uncertainties in decision-making are new workplace stressors that can potentially affect the mental health of professionals and are risk factors for the development of psychological distress, anger, and depression (Sanghera et al., 2020; Serrano-Ripoll et al., 2020).

There is strong evidence suggesting that trauma experienced during childhood or adolescence can have a particular etiological significance in determining the mental health outcomes of adults (McKay et al., 2020). However, little is known about the effects of early traumatic experiences during childhood or adolescence that can modulate health professionals’ responses to new stressors that have arisen since the beginning of the COVID-19 pandemic.

There is consistent evidence in the literature that an individual's resistance to persistent stress is closely related to his or her personal history of adverse childhood experiences (ACEs) and is associated with lower resilience (Ayyala et al., 2020; Barnes et al., 2020; Humphreys & Zeanah, 2015; Kessler et al., 2010). Some authors find that ACEs are also associated with an increased risk of anxiety in adulthood (Bandelow et al., 2004; Gal et al., 2011) and with other psychiatric problems, including substance use disorders (Bandelow et al., 2004; Gal et al., 2011; Norman et al., 2012; Pesonen et al., 2007; Räikkönen et al., 2011), personality disorders (Johnson et al., 1999; Lahti et al., 2012), psychosis (Varese et al., 2012), and suicidal tendencies (Dube et al., 2001; Norman et al., 2012). Considering that ACEs can be a predisposing factor for anxiety disorders, depression, and pathological stress (Norman et al., 2012) and that the COVID-19 pandemic is a strong and constant stressor (Shigemura et al., 2020), the probability that health workers who have experienced ACEs will suffer from these pathologies is very high (Siu, 2008; Xiang et al., 2020).

The objective of the study was to evaluate, in health professionals in Lima (Peru), the relationship between mental health problems (anxiety, depression, and perceived stress), and the care of patients with COVID-19 and early adverse events. Likewise, to evaluate whether early adverse events modulate the relationship the care of patients with COVID-19 and mental healt problems in these professional.

2 METHOD 2.1 Sample

Data from a sample of 542 health professionals working in the Lima region, mainly women (64.3%) aged between 24 and 87 years (Table 1), were collected through an online questionnaire using Google Forms. Participants were recruited through advertisements and social media. Due to the lack of health professionals for the care of patients with COVID-19, all health professionals, regardless of their training, were assigned to work directly or indirectly in the care of these patients. For this reason, the following were included in the study: doctors, nurses, midwives, medical technologists, nutritionists, pharmaceutical chemists, nursing technicians, psychologists, and biologists. The survey was conducted in May and June 2020 at the peak of contagion and mortality in Lima, Peru.

TABLE 1. General characteristics of health workers. 2020 Mental Health Survey Lima Variable N° % Female gender 344 63.5 Marital status Married 323 59.6 Single 219 40.4 Professional group Physician 302 55.7 Nurse 90 16.6 Midwife 64 11.8 Other 86 15.9 COVID-19 care 216  39.9  Change in service 138 25.5 Diagnosed with COVID-19 35 6.5 2.2 Measurements

The instrument consisted of 48 questions distributed among seven sections. The questionnaire was anonymous to ensure the privacy and confidentiality of the participants. The structure of the survey did not allow advancing to the next question until the previous question was answered. The variables included personal, occupational, occupational exposure to COVID-19 and recent COVID-19 diagnosis, ACE, anxiety, depression, and acute stress (AS).

Exposure to COVID-19 was measured by two questions that were incorporated into the questionnaire: one asked whether the health professional was involved in the direct care of patients with COVID-19, and the other asked whether the health worker was moved to a COVID-19 patient care area different from his/her usual area of work.

For the present study, the ACE data was obtained according to the procedure defined by J. Douglas Bremner in his Early Trauma Inventory Self Report-Short Form (ETI), which addresses the following types of traumas occurring before the age of 18 years (Posada et al., 2019):

General trauma investigates the occurrence of some type of negative, violent, or traumatic event in which the subject only witnessed violent events.

Physical abuse refers to the experience of physical abuse in childhood.

Emotional abuse refers to the experience of emotional abuse in childhood.

Sexual abuse refers a history of abuse, mistreatment, trauma, or physical invasion of a sexual nature.

Due to limitations of the survey, we only included four questions addressing each type of trauma (See Appendix). The ETI has been validated in Colombia (Posada et al., 2019).

Anxiety in the last 2 weeks was measured with the generalized anxiety disorder scale, version 7 (GAD-7), which has been validated in Peru (Zhong et al., 2015); the presence of anxiety is determined when the score is > 4 (Spitzer et al., 2006). Depression in the last 2 weeks was assessed with the Patient Health Questionnaire version 9 (PHQ-9) (“Cuestionario sobre la salud del Paciente (PHQ-9),” n.d.), which was also validated in Peru (Villarreal-Zegarra et al., 2019); the presence of depression is determined when the scores is > 4 (Kisely et al., 2020). Acute stress in the last month was measured with the Global Perceived Stress Scale version 10 (PSS-10) (Cohen et al., 1983). The median value of the obtained distribution was used as a cutoff point; values of 16 or higher were considered indicative of the presence of AS, and values less than 16 were considered indicative of the absence of AS (Guzmán-Yacaman & Reyes-Bossio, 2018). The PSS has been validated in our setting (Raygada Zolezzi, 2019).

2.3 Statistical analysis

The power and size of the sample was estimated before the study using Epidat v 3.1 (Junta de Gobierno de Galicia, n.d.) and considering maximum variability, a significance of 0.05, an accuracy of 5%, and a data loss of 10%. The minimum sample size required was 430 workers.

The statistical analysis of the data was performed with the statistical software Stata v 16. First, the frequency distributions and summary measures were analyzed, allowing us to identify the patterns and distribution of the variables. Second, we examined the relationships between the dependent variables (depression, anxiety, and acute stress) and the independent variables (caring for patients with COVID-19, change in service, being diagnosed with COVID-19, general trauma, physical abuse, psychological trauma, sexual abuse, gender, and age) using bivariate and multivariate logistic regression analyses. In the logistic regression analysis, two models were used: 1) a model without interactions; 2) a model with interactions between caring for patients with COVID-19 and a) each type of ACE, b) gender, and c) age. The latter model was used to evaluate whether the effect of caring for patients with COVID-19 on mental health was modified by ACEs, gender and/or age.

In case of finding a significant interaction between an ACE and the care of patients with COVID, stratified analyzes were performed, separating the sample into individuals with and without the ACE; and on the other hand, dividing it into health personnel who attend and who do not attend to patients with COVID-19.

Cronbach's α test was performed to determine the internal consistency of the questions on the questionnaire concerning depression, anxiety, and stress.

2.4 Ethics

The study procedures were conducted in accordance with the ethical standards for human experimentation. The study was approved by the “Carlos Alberto Peschiera Carrillo” National Cardiovascular Ethics Committee - INCOR - EsSalud (Certificate of Approval 09/2020).

3 RESULTS

A total of 542 health professionals completed the online questionnaire. The dropout rate related to abandoned or incomplete questionnaires was not estimated since the web platform did not allow it. The demographic characteristics of the health workers are summarized in Table 1.

The Cronbach's α values of the PHQ-9, GAD-7, and PSS in our sample showed high internal consistency and were 0.9081, 0.9284, and 0.8309, respectively.

The mean age of the respondents was 49.2 years (minimum 24 and maximum 87). A total of 63.5% were women. The women had a mean age of 47.5 years, and the men had a mean age of 52.3 years; the difference was statistically significant (F = 17.269, p < .001, df = 541). The number of participating physicians (55.7%) was larger than the number of nurses (16.6%) and midwives (11.8%).

The prevalence of anxiety, depression, and acute stress in the total sample was 54.2%, 44.5%, and 50.7%, respectively. In all three domains, women had significantly higher scores than men (Table 2).

TABLE 2. Distribution of adverse childhood experiences, anxiety and depression according to gender. 2020 Mental Health Survey Lima Total sample Men Women Statistical difference between men and women Variables N % N % N % 0R p Anxiety 294 54.2 77 38.9 217 63.1 2.69 (1.87–3.85) <.001 Mild-severe depression 241 44.5 64 32.3 177 51.5 2.22 (1.54–3.20) <.001 Acute stress 242 50.7 78 39.4 197 57.3 2.06 (1.44–2.94) <.001 General trauma 383 70.7 135 68.2 248 72.1 1.21 (0.82–1.76) NS Emotional abuse 324 59.8 99 50.0 225 65.4 1.89 (1.32–2.70) <.001 Physical abuse 170 31.4 65 32.8 105 30.5 0.90 (0.62–1.31) NS Sexual abuse 56 10.3 9 4.5 47 13.7 3.23 (1.60–6.94) <.001

Female health professionals had a significantly higher prevalence of emotional abuse than males did (OR = 1.89, p < .001). Men reported a significantly higher prevalence of sexual abuse than women (OR = 3.23, p < .001) (Table 2). There were no significant differences between genders in general trauma and physical abuse.

In the context of COVID-19, 25.5% of health workers were assigned to another service for pandemic-related reasons, while 39.9% of health workers treated patients with COVID-19. Of these workers, 6.5% reported having been diagnosed with COVID-19.

3.1 Bivariate and multivariate analysis

Anxiety was significantly more frequent in younger and female professionals (Tables 2 and 3). Anxiety was also strongly associated with reports of experiences of general trauma, emotional abuse and sexual abuse (but not physical abuse), as well as younger age, female gender and COVID care. However, the effect size was small for all variables (Table 3).

TABLE 3. Association between anxiety and the different variables according to the bivariate analyses and multivariate models with and without interactions Anxiety Bivariate analysis Model adjusted to include all variables Model with interaction Variable OR p All variables p Direct effect p With COVID care p General trauma 1.55 (1.07–2.25) .021 1.21 (0.78–1.86) NS 1.35 (0.79–2.32) NS 0.63 (0.25–1.58) NS Physical abuse 1.36 (0.94–1.96) NS 0.83 (0.52–1.31) NS 0.88 (0.50–1.60) NS 0.80 (0.31–2.09) NS Emotional abuse 2.27 (1.60–3.22) <.001 1.80 (1.17–2.77) .009 2.01 (1.17–3.44) .011 0.86 (0.34–2.15) NS Sexual abuse 2.79 (1.48–5.23) <.001 1.95 (0.99–3.82) NS 0.89 (0.38–2.15) NS 7.71 (1.62–36.80) .01 Female gender 2.69 (1.87–3.85) <.001 2.20 (1.51–3.22) .004 2.61 (1.60–4.24) <.001 0.61 (0.27–1.34) NS Age* 0.97 (0.96–0.99) <.001 0.98 (0.97–0.99) <.001 0.99 (0.97–1.01) NS 0.99 (0.97–1.01) NS Change in service 1.05 (0.74–1.47) NS 1.00 (0.69–1.44) NS 1.00 (0.70–1.50) NS COVID care 1.59 (1.12–2.26) .009 1.18 (0.78–1.78) NS 4.65 (0.67–32.20) NS COVID diagnosis 1.29 (0.64–2.59) NS 2.11 (0.82–5.43) NS 0.96 (0.45–2.10) NS

In the logistic regression models adjusted to include all variables, anxiety was significantly associated with female gender (OR = 2.20, p = 0.004) and emotional abuse (OR = 2.01 p = .011), and age maintained an inverse association (OR = 0.98, p < .001).

In the logistic regression models that introduced each of the traumatic events, including the interaction between COVID-19 care and the reported intensity of anxiety, it was found that a history of sexual abuse significantly increased self-reported anxiety in professionals working in COVID-19 care (OR = 7.71, p = .010) (Table 3 and Figure 1).

interaction associations between COVID-19 care and ACE, anxiety, and female gender. Note: The graph shows how the coexistence of childhood sexual abuse and COVID-19 care was significantly associated with anxiety (OR: 7.71)

In the stratified analysis, considering only individuals without a history of sexual abuse (n = 486), there was no significant relationship between the care of patients with COVID and anxiety, both in the bivariate analysis (OR = 1.33, 95% CI: 0.92−1.92) as in the one adjusted for sex and age (OR = 1.07, 95% CI: 0.70–1.63); on the other hand, considering only the participants with a history of sexual abuse (n = 56), a significant relationship was observed: crude OR = 7.33 (95% CI: 1.76–30.61) and OR adjusted for sex and age = 5.23 (95% CI: 1.17–23.47). Likewise, dividing the sample between the professionals who attended to patients with COVID (n = 216) and those who did not (n = 326), the first group did show a relationship between a history of sexual abuse and anxiety: crude OR = 7.27 (95 CI %: 2.13–24.76), OR adjusted for age and sex = 6.59 (CI: 1.90–22.85); while there was no significant relationship between those who did not attend to patients with COVID: crude OR = 1.31 (95% CI: 0.58–2.99), OR adjusted for age and sex = 1.04 (95% CI: 0.45–2.43).

3.2 Depression

Self-reported depression was significantly associated with the four types of traumatic experiences, younger age, female gender, and COVID-19 care (ORs between 0.97 and 3.23, significance p = .005 and < .001). The effect size was small for all variables (Table 4).

TABLE 4. Association between depression and different variables according to the bivariate analysis and multivariate models with and without interactions Depression Bivariate analysis Model adjusted to include all variables Model with interaction Associated variables OR p All variables p Direct Effect p With COVID care p General trauma 1.72 (1.17–2.52) .005 1.38 (0.89–2.14) NS 1.44 (0.82–2.52) NS 0.84 (0.33–2.11) NS Physical abuse 1.54 (1.07–2.21) .002 0.95 (0.61–1.50) NS 1.10 (0.59–1.90) NS 0.75 (0.30–1.90) NS Emotional abuse 2.07 (1.45–2.95) <.001 1.38 (0.89–2.13) NS 1.39 (0.80–2.40) NS 1.07 (0.42–2.64) NS Sexual abuse 3.23 (1.77–5.86) <.001 2.36 (1.24–4.49) .008 2.01 (0.83–4.85) NS 1.51 (0.41–5.59) NS Female gender 2.22 (1.54–3.20) <.001 1.81 (1.23–2.68) .003 1.96 (1.20–3.24) .009 0.81 (0.36–1.82) NS Age* 0.97 (0.95–0.98) <.001 0.98 (0.96–0.99) .005 0.98 (0.96–1.00) NS 0.98 (0.96–1.02) NS Change in service 1.26 (0.89–1.77) NS 1.22 (0.85–1.77) NS 1.21 (0.83–1.77) NS –.– COVID care 1.99 (1.40–2.82) <.001 1.36 (0.91–2.04) NS 3.12 (0.46–20.99) NS –.– COVID diagnosis 1.35 (0.68–2.68) NS 0.96 (0.46–2.02) NS 0.98 (0.17–2.28) NS –.–

In the logistic regression models adjusted to include all variables, depression was significantly associated with female gender (OR = 2.20, p = .004) and emotional abuse (OR = 2.01, p ≤ .011), although the effect size was small. Age also presented an inverse relationship, with younger age associated with an increased risk of depression (OR = 0.98, p = .005)

The multivariate analysis with interactions showed that female gender was significantly associated with depression, but there were no significant interactions associating traumatic events with COVID-19 care and depression in health professionals. (see Table 4)

3.3 Acute stress

A self-reported history of emotional abuse (OR = 2.21, p = .008) or sexual abuse OR = 3.23, p < .001) was statistically significantly associated with AS. Professional women presented higher scores for AS (OR = 2.69, p < .001) as did younger professionals overall (both men and women) (OR = 0.97, p = .020). No significant associations were found between the variable COVID-19 care and the intensity of AS.

In both the logistic regression models that included all variables and the logistic regression model with interactions, no significant interactions of ACEs were found for the association between COVID-19 care and current stress. Age maintained an inverse relationship that was statistically significant, and female gender was associated with the presence of AS (OR = 1.73, p = .004) (see Table 5).

TABLE 5. Association between acute stress and different variables according to the bivariate analysis and multivariate models with and without interactions Acute stress Bivariate analysis Model adjusted to include all variables Model with interaction Associated variables OR p All variables p Direct Effect p With COVID care p General trauma 1.14 (0.79–1.65) NS 0.89 (0.58–1.37) NS 1.03 (0.61–1.74) NS 0.66 (0.27–1.64) NS Physical abuse 1.31 (0.91–1.88) NS 0.91 (0.58–1.42) NS 0.97 (0.54–1.71) NS 0.86 (0.34–2.13) NS Emotional abuse 1.89 (1.34 −2.68) <.001 1.60 (1.05–2.46) NS 1.56 (0.92–2.65) NS 1.09 (0.44–2.67) NS Sexual abuse 2.21 (1.23–3.99) .008 1.68 (0.90–3.15) NS 1.41 (0.58–3.45) NS 1.39 (0.39–4.96) NS Female gender 2.06 (1.44–2.94) <.001 1.73 (1.19–2.52) .004 1.58 (0.98–2.54) NS 1.24 (0.58–2.71) NS Age* 0.98 (0.96–0.99) <.001 0.98 (0.96–0.99) .014 0.98 (0.96–1.00) .020 1.01 (0.98–1.05) NS Change in service 1.26 (0.89–1.78) NS 1.28 (0.89–2.52) NS 1.31 (0.91–1.89) NS –.– COVID care 1.18 (0.84–1.6