Sample descriptives are presented in Table 1. Overall, 962 HCWs took part in this study. The sample was 88.4% female, 85.6% Caucasian, and ranged from 18 to 80 years of age (mean [M] = 44.6, standard deviation [SD] = 12.8 years). Most respondents were located in Ontario (74.0%, n = 641) or Quebec (22.6%, n = 196), and included a higher proportion of nurses (28.1%, n = 270), followed by allied mental health specialists (22.2%, n = 214), allied health specialists (19.6%, n = 189), physicians (16.9%, n = 163), and administrators (13.1%, n = 126). A higher proportion of participants were working in psychiatry (11.3%, n = 40), emergency medicine (7.0%, n = 25), and cardiology/thoracic surgery (7.0%, n = 25) compared to other disciplines. Overall, 26.8% of respondents (n = 255) had been tested for COVID-19 prior to survey completion, and out of those, 5.5% indicated that they had tested positive (n = 14). Within the global sample, 31.8% (n = 291) reported that their work involved contact with patients diagnosed with COVID-19. These respondents had similar length of time elapsed since the start of the pandemic to survey completion and similar distributions in terms of sex, race, income level, and past trauma exposure compared to those who had not been in contact with patients with COVID-19 (Table S1, all p > .050). Conversely, those who were in contact with COVID-19 positive patients were significantly younger (39.6 + 11.3 vs 46.5 + 12.7 years old; F (1, 913) = 62.9, p < .001, np2 [partial eta squared, an estimate of effect size] = .064).
Moral distress and mental health classified by exposure to patients with COVID-19
A higher proportion of HCWs in contact with patients with COVID-19 reported having a current diagnosis of a mental disorder (27.4%, n = 76/277) compared to those who were not in contact with COVID-19 positive patients (18.5%, n = 110/593; Chi-squared = 8.9, p = .003).
After controlling for age and self-reported current mental disorder diagnoses, the severity of moral distress (F (1, 866) = 69.9, p < .001, np2 = .075), anxiety (F (1, 833) = 5.5, p = .020, np2 = .007), and depression (F (1, 805) = 5.5, p = .020, np2 = .007) were significantly higher in HCWs exposed to COVID-19 positive patients compared to those who were not (Fig. 2A). No significant effect was found for stress (PSS, F (1, 767) = 3.6, p = .058, np2 = .005).
The proportion of individuals falling within the second or third tertile of moral distress (i.e. higher levels of moral distress) and the proportion of those screening positive for stress, anxiety and depression were higher in respondents who were in contact with patients with COVID-19 compared to those who were not (Chi-squared > 9.1, p = .002, Fig. 2B).
Moral resilience moderates the relationship between exposure to potentially morally distressing events and moral distress
In the multiple linear regression model aiming to assess moderation, higher exposure to PMDEs (B = 2.98, 95% CI [2.90, 3.05]) and lower moral resilience (B = 3.12, 95% CI [0.04, 6.20]) were associated with more severe moral distress (Table 2). The R2 of the model integrating the interaction term (exposure by moral resilience) was significantly higher than the main effects model (R2 = .93 vs .92; F Change (1, 898) = 52.6, p < .001). The interaction term between PMDEs exposure and moral resilience was significantly associated with moral distress. As can be seen on Fig. 3, moral resilience moderated the positive association between exposure to PMDEs and moral distress. Specifically, compared to subgroups with higher moral resilience (second and third tertiles), the subgroup with the lowest moral resilience (first tertile) had a steeper cross-sectional worsening in moral distress as the frequency of PMDEs increased (i.e. individuals with more frequent PMDEs were more prone to have higher moral distress if their level of moral resilience was low). The multiple regression model also revealed that older age (B = 1.22, 95% CI [0.41, 2.02]) was associated with higher levels of moral distress. There was no significant independent association between moral distress and the time elapsed since the pandemic declaration.
Associations between moral distress/resilience and mental health
After controlling for relevant covariates, higher levels of moral distress were modestly but significantly correlated with more severe stress (PSS, n = 763, r = .29, p < .001), anxiety (GAD7, n = 829, r = .28, p < .001), and depression (QIDS-SR16, n = 801, r = .27, p < .001) symptoms. These correlations persisted after controlling for the degree of moral resilience, although correlation coefficients dropped to lower values (PSS, n = 758, r = .18, p < .001; GAD7, n = 823, r = .19, p < .001; QIDS-SR16, n = 796, r = .17, p < .001).
Higher moral resilience correlated with better mental health outcomes as reflected by lower stress (PSS, n = 763, r = .29, p < .001), anxiety (GAD7, n = 829, r = .28, p < .001), and depression (QIDS-SR16, n = 801, r = .27, p < .001) symptoms (Fig. S1).
Factors associated with moral resilience
Stronger moral resilience was significantly associated with: being a male (B = 0.12, 95% CI [0.03, 0.21]), older age (B = 0.08, 95% CI [0.05, 0.10]), not having a self-reported current diagnosis of a mental disorder (B = − 0.19, 95% CI [− 0.26, − 0.12]), sleeping more (B = 0.02, 95% CI [0.00, 0.04]), and higher level of support from one’s employer and colleagues (B = 0.12, 95% CI [0.06, 0.17]; Table 3). Conversely, there was no significant independent association between moral resilience and family support, HCW subtype, or the time elapsed since the pandemic declaration.