*3.2. Analyses*

The Chi-square test of independence (Table 1) conducted on the two samples revealed that NHCIs were generally younger (χ2(3) = 538.75, *p* < 0.001) and had had their diploma for a shorter time than NPPs had (χ2(2) = 753.33, *p* < 0.001). NHCI were also less frequently exposed indirectly to COVID-19, but more frequently exposed directly (χ2(2) = 98.85, *p* < 0.001).

Independent sample *t*-tests showed that NHCIs tended to receive more social support (t = 6.07, *p* < 0.001, *d* = 0.13) and were less resilient (t = −6.60, *p* < 0.001, *d* = 0.14) than NPPs, though effect sizes were very small. Moreover, the NHCIs reported seeking instrumental (t = 7.53, *p* < 0.001, *d* = 0.16) and emotional (t = 7.95, *p* < 0.001, *d* = 0.17) support, venting (t = 4.71, *p* < 0.001, *d* = 0.10), turning to religion (t = 4.85, *p* < 0.001, *d* = 0.10), and exhibiting behavioral disengagement (t = 5.58, *p* < 0.001, *d* = 0.12) more often than NPPs. On the contrary, NHCIs used active coping (t = −3.57, *p* < 0.001, *d* = 0.08), planning (t = −3.58, *p* < 0.001, *d* = 0.08), positive reframing (t = −5.85, *p* < 0.001, *d* = 0.12), and acceptance (t = −4.46, *p* < 0.001, *d* = 0.09) less often than NPPs.

Pearson's correlations among the variables are provided in Table 2. The strongest correlations were between QoL and perceived stress (*r* = −0.69), planning and active coping (*r* = 0.65), seeking instrumental and emotional support (*r* = 0.61), and seeking instrumental support and venting (*r* = 0.60). All other rs were below 0.60. With a sample size as large as 9898, any correlation larger than 0.04 would be significant at *p* < 0.0001. It is thus not surprising that almost all correlations were statistically significant. Consequently, the correlations were not interpreted based on obtained *p*-values alone.



Variables 5 to 18 are the dimensions measured by the Brief COPE Inventory. Bold font: *p* < 0.0001.

#### *3.3. Regression Analysis*

We performed multiple linear regression analyses, first on the full sample and then on each subsample (type of practice). In each analysis, QoL was the outcome variable, and the predictors were perceived stress, social support, resilience, coping styles, exposure to COVID-19, reassignment during the pandemic, and sociodemographic variables (Table 3). Adjusted R<sup>2</sup> values ranged from 0.62 to 0.63 in the three analyses.

The analyses revealed that perceived stress was associated with QoL in the full sample (β = −0.49, 95% CI: [−0.51, −0.48]), and in the NHCI and NPPs samples (β's = −0.49 and −0.50, respectively): Respondents reporting greater perceived stress also reported lower quality of life. Social support was positively associated with QoL in all analyses (β's ranged between 0.20 and 0.21, *p* < 0.005): the more respondents reported having social support, the higher their QoL. Lastly, resilience was not significantly associated with QoL (β's ranged between 0.01 and 0.02, *p* > 0.005).

**Table 3.** Association between variables and quality of life.



**Table 3.** *Cont.*

**\***: *p* < 0.005; Remark: some nurses who participated in the study had a mixed practice alternating between hospital/care institution and private work. They were included in the general analyses, but because of how few they were in number (*n* = 553) compared to the others, they were excluded from the stratified analyses. NHCI: nurses working at hospitals or care institutions; NPP: nurses in private practice; CI: confidence interval.

Higher level of coping-oriented venting (β's ranged between 0.03 and 0.05, *p* < 0.005 in the three analyses), as well as active coping (β's ranged between 0.01 [NHCI] and 0.04 [full sample], *p* < 0.005, positive reframing (β's ranged between 0.08 and 0.12, *p* < 0.005 in the three analyses, and acceptance (β's ranged between 0.06 and 0.08, *p* < 0.005), were associated with higher levels of QoL. Higher levels of coping oriented toward denial (β's ranged between −0.04 and −0.03, *p* < 0.005), blaming one-self (β's ranged between −0.08 and −0.06, *p* < 0.005), substance use (β's ranged between −0.04 and −0.03, *p* < 0.005), and behavioral disengagement (β's ranged between −0.05 and −0.06, *p* < 0.005 were associated with lower levels of QoL.

Direct and indirect exposures to COVID-19 compared with no exposure were not associated with QoL (β's ranged between −0.03 and −0.01, *p* > 0.005). Neither was reassignment during the pandemic (β's ranged between −0.01 and 0.00, *p* > 0.005). Age was negatively associated with QoL, with older participants reporting lower QoL (β's ranged between −0.07 to −0.10 for age classes compared to 18–29 years old, *p* < 0.005 for NHCI and the full sample).
