ABSTRACT
The aim of this study was to identify how previously existing burnout and its changes during the pandemic contributed to PTSD symptoms and psychological distress in a cohort of 388 healthcare workers (HCWs). Each HCW was surveyed in Sep 2019 (before COVID-19) and again in Dec 2020-Jan 2021 (during the pandemic) to assess burnout (MBI); and in the second wave only to assess PTSD (PCL-5-SF), psychological distress (GHQ-12) and resilience (CD-RISC-10). Changes in emotional exhaustion (EE) and depersonalisation (DEP) were stronger in HCWs with lower EE and DEP baseline values. HCWs with higher baseline poor personal accomplishment (PPA) improved more than those with lower baseline values. In multivariable-adjusted models, pre-pandemic EE and its changes were equally associated to both outcomes: standardised-ßs of 0.52 and 0.54 for PTSD, respectively; and 0.55 and 0.53 for psychological distress. Changes in DEP were associated with PTSD only (0.10). Changes in PPA had a higher association with psychological distress (0.29) than pre-pandemic PPA (0.13). Resilience was associated with lower psychological distress (-0.25). Preventive actions aimed at reducing EE, e.g., addressing organisational dysfunctions, are needed to mitigate the impact of future crises, whereas improving personal accomplishment levels is a key target to protect HCWs from mental health disorders during a pandemic.
ABSTRACT
It is unclear if the factor structure of the questionnaires that were employed by studies addressing the impact of COVID-19 on the mental health of Healthcare Workers (HCW) did not change due to the pandemic. The aim of this study is to assess the factor structure and longitudinal measurement invariance of the Maslach Burnout Inventory (MBI) and the factor structure of the General Health Questionnare-12 (GHQ-12), PTSD Checklist for DSM-5-Short Form (PCL-5-SF), Connor-Davidson Resilience Scale-10 (CD-RISC-10) and Post-Traumatic Growth Inventory-Short Form (PTGI-SF). Out of n = 805 HCWs from a University hospital who responded to a pre-COVID-19 survey, n = 431 were re-assessed after the COVID-19 outbreak. A Confirmatory Factor Analysis (CFA) on the MBI showed adequate fit and good internal consistency only after removal of items 2, 6, 12 and 16. The assumptions of configural and metric longitudinal invariance were met, whereas scalar longitudinal invariance did not hold. CFAs and exploratory bifactor analyses performed using data from the second wave confirmed that the GHQ-12, the PCL-5-SF, the PTGI-SF and the CD-RISC-10 were unidimensional. In conclusion, we found support for a refined version of the MBI. The comparison of mean MBI values in HCWs before and after the pandemic should be interpreted with caution.