BackgroundNurses at the frontline faced high risks of the COVID-19 infection, undertook heavy workloads of patient care, and experienced tremendous stress that often led to compassion fatigue.AimThis study was to explore the role of positive psychosocial resources (i.e., perceived social support and emotional regulation efficacy) in the relationship between role stress and compassion fatigue.MethodsA cross-sectional design was conducted in Hubei Province, China between May and September 2021. The Role Stress Questionnaire, the Perceived Social Support Scale, the Emotional Regulation Efficacy Scale, and the Professional Quality of Life Scale were used to measure key variables of interest. Nurse socio-demographic data were also collected. Structural equation modeling was used to explore the relationships, including potential mediating effect, among role stress, perceived social support, emotional regulation efficacy, and compassion fatigue.ResultsA total of 542 nurses participated in this investigation, and 500 were eventually enrolled in the analysis. The incidence of compassion fatigue among nurses was 94.2%, including 65.8% of nurses reporting at least moderate compassion fatigue. Univariate analysis showed that educational level, marital status, hospital rank, sleep time were the factors affecting compassion fatigue of the nurses. The structural equation modeling revealed that: Role stress had a direct positive effect on compassion fatigue; Perceived social support and emotional regulation efficacy partially mediated the link between role stress and compassion fatigue respectively; And there was a chain mediating role of perceived social support and emotional regulation efficacy between role stress and compassion fatigue.ConclusionThe incidence of compassion fatigue was high during the COVID-19 pandemic among bedside nurses in China. Improving social support and enhancing the efficacy of emotion regulation may help alleviate compassion fatigue directly and/or via buffering the impact of role stress.