Regression coefficient† | 95% CI | β | R square‡ | |

Constant* | 45.58 | 34.701 to 56.458 | – | 0.438 |

Working years | 2.853 | 1.189 to 4.516 | 0.202 | |

Psychological intervention or training during COVID-19 | 0.273 | 0.102 to 0.536 | 0.129 | |

Confidence about frontline work | 10.457 | 5.324 to 15.591 | 0.241 | |

Awareness of the risk of frontline work | 4.930 | 0.408 to 9.451 | 0.133 | |

Deliberate rumination | 1.323 | 0.997 to 1.649 | 0.496 |

F=31.626, p<0.001. Online supplemental appendix 4 in the supplementary material shows the assignment of the independent variables and the reference categories.

*The constant value represents the non-random part of PTGI, which is not explained by the independent variables.

†Linear regression coefficient represents the number of units the dependent variable will change when the independent variable changes per unit, holding all other variables constant. Take ‘psychological intervention or training received during COVID-19 and confidence about front-line work’ as examples, PTGI scores will increase 0.273 and 10.457, respectively, if frontline nurses answer yes to the questions. The influence order of these variables on frontline nurses’ PTGI is as follows: deliberate rumination, confidence about frontline work, working years, awareness of the risk of frontline work, psychological intervention or training during COVID-19 based on the β values.

‡R square was 0.438, and the adjusted R square was 0.425, indicating that approximately 42.5% of PTGI scores could be explained by the model.

PTGI, Post-traumatic Growth Inventory.