Objective To identify the key factors influencing the organizational resilience of primary medical institutions and establish a predictive framework, thereby providing insights and recommendations for enhancing emergency response capabilities and ensuring the stability and security of primary healthcare services.

Methods A survey was conducted between June and September 2023, involving 2 002 healthcare professionals from primary medical institutions across eight cities in Zhejiang province. The Boruta algorithm was employed for feature selection, and six machine learning models were used to assess the predictive power and significance of internal and external organizational factors on organizational resilience. SHAP analysis was employed to interpret feature importance, and partial dependence plots were established to reveal nonlinear relationships among variables.

Results Among the six machine learning models, the random forest algorithm demonstrated the best performance, achieving the area under the curve (AUC) of 0.954, the sensitivity of 0.889, the specificity of 0.872, the accuracy of 0.880, and the F1-score of 0.877, outperforming other models across most evaluation indicators. The top six influencing factors identified by the random forest model for organizational resilience of primary medical institutions were hospital reputation, government support, management systems, organizational commitment, production resources and operational capacity, and transformational leadership. These factors exhibited complex nonlinear relationships with organizational resilience.

Conclusions The established prediction model for organizational resilience of primary medical institutions has good prediction performance. It can help the managers and policy makers of primary medical institutions to identify the key factors affecting organizational resilience and provide a decision-making basis for promoting the improvement of organizational resilience.