Abstract: 【Purpose/significance】The development of public health emergencies and rumors are intertwined, bringing new challenges to network security and social stability. It's necessary to investigate the time-varying fluctuation trend and the optimal enhancement pathway of rumor refutation effectiveness with the development of public health emergencies.【Method/process】Taking rumor refuta‑ tion microblogs related to COVID-19 as an example, based on "three rules of the epidemic", 10 main conditions affecting the rumor refutation effectiveness index (REI) are obtained. To reduce data redundancy, five core conditions are extracted based on the inte‑ grated algorithm, and the extracted results are validated twice by SHapley Additive exPlanations (SHAP). And then, the fuzzy-set qualitative comparative analysis (fsQCA) is used to obtain the REI enhancement pathways. Finally, the time-varying parameter vector auto-regressive model (TVP-VAR) is used to analyze the time-varying fluctuation trend of REI under different pathways.【Result/con‑ clusion】According to results, the REI enhancement pathways are obtained for three different configuration modes: blogger-influential, blogger-active and emotion-oriented. And followers' stickiness is the core condition to determine REI enhancement. The pathway of "highly authoritative and interactive information source" has the best effectiveness of time-varying fluctuation trend of REI. Improving the followers' stickiness of high-impact microbloggers can effectively improve the rumor-refuting effectiveness and promote the spread of rumor-refuting information.【Innovation/limitation】The innovation is to discuss the strategy of REI enhancement from the perspective of configuration and analyze the time-varying law of REI for the first time, which provides new ideas to improve the rumor refutation effectiveness. In the future, the study can be extended to other social platforms and the dataset can be expanded to verify the universality of the study.