Abstract: OBJECTIVE To observe the compressive effect of the optimization of key links of etiological submission on the submission rate， submission quality， use of antibiotics， clinical treatment outcomes and medical costs in ICUs of 16 hospitals. METHODS A multi-stage stratified purposive sampling method was employed to select the intensive care units (ICUs) of 16 hospitals as the research subjects， which were divided into the control phase (January to November 2023) and the intervention phase (January to November 2024) according to the study period. The intervention comprised comprehensive strategies， including optimizing the specimen submission management mechanism， optimizing laboratory reporting workflows， enhancing testing quality， and developing information systems. The changes in key indicators were observed and compared before and after the intervention， and a mixed linear model was constructed to analyze the impact of optimizing the specimen submission process on use of antibiotics， clinical outcomes， and medical costs. RESULTS The targeted etiological submission rate prior to antimicrobial therapy in ICUs was 86.01% after the intervention， higher than 74.72% before the intervention. The pre-analytical turnaround time (TAT) for blood cultures was 0.26 hour after the intervention， shorter than 0.50 hour before the intervention; the qualified rate of sputum specimens was 87.18% after the intervention， higher than 80.82% before the intervention; the proportion of bilateral dual-set blood cultures was 94.64% after the intervention， higher than 77.33% before the intervention; the positive rate of blood culture was 23.84% after the intervention， higher than 18.75% before the intervention; the per capita cost of antimicrobial agents was 4 550 yuan after the intervention， less than 6 460 yuan before the intervention(all P＜0.05). Correlation analysis revealed that the shorter pre-analytical TAT and higher proportion of sterile specimen submissions were associated with the reduced antimicrobial use density; the higher qualified rate of sputum specimens and high proportion of bilateral dual-set blood cultures were associated with the reductions in both the per capita antimicrobial costs and total hospitalization expenses. Furthermore， the increase of the proportion of sterile specimen submissions was associated with the decrease of the mortality rate (all P＜ 0.05). CONCLUSIONS Optimizing the multi-links of etiological submission may effectively improve the quality of etiological submission from ICU patients， provide guidance for precise use of antibiotics， and reduce the medical cost. It may have positive effect on the prognosis of the patients.