Abstract:Abstract: Objective To compare the effect of pump-assisted sampling and conventional flushing sampling method in the quality monitoring of endoscope disinfection. Methods From April 2016 to October 2017, the basic information of endoscope was investigated, and the method of random sampling was used to sample the flushing water of the endoscope biopsy channel to detect the total number of bacterial colonies. The propensity score matching method was used to matching the bias that may affect the results. Results 184 endoscope samples were collected, involving 59 endoscopy units, with a sample qualified rate of 82.07%. The qualified rate of the samples collected by pump-assisted method was lower than that of conventional sampling method (65.52% vs. 89.68%). The 78 pairs of 1:1 matching data obtained from the propensity score were analyzed by multivariate unconditional logistic analysis. The pump-assisted method was a risk factor (OR = 6.058, 95%CI 1.681 to 21.837). The tertiary hospital was a protective factor (OR = 0.075, 95%CI 0.020 to 0.279). Conclusion The qualified rate of reprocessing of endoscopes under investigation was low. The pump-assisted method can improve the sensitivity of sampling and monitoring. The tertiary hospitals have better implementation of reprocessing of endoscopes.