Abstract: ObjectiveTo investigate causes of cases of foodborne disease and provide reference evidence for preventing the occurrence and handling similar incidents.MethodsA field epidemiological study method and a case-control study method were conducted. Case definitions were defined and case searches were performed. Suspected meals were searched according to clinical case data and meal history information, and samples of suspicious foods, cases and environment were collected for laboratory testing.ResultsA total of 20 cases were found, all of whom were employees of a company in Zhuhai City, Guangdong Province, with an attack rate of 18.7% (20/107). Clinical manifestations included abdominal pain (100%, 20/20) and diarrhea (95%, 19/20). Epidemic chart showed a point-source exposure patterns. The field epidemiological survey results showed that lunch food on September 6 was a suspicious meal (RR=8.63, 95% CI：1.04-71.53). The case-control analysis results showed that "crispy roasted duck" was a suspicious food (OR=5.11, 95% CI：1.50-17.37). The results of dose-response relationship showed that the more the intake amount of crispy roasted duck, the higher the risk of onset (?2trend=5.19, P<0.05). Dominant and isobiochemical Proteus mirabilis was detected from 3 cases and 1 preserved specimen of crispy roasted duck, with a detection rate of 10.8% (4/37).ConclusionWhen the pulsed field gel electrophoresis and other molecular typing methods were unavailable for the laboratory, the field epidemiological investigation for the differentiation of Proteus mirabilis diarrhea strains and the normal intestinal tract carrier strains still has some guiding significance. It is recommended to include pathogen monitoring of Proteus mirabilis into PulseNet China, a national molecular classification network system for foodborne disease surveillance.