Shigella sonnei (S. sonnei) are Enterobacteria with a strictly human reservoir; they are extremely contagious, and the primary mode of transmission is person-to-person. Bloody diarrhea is a highly suggestive clinical sign, and stool culture using fresh stool samples can confirm the diagnosis. An S. sonnei outbreak occurred at a Specialized Medical-Educational Institute for children with multiple disabilities in Proisy between November 28, 1998, and March 21, 1999. The Aisne Departmental Directorate of Health and Social Affairs and the North Picardy Interregional Unit were called upon to intervene. The objective of the study was to describe the outbreak in order to identify the mode(s) and source(s) of transmission and to propose appropriate control measures. The study population included the children and staff of the Institute present between November 28, 1998, and March 21, 1999, totaling 270 people. The investigation was descriptive and retrospective. Three types of cases were defined: confirmed cases (positive stool culture with diarrhea or fever), probable cases (blood in the stool and negative stool culture), and possible cases (non-bloody diarrhea and fever, negative stool culture). The epidemic curve spanned the period from November 28, 1998, to March 21, 1999, with several successive waves. The overall attack rate was 17% (45/270); it was 33% (35/106) among children and 6% (10/164) among staff. There were 29 confirmed cases, 6 probable cases, and 10 possible cases. The cases among children involved only boarding students. In the younger children’s building, 13 children were ill, with 8 confirmed cases, 3 probable cases, and 2 possible cases, representing an attack rate of 52% (13/25). In the older children’s building, 22 children were ill, with 14 confirmed cases, 3 probable cases, and 5 possible cases, representing an attack rate of 37% (22/60). Among staff, 90% (9/10) of cases were among those responsible for changing the children (p<0.01). In each building, cases among children and staff were concentrated on a single floor. The average duration of symptoms was 8.5 days (minimum 1, median 8, maximum 25). Six children (both younger and older) were hospitalized. The reason for hospitalization was acute dehydration. One child presented with hemolytic uremic syndrome. Three antibiotics were most commonly used as first- or second-line therapy in children: the amoxicillin-clavulanic acid combination (Augmentin®, 21 times), ciprofloxacin (Ciflox®, 20 times), and cefpodoxime (Orelux®, 13 times). The use of multiple treatments was common, as after treatment with amoxicillin-clavulanic acid (Augmentin®), a second course of antibiotics was necessary in 43% of cases (9/21); after treatment with cefpodoxime (Orelux®), a secondary course of antibiotics was initiated in 46% of cases (9/13). The spread of the epidemic curve and the distribution of cases among children and staff responsible for personal care in the same buildings and floors support the theory of person-to-person transmission. Factors that may explain the duration of the epidemic include: difficulty in enforcing isolation measures, delayed reporting, and Shigella antibiotic resistance. In facilities for children with intellectual disabilities, since the risk of person-to-person Shigella transmission is very high, staff must be made aware of warning signs, and the alert should not wait for bacteriological confirmation of cases but should be based on the presence of clinical signs (sudden onset of bloody diarrhea and high fever) and a suggestive context (living in a communal setting). Control measures must be implemented promptly (isolation, handwashing before and after care and diaper changes, disinfection of premises). Finally, to improve the quality of specimens, the use of appropriate media is recommended in non-strictly hospital settings (TGV media, 7f per unit). (R.A.)