Les infections nosocomiales sont un fleau pour les services de pediatrie, notamment en periode hivernale, avec la superposition des epidemies de bronchiolites et de gastroenterites chez les nourrissons.Materiel et methodes - Une enquete prospective a ete menee entre le 2 novembre 1999 et le 31 mars 2000 dans les deux unites accueillant des enfants de moins de trois ans. Une fiche descriptive individuelle a ete remplie. Une recherche virologique systematique a ete prescrite pour identifier le virus respiratoire syncytial et le rotavirus. Une representation geographique a ete concue afin d'identifier les secteurs a risque d'infection nosocomiale. Un appel telephonique des familles dix jours apres la sortie a ete effectue.Resultats - Durant la periode etudiee, 687 nourrissons ont ete hospitalises. Les bronchiolites et les gastroenterites representaient 458 nourrissons, l'age moyen etait de 5,4 mois et la DMS de 5,3 jours. Aucune bronchiolite nosocomiale n'a ete constatee. La prevalence des gastroenterites nosocomiales a ete de 10 % (68 cas, dont 9 identifies lors de l'appel telephonique). L'infection nosocomiale est survenue chez des enfants plus jeunes (6,6 mois vs 11,2 mois, p < 0,01). Elle a ete responsable d'une augmentation de la DMS (7,7 jours vs 4,1, p < 0,05). Parmi les bronchiolites, 16 % ont developpe une gastroenterite nosocomiale (risque relatif : 2,65, p < 0,01). L'etude geographique du service a permis d'identifier des zones a risque.Conclusion. - L'incidence des infections nosocomiales, en periode d'epidemie, nous a conduits a etablir une demarche assurance-qualite et nous permettra dans la perspective de futurs locaux d'integrer les donnees issues de la geographie.
Nosocomial infections are a preoccupation in a pediatric hospital mainly during the winter with bronchiolitis and gastroenteritis epidemics. We have examined the risk factors of nosocomial infections.Material and methods. - A prospective study was conducted between November, 1999 and March, 2000 in the infants units of the Le Havre hospital. We systematically listed the admissions and contacted the family after their discharge by phone. A geographic information system was implemented to display the epidemiological data; this software is able to illustrate the sectors at risk.Results. - During the study, 687 infants were hospitalized of whom 458 for bronchiolitis and community-acquired gastroenteritis. Mean age was 5.4 months old. No nosocomial bronchiolitis occured. Prevalence of nosocomial gastroenteritis was 10% (68 cases including nine after discharge). Infants with nosocomial infection were younger than those with community-acquired infection (6.6 months vs. 11.2 months, P < 0.01). The mean length of stay was longer in nosocomial infection (7.7 vs. 4.1 days, P < 0.05). Among the infants with bronchiolitis, 16% have developed nosocomial intestinal infections (RR = 2.65, IC: 1.59-4.4; P < 0.01). The geographic analysis pointed the area with nosocomial risk (bedroom without water, nearness of nurse office and games room).Conclusion. - Geographic information system is a part of the quality control system and may have some interaction effect on final decision making. Incidence of nosocomial infections showed the need for a prevention strategy in a pediatric hospital.