We used four criteria to define true bloodstream infections after isolation of coagulase-negative staphylococci (CNS) from Isolator (Wampole Laboratories, Cranbury, NJ) blood cultures: (1) the patient's temperature was > or = 38 degrees C, (2) appropriate treatment was administered, (3) the physician diagnosed bloodstream infection or criteria for nosocomial bloodstream infection were met, and (4) at least one clinical sign or laboratory value was consistent with infection. Sixty (26.4%) of 227 episodes met these four criteria. By logistic regression, variables associated with meeting the definition of infection were admission to a service other than the surgical intensive care unit, the biotype of the Staphylococcus epidermidis isolates, the log of the weighted average of the total number of bacteria per milliliter of blood in all positive cultures, resistance to at least six antimicrobial agents, and the positivity of a BACTEC blood culture specimen that was drawn with the first positive Isolator culture specimen. In a high-risk population, 26% of Isolator blood cultures positive for CNS represented infections, a rate two to four times greater than that reported in the literature. Information regarding the species, biotype, antibiogram, and number of organisms per milliliter of blood might help physicians distinguish between CNS bloodstream infections and contamination.

The authors performed a matched historic cohort study to determine the attributable mortality and risk factors for nosocomial pneumonia in bone marrow transplant (BMT) recipients. All patients with nosocomial pneumonia at a university tertiary care center were identified by a prospective surveillance system between 1980 and 1988. Control patients were selected from the population of BMT patients. The crude mortality for 55 patients with nosocomial pneumonia was 74.5% (95% confidence interval [CI95], 63% to 86%). The excess or attributable mortality was 61.8% (CI95,43.7% to 80%). Aspergillus species represented the most frequent etiologic agent in this series, causing 20 of the 55 (36%) episodes. The attributable mortality of Aspergillus species pneumonia alone was 85% (CI95, 58.6% to 100%). For death in the hospital, the risk ratio for all 55 case patients relative to control patients was 9.5 (CI95, 4.1 to 22.1). To evaluate several risk factors simultaneously, a multiple logistic regression analysis using a conditional likelihood method was performed. A mathematical model with three variables best predicted nosocomial pneumonia in our patients: the occurrence of other nosocomial infections before the diagnosis of pneumonia, allogeneic BMT, and the use of methotrexate. The presence of other nosocomial infections before the diagnosis of pneumonia remained a significant independent risk factor, with an odds ratio of 13.27 (CI95, 2.51 to 70.2) after adjustment for the use of methotrexate and allogeneic BMT. Most importantly, effective methods for preventing nosocomial pneumonias in BMT recipients will have an enormous effect on crude mortality. Cancer 1992: 69:2653‐2662.

No abstract

BackgroundPrevious research has shown that family interactions are associated with depressive symptoms in children. However, detailed classifications of family interaction types have not been studied thoroughly. This study aims to understand the types of family interactions children experience and to identify the specific types of family interactions that are associated with a higher risk of depressive symptoms in children.MethodsData used in the study was collected as part of the Child and Adolescent Behavior in Long term Evolution (CABLE) project in 2003. CABLE is a longitudinal cohort study that commenced in 2001 and collects data annually from children in Taipei city and Hsinchu county in northern Taiwan. The data analyzed in this study was that obtained from the sixth graders (aged 11 to 12 years old) in 2003. Of the 2,449 sixth graders, 51.2% were boys and 48.8% were girls. Factor analysis and cluster analysis were used to investigate the types of family interactions. One way ANOVA was used to establish the relationship between family interaction types and children's self-reports of depressive symptoms.ResultsBased on the results of factor analysis, the latent factors for family interactions included supporting activities, psychological control, parental discipline, behavioral supervision, and family conflict. After conducting cluster analysis using factor scores, four types of family interactions were revealed: supervised (29.66%), disciplined (13.56%), nurtured (40.96%) and conflict (15.82%). Children from the disciplined or conflict families were more likely to report depressive symptoms. Children from the nurtured families were least likely to report depressive symptoms.ConclusionFamily interactions can be classified into four different types, which are related to children's self-reports of depressive symptoms. The creation of a family interaction environment that is beneficial for children's mental health is an important issue for health education and health promotion professionals.