To assess the association between food insecurity and depression symptom severity stratified by sex, and test for evidence of effect modification by social network characteristics.

A population-based cross-sectional study. The nine-item Household Food Insecurity Access Scale captured food insecurity. Five name generator questions elicited network ties. A sixteen-item version of the Hopkins Symptom Checklist for Depression captured depression symptom severity. Linear regression was used to estimate the association between food insecurity and depression symptom severity while adjusting for potential confounders and to test for potential network moderators.

In-home survey interviews in south-western Uganda.

All adult residents across eight rural villages; 96 % response rate (n 1669).

Severe food insecurity was associated with greater depression symptom severity (b=0·4, 95 % CI 0·3, 0·5, P<0·001 for women; b=0·3, 95 % CI 0·2, 0·4, P<0·001 for men). There was no evidence of effect modification by social network factors for women. However, for men who are highly embedded within in their village social network, and (separately) for men who have few poor contacts in their personal network, the relationship between severe food insecurity and depression symptoms was stronger than for men on the periphery of their village social network, and for men with many poor personal network contacts, respectively.

In this population-based study from rural Uganda, food insecurity was associated with mental health for both men and women. Future research is needed on networks and food insecurity-related shame in relation to depression symptoms among food-insecure men.

To investigate an outbreak of invasive disease due to Enterobacter cloacae and Serratia marcescens in a surgical intensive care unit (ICU).

Pulsed-field gel electrophoresis (PFGE) analysis of restriction fragments was used to characterize the outbreak isolate genotypes. A retrospective cohort study of surgical ICU patients was conducted to identify risk factors associated with invasive disease. Unit staffing data were analyzed to compare staffing levels during the outbreak to those prior to and following the outbreak.

An urban hospital in San Francisco, California.

During the outbreak period, December 1997 through January 1998, there were 52 patients with a minimum ICU stay of ≥72 hours. Of these, 10 patients fit our case definition of recovery of E cloacae or S marcescens from a sterile site.

PFGE analysis revealed a highly heterogeneous population of isolates. Bivariate analysis of patient-related risk factors revealed duration of central lines, respiratory colonization, being a burn patient, and the use of gentamicin or nafcillin to be significantly associated with invasive disease. Both respiratory colonization and duration of central lines remained statistically significant in a multivariate analysis. Staffing data suggested a temporal correlation between understaffing and the outbreak period.

Molecular epidemiological techniques provided a rapid means of ruling out a point source or significant cross-contamination as modes of transmission. In this setting, patient-related risk factors, such as respiratory colonization and duration of central lines, may provide a focus for heightened surveillance, infection control measures, and empirical therapy during outbreaks caused by common nosocomial pathogens. In addition, understaffing of nurses may have played a role in this outbreak, highlighting the importance of monitoring staffing levels.

To assess the efficacy of an infection control program as measured by tuberculin skin-test (TST) conversion rates in medical house staff.

Observational study.

University-based hospital in New York City serving a large indigent population.

Medical house staff.

TST conversions were measured every 6 months in medical house staff from June 1992 to June 1994. Compliance with the isolation policy was measured by identifying room locations 24 hours after admission of patients who had Mycobacterium tuberculosis recovered from respiratory specimens.

The TST conversion rate decreased from 5.8 to 0, 2.3, and 0 per 100 person years of exposure in successive 6-month periods. The estimated annual TST conversion rate among interns fell from 7 per 100 person years in June 1992 to 0 per 100 person years in June 1993 and 0 per 100 person years in June 1994 (P=.029). The proportion of patients with pulmonary tuberculosis who were isolated in negative-pressure rooms increased from 38% to 75% over the study period (P=.01).

Development of a multifaceted infection control program can decrease the risk of nosocomial tuberculosis infection in medical house staff.