Methodologic challenges and lessons learned through conducting a hospital-based Salmonella Javiana case-control study

Abstract

Determine the food(s) most likely contaminated by Salmonella Javiana associated with a salmonellosis outbreak involving 2 hospitals in North Carolina that were within 50 miles of each other in November 2021. A 2:1 matched case-control study was conducted. Food histories were obtained from hospital food orders and potential confounder covariates were collected from patient medical records. Attack rates and conditional logistic regression odds ratios (OR) were estimated at the 80% confidence interval (CI) for each food exposure and salmonellosis. There were 21 cases and 42 controls included. Fruit cups had the strongest association with salmonellosis (matched and adjusted OR=7.9 80% CI: 2.7, 23.6). Hospital-specific ORs varied for several food items, but attack rates analyses provided additional evidence that fruit cups were a likely common source. Our analyses implicated fruit cups in an outbreak of salmonellosis in 2 hospitals. Other methodologic challenges included selection of controls among sick patients, heterogeneity of food exposures, reliance on food orders rather than foods consumed, and retention of food history records. Understanding and anticipating these challenges through changes to policies and operational procedures is critical for conducting efficient and effective case-control studies in the hospital setting.