The serotype case-case design: a direct comparison of a novel methodology with a case-control study in a national Salmonella Enteritidis PT14b outbreak in England and Wales

Retrospective assessment of rapid outbreak investigation for gastrointestinal diseases using only cases and background exposure data

For outbreaks of gastrointestinal disease, rapid identification of the source is crucial to enable public health intervention and prevent further cases. Outbreak investigation comprises analyses of exposure information from cases and, if required, undertaking analytical epidemiological studies. Hypothesis generation has been reliant on empirical knowledge of exposures historically associated with a given pathogen. Epidemiology studies are resource-intensive and prone to bias, one of the reasons being the difficulties in recruiting appropriate controls. For this paper, the information from cases was compared against pre-defined background exposure information. As exemplars, three past outbreaks were used, one of common and two of rare exposures. Information from historical case trawling questionnaires was used to define background exposure having removed any exposures implicated with the outbreak. The case-background approach showed good sensitivity and specificity, identifying correctly all outbreak-related exposures. One additional exposure related to a retailer was identified and four food items where all cases had been exposed. In conclusion, the case-background method, a development of the case-case design, can be used to assist with hypothesis generation or when a case-control study may not be possible to carry out.

Outbreak of Salmonella enteritidis phage type 1B associated with frozen pre-cooked chicken cubes, Finland 2012

In August to October 2012, a nationwide outbreak of Salmonella enteritidis phase type (PT) 1B with 53 cases occurred in Finland. Hypothesis generating interviews pointed toward ready-to-eat chicken salad from a Finnish company and at the same time Estonian authorities informed of a S. enteritidis PT 1B outbreak linked to chicken wrap prepared at an Estonian restaurant. We found that chicken salad was associated with the infection (odds ratio (OR) 16·1, 95% confidence interval (CI) 1·7-148·7 for consumption and OR 17·5. 95% CI 4·0-76·0 for purchase). The frozen pre-cooked chicken cubes used in Finnish salad and in Estonian wraps were traced back to a production plant in China. Great Britain made two Rapid Alert Systems for Food and Feed notifications on chicken cubes imported to the UK from the same Chinese production plant. Microbiological investigation confirmed that the patient isolates in Estonia and in Finland were indistinguishable from the strains isolated from chicken cubes in Estonia and in the UK. We recommend that despite certificates for tested Salmonella, food items should be analyzed when Salmonella contamination in outbreak investigations is suspected. In outbreak investigations, electronically implemented case-case study saves time, effort, and money compared with case-control study.

Epidemiological analysis of Salmonella clusters identified by whole genome sequencing, England and Wales 2014

The unprecedented level of bacterial strain discrimination provided by whole genome sequencing (WGS) presents new challenges with respect to the utility and interpretation of the data. Whole genome sequences from 1445 isolates of Salmonella belonging to the most commonly identified serotypes in England and Wales isolated between April and August 2014 were analysed. Single linkage single nucleotide polymorphism thresholds at the 10, 5 and 0 level were explored for evidence of epidemiological links between clustered cases. Analysis of the WGS data organised 566 of the 1445 isolates into 32 clusters of five or more. A statistically significant epidemiological link was identified for 17 clusters. The clusters were associated with foreign travel (n = 8), consumption of Chinese takeaways (n = 4), chicken eaten at home (n = 2), and one each of the following; eating out, contact with another case in the home and contact with reptiles. In the same time frame, one cluster was detected using traditional outbreak detection methods. WGS can be used for the highly specific and highly sensitive detection of biologically related isolates when epidemiological links are obscured. Improvements in the collection of detailed, standardised exposure information would enhance cluster investigations.

Control selection methods in recent case-control studies conducted as part of infectious disease outbreaks

Successful investigation of national outbreaks of communicable disease relies on rapid identification of the source. Case-control methodologies are commonly used to achieve this. We assessed control selection methods used in recently published case-control studies for methodological and resource issues to determine if a standard approach could be identified. Neighbourhood controls were the most frequently used method in 53 studies of a range of different sizes, infections and settings. The most commonly used method of data collection was face to face interview. Control selection issues were identified in four areas: method of identification of controls, appropriateness of controls, ease of recruitment of controls, and resource requirements. Potential biases arising from the method of control selection were identified in half of the studies assessed. There is a need to develop new ways of selecting controls in a rapid, random and representative manner to improve the accuracy and timeliness of epidemiological investigations and maximise the effectiveness of public health interventions. Innovative methods such as prior recruitment of controls could improve timeliness and representativeness of control selection.