Disparities in the Incidence of Salmonellosis Overall and By Leading Serogroups with Census Tract-Level Poverty, Connecticut 2000-2011

BACKGROUND:  In the US, most public health agencies do not collect data on socioeconomic status (SES) of persons affected by reportable diseases. Thus it is unknown whether distinct SES groups are differentially affected by foodborne disease. However, for diseases for which residential address is collected, it is possible to create area-based SES variables by using the census tract of case residence and its SES characteristics. As part of an initiative by the Connecticut Emerging Infections Program to describe and monitor disparities over time in incidence of diseases for which it conducts surveillance, we analyzed 12 years of surveillance data on salmonella to determine the incidence by census tract-level poverty of all salmonella combined and of the most prevalent serogroups.

METHODS:  Reported salmonella case addresses from 2000-2011 were geocoded to the census tract of residence. Census tract poverty was defined based on Public Health Disparities Geocoding Project recommendations as the percentage of residents in a census tract living below the federal poverty level in four categories: 0-<5%, 5-<10%, 10-<19% and 20+%. For cases reported 2000-2005, census tract-level poverty was determined from the US 2000 Census. For cases reported 2006-2011, it was determined from the aggregate 2006-2010 American Community Surveys. Age-adjusted salmonella incidence was determined overall and for each of the nine most prevalent serogroups using the average of the Census 2000 and 2010 denominators. The association of incidence with census tract-level poverty was measured by chisquare for trend.

RESULTS:  A total of 5484 salmonella cases were reported of which 5204 (94.9%) were geocoded (2,469 from 2000-2005, 2,735 from 2006-2010). Statistically significant associations with census tract poverty were found for salmonella overall and for three of nine serotypes. The directions of the associations, however, were not the same. For Salmonella Heidelberg, those living in the highest poverty category had 1.79 times higher incidence than those in the lowest poverty category. For salmonella overall and for Salmonella Enteriditis and Salmonella Newport, those in the lowest poverty category had incidences 1.11, 1.53 and 1.75 times higher, respectively, than those in the highest poverty category. All of these relationships were consistent over the two 6-year time periods. 

CONCLUSIONS:  In Connecticut, salmonella incidence is associated with census tract poverty with different serogroups having different relationships to poverty. To direct prevention efforts, a better understanding is needed of serogroup-specific risk factors and their association with neighborhood poverty for S. Heidelberg, S. Enteriditis and S. Newport.