Assessment of Medical Claims Data for Public Health Reporting of Salmonella

BACKGROUND:   The Tennessee Department of Health (TDH) conducts active population-based surveillance for culture-confirmed cases of Salmonella as part of the Foodborne Diseases Active Surveillance Network (FoodNet). Nationally, only about 4% of the estimated 1 million Salmonella cases that occur each year are reported to public health. Administrative data from large health insurance companies captures disease-specific data which could potentially be utilized for public health surveillance. To determine the utility of using medical claims data for public health reporting, we assessed medical claims data from BlueCross BlueShield of Tennessee (BCBST) members, representing about 50% of the state population. 

METHODS:   International Classification of Diseases, 9^th Edition (ICD-9) codes were used to identify medical claims for Salmonella from BCBST data during January 2007 through December 2011. A BCBST case was defined as a medical claim with diagnosis code indicative of non-typhoidal Salmonella (ICD-9 code 003.0). BCBST cases were matched to FoodNet Salmonella cases, based on first name, last name, birthdate, sex, pathogen, and a difference of ≤60 days between claim service date and FoodNet specimen collection date. Matched cases were defined as BCBST cases reported to FoodNet. Non-matched cases were defined as BCBST cases not reported to FoodNet. Non-matched cases from 2007 to 2011 were reviewed and validated using medical records. 

RESULTS:   During the study period, 320 (67.7%) of 473 BCBST Salmonella cases matched FoodNet cases. All 320 (100%) matched cases from 2007-2011 were valid matches. Of the 153 non-matches, 38 (24.8%) were actually reported to TDH FoodNet but were mismatched due to differing demographic information in administrative claims data and state surveillance data. Medical records were available for review on 45 non-matched Salmonella cases. Of these, 15 (33.3%) had documentation of a negative stool culture and 30 (66.7%) had no record of a laboratory culture included in the medical chart. 

CONCLUSIONS:   Two-thirds of the BCBST Salmonella cases matched FoodNet cases. All culture-confirmed BCBST Salmonella cases were reported to FoodNet. Non-matched BCBST cases included clinical diagnoses which were not submitted for culture or confirmed by culture. Public health surveillance depends on stool cultures for patients with acute gastroenteritis, although this is not always indicated for clinical management. FoodNet surveillance captured all culture-confirmed cases of Salmonella among BCBST members. Use of medical claims data has the potential to assess and potentially enhance population health surveillance but needs to be evaluated on a pathogen-specific basis.