Comparison of newspaper information to official data sources for violent deaths

Objectives. The primary objective of our study was to evaluate newspapers as a surveillance tool for violent deaths. Methods. Deaths resulting from violence were identified for the National Violent Death Reporting System (NVDRS) project. The Utah articles reporting these deaths were compared with NVDRS. We adopted the NVDRS definitions and coding manual to abstract information from the articles. Results. There were a total of 852 articles covering 131 violent deaths (17% of NVDRS deaths). While homicides accounted for 7% of all violent deaths in Utah, over half of article reported violent deaths were homicides. Almost all (94%) of homicides and homicide/suicides were reported in at least one article. In contrast, only 12% of suicides and 6% of deaths the medical examiner classified as of undetermined intent were reported, even though these account for over 90% of all violent deaths. Despite the large number of articles reporting on violent deaths, few of these articles (5%) contained public health information. There were five basic types of public health information: domestic violence related, substance abuse, aging resources, gangs, and general homicide facts. Discussion. Overall our study found that there is a bias in story selection in newspapers. Homicides are reported in a disproportionate number relative to other types of death. These findings suggest if suicides were reported in newspapers in proportion to the degree in which they occur, according to the ME, the general public could have a more accurate sense of the scope these types of deaths that affect our public health.