Effects of misclassification biases on veteran suicide rate estimates

Rapidly rising suicide rates among current and former U.S. military personnel have generated nationwide concern. Some studies reveal that veterans have higher suicide risk than nonveterans while others found no connection between military service and suicide. These conflicting findings have been attributed to possible misclassification biases. The Blue Ribbon Panel on Suicide Prevention in the Veteran Population identified two main sources of misclassification: (1) the validity of veteran status ascertainment based on death certificate information and (2) the high prevalence of undetermined deaths among nonveterans relative to veterans. Few, if any, studies have attempted to systematically test for the effect of these misclassification biases on suicide rates in the veteran and nonveteran populations. Therefore, the aims of this study were (1) to assess the validity of veteran status obtained from the death certificate and (2) to test the effect of including undetermined deaths in the estimation of suicide rates among veterans and nonveterans. Analyses of data from the National Mortality Followback Survey, the National Violent Death Reporting System, and the American Community Survey revealed that (a) there was a 93% agreement between death certificate and proxy derived information on veteran status and (b) there were small differences between unadjusted (suicides only) and adjusted (suicide plus undetermined deaths) suicide rates for both men and women. The presentation will discuss the implication for health care providers and policy analysts who need precise epidemiological estimates of suicide rates.