reasons, multiple centers would be optimal to enhance the science of suicide. The integration of data across laboratories can provide an ongoing picture of the key factors influencing national suicide rates such that studies of risk and protective factors can be optimized, and permit rational prevention and treatment planning. The national impact of treatment and prevention interventions shown to be effective within a network can be estimated. This will permit translation into national implementation, and with systematic cross-cultural comparisons, global extension of United States studies would become more feasible.
To obtain optimal data for the understanding, prevention, and treatment of suicide and suicidal behavior, a large population base is essential. The committee proposes a coordinated network of Population Laboratories that would allow stratified and repeated longitudinal surveys to provide more accurate data on rates of suicidal behavior, as well as long-term data on ethnographic, social, psychiatric, biological, and genetic measures necessary for increased success in prevention. Data on diagnoses associated with suicides would be obtained through the psychological autopsy method by the population laboratories for all suicides within their population, which would be enriched by highly focused ethnography. Similarly, data would be obtained on suicide attempts in the course of stratified population surveys that would be more complete than that obtained from reports generated from emergency rooms or health care providers. Thus, the population laboratory rates would correct underestimations of national rates through these registries of suicides and attempted suicides. The population laboratories would be the source of data on rates of suicidal ideation. Accurate ascertainment is essential for measurement of relative impact of risk and protective factors, and of preventive interventions.
Drawing smaller samples from these large population centers will allow the examination of risk and protective factors in far greater specificity. Multiple risk factors must be measured in the same high-risk group by multi-disciplinary groups of scientists to determine their interaction as well as their relative importance. This differs from the overwhelmingly typical approach of measuring only a few risk or protective factors in unrepresentative convenience samples. Sampling from within population laboratories allows measurement of generalizability. Deliberate sampling within ethnic and social subgroups as well as from groups with specific mental disorders can generate data applicable to at-risk groups all over the United States. In the course of obtaining data on completed suicides, the population laboratories can collect tissue samples from each indi-