Enhancing Surveillance of Suicide Ideation and Suicide Attempt Through Integration of Data from Multiple Systems

In 2012, the U.S. surgeon general and National Action Alliance for Suicide Prevention released an update to the National Strategy for Suicide Prevention (U.S. Department of Health and Human Services, Office of the Surgeon General, and National Action Alliance for Suicide Prevention, 2012). The updated strategy included 13 goals and 60 objectives designed to advance suicide prevention by recommending new initiatives ranging from enhancements to communication strategies to innovations in the process for identifying those with increased risk for suicide. Providing a foundation for these renewed efforts were recommendations for enhancements to data systems supporting research and surveillance of suicide (Data Surveillance Task Force, 2014).

The most recent report on research conducted by investigators from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS) is an assessment of the incidence and characteristics of suicide ideation and attempt using information from integrated administrative and medical databases (Ursano et al., this issue). Included in this important report is a template for integrating information from multiple sources, which includes consideration of data quality and completeness, an assessment of the relationships between individual characteristics and probability of different forms of nonfatal suicide behaviors, and relationships between the rate of nonfatal events and suicide. Analyses such as those conducted by Dr. Ursano and colleagues provide a crucial example of the value of data on nonfatal suicidal behaviors and opportunities to enhance existing prevention programs.

While research has documented continued excess risk for suicide among those with a history of one or more attempts (Suominen et al., 2004), limitations associated with the availability of comprehensive data on nonfatal suicide events diminish our ability to understand the cumulative effect of repeated nonfatal behaviors and to identify opportunities for intervening among those at greatest risk. While existing data are sufficient for identifying broad groups of high-risk populations, such as those discharged from inpatient psychiatric care (Valenstein et al., 2009), recent efforts to model the distribution of risk within these groups (Kessler et al., 2015) provide evidence of an opportunity to refine understanding of risk within boundaries populations and the possibility of strategies informed by a more nuanced understanding of differences in the characteristics of risk. Such an approach may both enhance opportunities for successful intervention through identification of strategies appropriate for different subpopulations (Caine, 2013) or through the use of interventions at multiple levels consistent with a public health–oriented approach to intervention delivery (Frieden, 2010).

The development of data systems including reliable information on the incidence and characteristics of nonfatal suicide behaviors is not an easy task. Despite evidence of good quality (LeMier, Cummings, & West, 2001), questions related to the reliability of information on suicide attempt derived from external cause injury codes may have limited confidence in results studies using data uniquely obtained from this source. The characteristics of data on suicide ideation and outcomes among those with one or more reports of ideation using information from International Classification of Disease (ICD) codes have not been extensively studied. There is, however, tremendous value in better understanding the course of suicide risk, and the ability to differentiate between ideation as a more distal indicator of distress and a warning of escalating crisis has the potential to support suicide prevention by directing programs to those with different needs.

By developing a framework for categorizing suicide ideation and attempt in groups including both definite or probable events, depending on data source and reporting requirements, Dr. Ursano and colleagues have provided a crucial template for reliably integrating data on nonfatal suicide behaviors into studies of self-harm and evaluation of suicide prevention programs. Additional research is needed to compare results reported by Ursano and colleagues with results from studies using data obtained from other systems and to better understand the relationships between suicide ideation, suicide attempt, and suicide mortality.

The need for timely and comprehensive data to support surveillance of suicide and development of targeted prevention programs was noted in the 2012 National Strategy and in reference to initiatives implemented by other federal systems (Kemp & Bossarte, 2012). As noted in the 2012 National Strategy, suicide prevention programs have been limited by the availability of timely data to document changes in the rate of suicide, evaluate the impact of prevention programs, or identify emerging risk populations. The comparatively rapid availability of data from internal efforts such as the Department of Defense Suicide Event Report (DoDSER) and information available from electronic medical records provide an opportunity to both identify and care for individuals with increased risk for suicide and to enhance our ability to identify areas or groups with rapidly changing risk. As the relationships between the rate of nonfatal suicide behaviors and suicide are better understood, these systems may provide an important and timely indicator of potential changes in the rate of suicide among defined populations and serve as a measure of the impact of prevention programs.

The integration of data from multiple systems provides a unique opportunity to enhance suicide prevention programs by improving the quality of available information, expanding the amount of information available on outcomes associated with increased risk for death, and enhancing understanding of other categories of self-directed harm. While recent studies have produced invaluable information the rate of suicide in the United States has increased in recent years and there is evidence of particularly elevated risk among some high-risk groups. Additional research is needed to discriminate between those nonfatal events that represent an opportunity to intervene on increased risk of death from suicide and those for which opportunities for different forms of intervention may exist. Continued support for efforts such as those resulting from STARRS and new initiatives informed by findings of these studies are needed to achieve the goals included in the 2012 National Strategy for Suicide Prevention.