As has been noted earlier in this chapter, some suicide prevention programs, in particular universal education programs, have targeted outcome variables whose relationship to youth suicide has not been precisely identified. Evaluation of program outcomes must include careful consideration of the theoretical relevance of program goals and expected outcomes.

Theory-driven prevention strategies, programs, and treatments are most likely to inform the field in a cumulative manner. Specific variables believed to contribute to youth suicidal behavior (e.g., impaired problem-solving) need to be justified theoretically and addressed in the

prevention strategy. Following the intervention, change in the variable must be specifically measured to determine if it functions, in fact, as a mediator of suicide-related outcomes. Without a theoretical base, findings from many studies are difficult to integrate, leaving the field with an absence of information as to what actually worked, and what directions (and theoretical models) are worthy of further investigation.

Relatively few suicide prevention programs have systematically studied adequate numbers of representative at-risk youth to allow meaningful conclusions to be reached about program effectiveness, and only rarely have appropriate comparison groups been simultaneously studied. Further, most outcome studies have had access to program participants for a short period of time, which precludes attention to long-term effects of the program, including adverse effects.

Evaluation reports on the outcomes of youth suicide prevention programs have most frequently been issued by the program developers themselves. Third-party evaluation, rare in this field, may provide a more objective assessment of program accomplishments.

Few youth suicide prevention programs have had the necessary personnel or financial resources to conduct independent program evaluations. If the field is to move forward, however, mechanisms need to be established that mandate and support comprehensive, well-designed outcome studies as a regular part of prevention programming. In regard to school-based programs in particular, effective evaluation requires follow-up of students who have participated in curricular or screening activities, to determine long-term outcomes. To date, sufficient resources for such research programs have not been available.

We have noted that much more attention also needs to be given to evaluating the outcome of treatment programs for suicidal youth. Since universal and selective suicide prevention programs focus heavily on encouraging help seeking and on identifying vulnerable youth and referring them to treatment, their impact on reducing youth suicide depends ultimately on the effectiveness of the treatments that are available to such young people. Thus, the single highest priority must be given to determining the relative efficacy and effectiveness of all currently employed treatments and indicated interventions for suicidal youth.

As has been noted, randomized controlled trials (RCT) of treatments used for suicidal youth are seriously lacking. These are clearly needed to determine the impact of brief interventions with young suicide attempters presenting to emergency departments; psychotherapeutic strategies for suicidal youth and pharmacological treatments for young suicide ideators and attempters; as well as hospitalization, partial hospitalization, and posthospitalization support programs for youth. In addition to studies focusing on individual treatments, simultaneous evaluations of multiple treatment approaches (e.g., psychotherapy and medication) are especially needed. Adverse effects of treatment, including the potential of certain antidepressant medications to induce suicidality among vulnerable youth, also need to be more closely evaluated. It is also important to encourage psychiatric treatment studies to include a systematic assessment of suicidal behavior, even if this is not their primary focus.

In addition to the general evaluation concerns noted above, treatment evaluations need to give particular attention to building appropriate safeguards into treatment trials involving high-risk youth populations. Maintaining troubled youth in treatment trials over an adequate period of time to observe both short-and long-term outcomes is a particular challenge. Time-limited treatments, while easier to evaluate, leave unresolved questions about long-term effectiveness. Although control or comparison groups are essential, the inclusion of such groups necessitates ethical consideration of appropriate “control”

treatments. Few studies involving treatments for suicidality among youth have adequately defined or measured the therapeutic effects of treatment as usual.

The ultimate criteria for effectiveness in suicide prevention remain reduction in suicide attempts and suicide deaths, events for which the population base rate is low. The primary limitation of virtually all studies of the effectiveness of treatments for suicidal patients has been their relatively small size and thus their limited power to detect significant differences between or among alternative strategies (Hawton & Sinclair, 2003). Enrolling adequate numbers of appropriate participants into treatment trials can best be achieved through a number of centrally coordinated treatment research centers that can pursue common studies of treatment effectiveness. We know, for example, that antidepressant medication is effective against depression, and there is suggestive but not conclusive evidence that it reduces suicide in both adolescents and adults. Treatment research centers would make it possible to determine if this is so, to identify which medications are most effective, and to determine what degree their effectiveness is increased by combining them with various forms of psychotherapy. The formation of such centers was a primary recommendation of a recent Institute of Medicine report on suicide (Goldsmith, Pellmar, Kleinman, & Bunney, 2002).

Youth who engage in suicidal behavior vary considerably with respect to specific forms of psychopathology, substance abuse, and other psychosocial problems, and treatment trials must address this variability (Hawton & Sinclair, 2003). Particular protections must be developed to allow inclusion in such trials of suicidal youth with serious alcohol and drug problems, which confers considerable risk for subsequent attempts and suicide death.

In assessing the outcomes of treatment studies primary measures should focus on suicide-related outcomes, specifically suicidal ideation and behavior, and should also include as secondary outcomes measures of mood and social functioning. Based on their review of RCTs that have used repeated suicidal behavior as an outcome, Hawton and Sinclair (2003) have emphasized the importance of including measures of deliberate self-harm that did not lead to medical treatment. Their analysis also points to the importance of including measures of the costs of various intervention programs, and their cost-effectiveness, in assessing the impact of treatment strategies on reducing suicidality.

Finally, longitudinal studies are needed to follow up suicidal or at-risk youth through their young-adult, middle-adult, and later life years. It is clear that psychopathology can be lethal, and sustained attention to the problems evidenced by this vulnerable population is needed across the lifespan.