Estimating Suicidality as an Outcome Measure in Clinical Trials of Suicide in Schizophrenia

Abstract

A Suicide Monitoring Board (SMB) evaluated putative events of suicidal behavior to determine whether they should be counted as outcome data (endpoints) in the International Suicide Prevention Trial (InterSePT) study. The InterSePT study compared clozapine and olanzapine prospectively, on reduction of suicidality in schizophrenia and schizoaffective disorder. SMB members, blinded to patient identification, evaluated packages of clinical information and the forms used to summarize the putative events and their context. The SMB members met regularly by international teleconference to discuss ratings they found problematic and to achieve consensus. Exactly 243 of the 980 patients enrolled (24.8%) were considered to have experienced 577 suicidal events. Of these events, 483 (83.7%) were confirmed by the SMB as real or valid endpoints. Before the consensus discussions, acceptable levels of agreement (K = 0.52) were reached among the SMB members regarding ratings considered as valid suicidal events; agreement increased (K = 0.64) when consensus ratings and those of on-site psychiatrists also blinded to treatment were also included. This article discusses our experience of evaluating schizophrenia patients for suicidality in general and in particular from reviewing chart information. The process of making clinical judgments regarding the seriousness of suicidal behavior in schizophrenia patients and their suicidal risk warrants further study.

Suicidal Behavior and Medication Adherence in Schizophrenic Patients

“…The evaluation of the risk factors and predicting the suicidal outcome in the schizophrenic 1 1 1 patient population is a tough task for the physician. However, knowing these possible risk factors helps the physician better evaluate and address patient beliefs and values related to the disorder [8][9][10][11].…”

“…There are several risk factors accountable for the suicidal behavior in this patient population [2][3][4][5][6]. The most common ones are white ethnicity, younger age, unmarried, higher intellectual level, social isolation, gradual onset of illness, hopelessness, co-morbid substance abuse, previous hospitalizations, previous suicide attempts, family history of suicide, family stress, impulsivity, positive symptoms of schizophrenia, unemployment, prolonged illness, the greater insight of the illness and others [7][8][9][10][11]. We conducted this review article to determine the significance of antipsychotic medication compliance in Schizophrenic patients to reduce suicidal ideation and hospitalizations.…”

Classification Algorithm for the Determination of Suicide Attempt and Suicide (CAD-SAS)

“…Agreement was somewhat higher (κ = 0.52) in a similar study by Sakinofsky, Heilä, and Krishnan (2004), where suicidology experts evaluated packages of clinical information to determine whether or not a given incident was a suicide attempt. Concordance increased (κ = 0.68) following a consensus discussion, which suggests that interrater reliability can be improved by resolving ambiguities and conflicting interpretations in more complex cases.…”

“…Differences in the range of evidence considered in the assessment can be a source of discrepant judgments (Rosenberg et al, 1989). In addition, intent may be concealed or denied to help the person manage guilt, shame, or regret that often arise after a suicide attempt (Sakinofsky et al, 2004; Wagner et al., 2002). Likewise, one may claim suicidal intent even in its absence, to bring about change in their circumstances (O’Carroll et al, 1996).…”

“…Although the lethality of the method may seem to be a more objective indication of how serious the person’s resolve to die is, this is not the case unless the individual is aware of the potential medical consequences of the act (Beck, Beck, & Kovacs, 1975). However, certain high-risk groups, including young people (Harris & Myers, 1997), patients with psychotic illness (Sakinofsky et al, 2004), and those with cluster B personality disorders (Stanley, Gameroff, Michalson, & Mann, 2001), are particularly prone to misestimating the lethality of the means. Taken altogether, the more ambiguous, inconsistent, or incomplete the facts of the case are, the greater the scope for discrepant ratings (Wagner et al, 2002).…”