The Hazardous Unknown of Differential Disclosure: Investigating Patterns and Correlates of Disclosing Different Kinds of Suicidal Cognition

Background/Purpose: Revealing suicidal thinking is pivotal to getting help but revelation is not an all-or-nothing practice. Recognizing that different kinds of suicidal cognition (i.e., active thoughts, plan, intent) confer different levels of suicide risk raises the importance of understanding disclosure practices across cognition types. Despite that differential disclosure across cognition type can affect suicide risk assessment and management, prior work has largely neglected this topic. This study’s aims were to (1) describe disclosure prevalence of active suicidal thoughts, intent, and plan (2) examine the association between individual and relationship factors and intended disclosure of each kind of cognition.

Methods: Egocentric network and survey data were collected from 45 young adults/adults (Level 2 sample size) with a serious mental illness and a history of suicidal crisis who nominated 347 social network members (Level 1 sample size). Participants were asked about their intended disclosure of 3 suicidal cognition types (i.e., active thoughts, plan, intent) to their social network members. These outcome variables were dichotomized into “likely” or “unlikely” disclosure. Information was gathered about individual participants (e.g., psychiatric symptoms, stigma) and their social networks. Proportions were calculated to determine disclosure prevalence of suicidal cognition at the individual- (i.e., proportion of participants disclosing) and relational-level (i.e., proportion of network members as confidants). Two-level hierarchical logistic models (HLM) were used to examine associations between individual and relational factors and intended disclosure; univariable and multivariable analyses were employed for each cognition type.

Results: Intended disclosure varied by cognition type, with a higher proportion of people disclosing less serious (i.e., active thoughts) than more serious (i.e., intent/plan) types of suicidal cognition (96% vs. 82%) and a higher proportion of network members viewed as confidants for less serious than more serious cognition (43% vs. 30%). Multivariable HLM analyses revealed similar correlates for disclosure across cognition type. No individual factors were associated with disclosure.  For relationship factors, network member social support and stigma were associated with intended disclosure. For social support, the odds ratios by cognition type were as follows: active thoughts (OR=2.22; 95% CI: 1.68, 2.94); intent (OR=2.00; 95% CI: 1.43, 2.80); and plan (OR=2.08; 95% CI: 1.48, 2.93).  For stigma, the odds ratios by cognition type were as follows: active thoughts (OR=2.04; 95% CI: 1.49, 2.79); intent (OR=3.70; 95% CI: 2.14, 6.38); and plan (OR=4.07; 95% CI: 2.35, 7.05).  Having more social network members was associated with less disclosure for suicide plan (OR=0.81; 95% CI: 0.70, 0.93) but not active thoughts or intent. 

Conclusions/Implications: These findings suggest that participants adopt a more restrictive approach to disclosure for more serious types of suicidal cognition. This is concerning given that it can lead to an underestimation of risk in riskier clinical situations.  Clinicians should conduct more comprehensive assessments (e.g., multi-method approaches) of suicidal plan/intent to capture underreporting.  Programming targeting social support and stigma within networks can help to develop more confidants for disclosure. The paradoxical finding that more social network members meant less intended disclosure of suicide plan is a reminder not to equate greater network size with safety.