Rumination and Pessimistic Certainty as Mediators of the Relation Between Lifetime Suicide Attempt History and Future Suicidal Ideation

Abstract

This study examined whether rumination and certainty about pessimistic future-event predictions (P-Certainty) would mediate the relation between lifetime suicide attempt history and future suicidal ideation. Young adults, ages 18–25 (N = 143), with a suicide attempt history (n = 32) or no previous suicide attempt history (n = 111) at baseline, were followed up 2–3 years later and completed measures of rumination, depressive symptoms, hopelessness, suicidal ideation, and pessimistic future-event certainty. Lifetime suicide attempts at baseline were associated with higher suicidal ideation at follow up, and this relation was mediated by rumination and P-Certainty. Suicide attempters may be vulnerable to later ideation due to higher levels of rumination and also certainty in their pessimistic future expectations.

Keywords:

• future-event predictions
• rumination
• suicidal ideation
• suicide attempts

Notes

¹We did not examine predictors of a suicide attempt at follow up due to the small number of individuals who actually endorsed an attempt at follow up (N = 5 out of 143; 3 participants who endorsed an attempt within the previous 2 years, and 2 individuals who did not endorse an attempt at baseline but who did so at follow up, and who were followed up more than 2 years after their initial study session). Furthermore, given that all attempters at follow up had not endorsed an attempt at baseline, there was not enough variability in the data to enable inclusion of variables such as suicide attempt history in a model predicting future attempts. However, both rumination, r(141) = .28, p < .01, and predictive certainty, r(139) = .29, p < .01, at follow up were positively associated with having made a suicide attempt during the follow-up period, suggesting that there may be a relation between these variables and risk for suicidal behavior.

Note. *p < .05, **p < .01 (1) Indicates measures at baseline. (2) Indicates measures at follow-up.

P-Certainty = Certainty about Pessimistic Future-Event Predictions; Depressive Sxs = Depressive Symptoms, measured by items 1–8 on the PHQ-9.

Note. *p < .05, **p < .01.

(1) Indicates measures at baseline.

(2) Indicates measures at follow-up.

Note. **p < .01; *p < .05.

(1) Indicates measures at baseline.

(2) Indicates measures at follow-up.

Correlations for suicide attempters (n = 32) are reported above the diagonal and for non-attempters (n = 111) are reported below the diagonal.

²Lifetime history of mental health treatment was also significantly and positively correlated with study variables, including number of previous lifetime suicide attempts, r(141) = .24, p < .01, rumination at baseline, r(141) = .29, p < .01, and follow up, r(141) = .29, p < .01, and depressive symptoms at baseline, r(141) = .35, p < .01, and follow up, r(141) = .23, p < .01. However, it was not significantly associated with predictive certainty at baseline, r(141) = .15, p = .08, or suicidal ideation at follow up, r(141) = .14, p = .09.

Note. *p < .05, **p < .01.

(1) Indicates measures at baseline.

(2) Indicates measures at follow-up.

b = unstandardized regression coefficient.

β = Standardized regression coefficient.

³We also conducted the same analyses by splitting total P-Certainty scores into predictive certainty about the presence of negative outcomes and certainty about an absence of positive outcomes, and these were included as separate mediators (along with rumination). We found that predictive certainty when anticipating negative future outcomes (95% CI = 0.04–0.97), but not certainty when anticipating an absence of positive future outcomes (95% CI = −0.02–0.32), mediated the relation between number of previous suicide attempts and future suicidal ideation. However, because there was little variability in predictive certainty about an absence of positive outcomes (M = 0.28, SD = 0.78, range = 0–5), relative to certainty about negative outcomes (M = 1.15, SD = 2.06, Range = 0–9), the present analyses used the total score for predictive certainty involving both the anticipation of negative future outcomes and an absence of positive future outcomes.