Modeling the Dynamics of Perceived Burdensomeness, Thwarted Belongingness, and Suicidal Ideation in Continuous Time

Introduction The Interpersonal Theory of Suicide (ITS) posits that perceived burdensomeness (PB) and thwarted belongingness (TB) are central to suicidal ideation (SI). The ITS proposes that either PB or TB alone are sufficient to develop passive SI; however, both are required to develop active SI. Previous studies have failed to consistently find a proximal relationship between PB, TB, and SI, possibly due to reliance on discrete time modeling. This study aims to use continuous time modeling to examine the relationship between PB and TB with passive and active SI. Methods Three independent samples of participants with recent suicidal thoughts and behaviors were recruited. In the first (N = 49) and second (N = 35) samples, participants completed 28-day and 30-day ecological momentary assessment (EMA) periods, receiving six and four daily surveys, respectively. In the third sample (N = 57), participants completed 28-days of EMA receiving either four daily surveys with one week of 10 daily surveys, or 3-15 daily surveys, varying randomly. Each EMA survey assessed PB, TB, passive SI, and active SI using two items per construct. Continuous time residual dynamic structural equation models (CT-RDSEM) were used to examine the relationship between PB and TB with passive SI and active SI separately. Dataset was introduced as a time-invariant covariate. Auto- and cross-effect drift rates were estimated to capture the influence of a variable on the instantaneous rate of change on another variable. In this framework, positive effects indicate shifts away from equilibrium, whereas negative effects indicate resistance to shifts away from equilibrium. Results There were significant negative drift rates of passive SI on itself (-1.02, 95% CI [-1.11, -0.94]) and TB on passive SI (-0.045, 95% CI [-0.095, -0.007]). PB did not show a significant drift rate on passive SI. For active SI, there were negative significant drift rates of active SI on itself (-0.97, 95% CI [-1.06, -0.84]), PB on active SI (-0.067, 95% CI [-0.15, -0.02]), and TB on active SI (-0.06, 95% CI [-0.16, -0.01]). Dataset did not significantly influence drift rates. Discussion These findings suggest that both passive and active SI exhibit significant auto-effects, reducing their own rates of change and resisting shifts from equilibrium. For passive SI, TB showed a cross-effect, while for active SI, both PB and TB demonstrated cross-effects. These cross-effects similarly work to maintain current SI levels rather than drive increases (or decreases). In contrast to the ITS, which posits that PB and TB directly relate to the development of SI, continuous time modeling results suggest that PB and TB may play a role in maintaining a person's current level of SI, rather than being a driving factor for increases in SI.