4 - (SYM 62) The Time-varying Effects of Momentary Cognitive Processes and Suicidal Ideation

Suicidal ideation (SI) and other dynamic/complex cognitive processes (e.g., rumination) have historically proven difficult to assess. Ecological momentary assessment (EMA) allows researchers to assess participant thoughts, feelings, and behaviors through repeated assessments in daily life. EMA enhances external validity, allows for characterization of dynamic processes and disentanglement of factors at the between-person (how much an individual differs from other participants) and within-person (how much the individual differs from their personal average) levels.

Recently, novel analytic methods have emerged to complement novel data collection methods (e.g., EMA). One such method that is particularly well suited to the analysis of dynamic processes is time-varying effects modeling (TVEM). TVEM is an extension of multiple linear regression that examines how associations between variables change over time in order to study dynamic processes. It accommodates uneven spacing between assessments, makes no a priori assumptions about the trajectory shape (i.e., linear, quadratic, cubic, etc.), and allows the strength of the relationship between variables to change over time.

The present study recruited adult psychiatric inpatients (n = 178) hospitalized for suicidal thoughts (n = 173) or behaviors (n = 35) or non-suicidal psychiatric reasons (n = 26). Following discharge, they completed 21 days of EMA (assessing affect, cognition, and STBs at four random times daily). We present the results from TVEM analyses examining the temporal patterns and momentary coupling of participants’ cognitive processes (rumination, emotion reactivity [ER] & distress [in]tolerance [DT]) and SI (frequency, intensity, & duration).

Within-person levels of all cognitive processes were significantly related to SI at all time points (βs ranging from 0.18 to 0.75), with the coupling remaining stable for the first two weeks and increasing slightly over the final week. Between-person levels of cognitive processes were significant for the first and final week (βs ranging from 0.07 to 0.60), but not the middle week of EMA monitoring. Results indicate that while both between- and within-person levels of rumination, ER, and DT are associated with SI, within-person levels are more robustly related—potentially reflecting greater utility as a risk factor for acute suicide risk. Theoretical and clinical implications will be discussed.