Momentary Assessment.Tracking Suicidal Ideation

Suicide risk is high amongst patients in mental health services in the postdischarge period (Large \& Kapur, 2018; Chung et al 2019). However, our current knowledge and understanding does not help us in preventing suicide as an outcome. It is not possible to predict suicide in high risk populations, and- more importantly- we also are not able in any efficient way to discriminate between those who are at higher risk for suicide and those who are in less high risk of suicide in this population (Franklin et al 2017, Large, 2018).

The most limiting feature of current procedures for suicide risk assessment is their very broad based, coarse characteristics, with a notable lack of incorporated knowledge of how ongoing daily life experiences influence suicide risk. If clinical judgments on suicide risk are to reach a higher level of certainty, this can only be achieved by addressing the specific, proximal factors that maximally relevant to the patients and their situation (Nock et al, 2009).

The field of suicidology needs to determine how, when and for whom continued life suffering after discharge are accompanied by increased suicidal risk. Further important tasks are to identify the post-discharge factors that protect these patients. In this context, the aim is to increase the knowledge base about fluctuations in suicide ideation and its association with relational and contextual factors after hospital discharge in suicide attempters and patients hospitalized to psychiatric institutions due to severe suicide risk. Moreover, the researchers aim to determine "profiles" of such associations for different patient subgroups. These profiles of associations will be regressed against the presence vs absence of novel suicide risk behavior including possible self-harm and suicide episodes as a subsequent follow up point months later. Hence, the core data from the project will be seen in relation to both suicidal thoughts and subsequent suicidal behavior.

Below, four Research Questions (RQ) including specific hypotheses are described.

RQ1: Delineate fluctuations in suicide ideation

1. In the overall patient group considered together, suicide ideation (degree and endurance) will fluctuate markedly within a day and over the assessment period
2. Patient characteristics and patient subgroups can be identified, with different fluctuations in suicide ideation, including absence of suicide ideation at most time points, stable levels of suicide ideation, highly fluctuating levels, and duration of suicide ideation

RQ2: Identify factors that are associated with/ influence suicide ideation in the EMA assessment period

1. Associations with patient characteristics at baseline: Overall, suicide ideation will be at higher levels and fluctuate more profoundly, in individuals with high baseline scores on the following: i) psychological pain, stress, agitation or inner turmoil, hopelessness and self-hate, ii) high severity of their recent suicide attempt, and iii) high levels of depression and general symptom severity, in addition to (iv) in individuals who live socially isolated/ alone versus those who live together with a partner or family
2. Associations with relational and contextual measures during EMA assessment:
3. Suicide ideation will be relatively high as well as enduring in times of i) psychological pain, stress, agitation or inner turmoil, hopelessness and self-hate, ii) interpersonal conflict, iii) when they are alone or with friends and negative feelings dominate over positive feelings, and iv) when perceived control over thoughts is low ii. Suicide ideation will be absent when i) patients are in relations with other people and (ii) when they experience positive feelings

RQ3: Generate profiles for subgroups

1. Based on findings for RQ1-RQ3, different profiles of associations between suicide ideation and other EMA measures can be delineated for different subgroups of patients, and that can be reliably linked to patient characteristics at baseline
2. Explore differences in suicide ideation and related cognitions/ other EMA measures in the group with and without a suicide attempt
3. High risk profiles can be identified by linking associations between suicide ideation and other EMA measures to subsequent suicide ideation, symptom severity, self-harm, and/ or suicide attempts at 3 weeks and 3 months follow-up

RQ4: How do participants experience and evaluate their participation and completion of the EMA assessment protocol?

This is a longitudinal observational study. The researchers will adopt the method of Ecological Momentary Assessment to assess how participants cope in the postdischarge period. Participants will receive 5 surveys daily in the first 10 days after leaving the hospital. Additionally, participants will meet to three research interviews; baseline interview before discharge from the hospital, after three weeks and after three months

Measures (see section for outcome measures)

The multi-level EMA data (assessments nested within days and participants) will be analyzed in mixed effect regression models, focusing on linear changes in suicidal ideation over time and on how this is associated with independent variables. To estimate within-subject and between-subject variation, the patient identifier and time will be the random variables in the mixed effect model. In the random part of the mixed model, the residual is the estimate for within-subject variation. In prospective analyses (from T to T+1), time-lagged variables will be used as predictors. Data will be centered prior to analysis, with centering model based on the hypothesis to be tested.

Statistical power in multilevel data depends on number of both individuals and assessments.

With 5 daily assessments over 10 days in 50 participants, power is 100% to detect large effect sizes and 80% to detect medium effect sizes (as modeled by Power Curves for Multi-level Studies (shinyapps.io).

References- Chung, D., et al., Meta-analysis of suicide rates in the first week and the first month after psychiatric hospitalisation. BMJ Open, 2019. 9(3): p. e023883.

Franklin, J.C., et al., Risk factors for suicidal thoughts and behaviors: A meta-analysis of 50 years of research. Psychol Bull, 2017. 143(2): p. 187-232.

Large, M. and N. Kapur, Psychiatric hospitalisation and the risk of suicide. Br J Psychiatry, 2018. 212(5): p. 269-273.

Large, M., The role of prediction in suicide prevention. Dialogues Clin Neurosci, 2018. 20(3): p. 197-205.

Nock, M.K., M.J. Prinstein, and S.K. Sterba, Revealing the form and function of self-injurious thoughts and behaviors: A real-time ecological assessment study among adolescents and young adults. J Abnorm Psychol, 2009. 118(4): p. 816-27.