53. TRAJECTORIES OF CHANGE IN SUICIDE RISK, THERAPEUTIC ALLIANCE, AND TREATMENT OUTCOMES IN PSYCHOTHERAPIES FOR LATE-LIFE DEPRESSION

Introduction

Suicide rates peak in late adulthood, with rates continuing to increase throughout the past decade. Psychotherapies for depression are effective in reducing suicide risk in late-life. The therapeutic relationship is a key predictor of reduction in suicidality during psychotherapy for adult depression. However, little is known about the association between the therapeutic relationship and late-life suicidality. This study aims to examine the relationship between the therapeutic alliance and suicidality during psychotherapy for late-life depression.

Methods

In a recently completed psychotherapy trial for late-life depression, we recruited 60 older adults between the ages of 50 and 85 (mean = 63.3, SD = 7.54) with major depression (MADRS ≥ 20). Patients received 9-weeks of psychotherapy (either Engage and Connect behavioral activation therapy or Symptom Review and Psychoeducation intervention). Depression severity and suicidal ideation (SI) were measured at weeks 0, 3, 6, and 9 using the Montgomery-Asberg Depression Rating Scale. Quality of the therapeutic relationship was measured with the self-reported Working Alliance Inventory from the patient. We categorized individuals into three SI groups: 1) persistent SI (an increase or no change in ideation throughout the study); 2) remitted SI (a decrease in ideation throughout the study); and 3) absent SI (no ideation endorsed at any point during the study). Using linear mixed-effects models, we tested whether therapeutic alliance from the patient perspective differed between the three SI groups over time.

Results

We found that throughout the study, patients with persistent and remitted suicidal ideation consistently had significantly lower therapeutic alliance scores compared to those without suicidal ideation (F[2,58] = 12.52, p < .001). We also found a significant interaction, such that the difference in therapeutic alliance between the group with persistent SI and those with no SI increased over time (F[16,422] = 2.22, p < .01). Alliance in both the remitted and absent SI groups increased throughout treatment (as of week 5, t’s > 2, p’s < .05; aside from week 7 in the remitted SI group, p = .12). In contrast, alliance decreased throughout therapy in the group with persistent SI (as of week 6, t’s > 2, p’s < .05).

Conclusions

The preliminary results indicate a relationship between persistent suicidal ideation and therapeutic alliance. Patients with persistent suicidal ideation reported lower alliance scores throughout the study. In addition, those with persistent SI showed consistent difficulty developing and sustaining a strong alliance over the course of treatment. These findings could inform personalization of psychotherapy to include alliance-building interventions with this subgroup. Further, future research can examine mediators and moderators of this relationship in psychotherapy for late-life depression.