Effects of Loneliness in the Treatment of Depression

Loneliness and depression are widespread and severely debilitating health conditions. Notably, loneliness and depression are closely intertwined, with individuals suffering from depression being particularly vulnerable to loneliness. However, little is known regarding the clinical significance of loneliness in the treatment of depression and the biopsychosocial mechanisms underlying this association.

To examine the clinical relevance of the interplay between loneliness and depression, a prospective, noninterventional longitudinal design will be adopted. The study will be performed at the University Hospital and the community hospital in Erlangen, Germany. Every patient admitted to the hospital with depressive symptoms is eligible for screening according to the inclusion and exclusion criteria. Three main measurement points (T0, T1, and T2) and brief interim measurements conducted at two-week intervals (t0.1 to t0.X and t1.1-t1.6) will be utilized. After screening and clinical interviews, participants will complete the baseline measurements (T0). Thereafter, each participant will adhere to an individual period corresponding to the duration of that specific patient's inpatient stay at the clinic until the second measurement is obtained (T1). Follow-up will occur three months after T1.

Given that the effect sizes of the interaction between loneliness and depression in a clinical population are currently unknown, the sample size for this study was determined based on considerations of practicability and the population generalizability of the obtained results, as well as statistical plausibility.

To ensure the external validity of the results, it is important that both severely lonely depressive patients and less lonely depressive patients, as well as different disease trajectories and patient characteristics (e.g., age, sex, and number of comorbidities), can be identified. A sample size of approximately 200 participants is expected to ensure sufficient heterogeneity. This sample size is practically feasible due to the fact that, even when considering an inclusion rate of 50% and a dropout rate of 30% for the primary diagnosis of depression at the University Hospital and the community hospital in Erlangen, the recruitment potential clearly exceeds the target number of cases within an estimated period of two years.

Without the knowledge of which measured level of loneliness at baseline represents a strong level of loneliness and which measured level represents a weak level of loneliness in a depressed patient population, a median split of loneliness at baseline will be performed to calculate the statistical significance of the interaction between loneliness and depression in the inpatient and follow-up course (T0, T1, and T2). To detect a significant interaction effect of a mixed analysis of variance (ANOVA), when considering an α error of 0.05, a power of 0.9 (1-β error probability) and a correlation value among repeated measures of 0.5 for the groups defined as highly lonely and slightly lonely groups, the number of cases was calculated by using G*Power Version 3.1.9.7. Assuming a small effect size (f= 0.10), the desired sample size would be determined at n= 214 individuals; moreover, assuming a medium effect size (f= 0.25), the sample size would be n= 36 individuals. This indicates that there is statistical plausibility for detecting interaction effects between loneliness and depression with the target number of cases.