Self-Directed Passive Aggressive Behaviour as an Essential Component of Depression: Findings from two observational studies

Background Self-directed passive aggressive behaviour is dened as self-harm by inactivity or omission. Based on the self-control model of depression suggesting depressive symptoms to derive from distorted self-monitoring, self-evaluation and reduced self-reward as well as increased self-punishment and reduced self-reward, a relationship between self-directed passive aggressive behaviour and depression had been assumed. First evidence for this notion derives form a recent study, demonstrating a correlation between self-directed passive aggressive behaviour and depressive symptoms. However, it remains unclear if patients with depressive disorders report more self-directed passive aggressive behaviour than patients without depressive disorders and if self-directed passive aggression mediates the associations between distorted self-monitoring and self-evaluation with depressive symptoms. Methods Study 1 compared self-directed passive aggressive behaviour levels between 220 psychotherapy outpatients with (n = 140; 67.9% female; M age = 40.0) and without (n = 80; 65.0% female; M age = 36.2) depressive disorders. Diagnoses were made based on the Structured Clinical Interview for DSM IV. Study 2 examined self-directed passive aggressive behaviour as mediator of the relationship between distorted self-monitoring and self-evaluation and depressive symptoms in 200 Psychology students. Results Compared to outpatients without depressive disorders, outpatients with depressive disorder reported signicantly more self-directed passive aggression (d = 0.51). Furthermore, Study 2 veried self-directed passive aggressive behaviour as partial mediator of the relationship between dysfunctional attitudes (ab cs = .22, 95%-CI = .14 − .31), attributional style (ab cs = .20, 95%-CI = .13 − .27), ruminative response style (ab cs = .15, 95%-CI = .09 − .21) and depressive symptoms. Conclusion Self-directed passive aggressive behaviour partially mediates the association between distorted self-monitoring and self-evaluation with depressive symptoms and might represent a core component of depressive disorders. Trial

and attributional style (Abramson, Metalsky, & Alloy, 1989; Alloy, Abramson, Metalsky, & Hartlage, 1988;Liu, Kleiman, Nestor, & Cheek, 2015) as major psychological factors. In the general population, lifetime prevalence of depressive disorders is high, ranging from 6.6% in Japan to 21.0% in France (Bromet et al., 2011). Ten to seventeen percent of individuals with depressive disorders develop a chronic course (Steinert, Hofmann, Kruse, & Leichsenring, 2014), spending approximately 20.8 percent of their lifetime in depression (Vos et al., 2004). Although cognitive behavioural psychotherapy is effective in the treatment of depression in the short-term, 54% of initial responders relapse in a 2-year period after treatment (Vittengl, Clark, Dunn, & Jarrett, 2007). Furthermore, depressive disorders are associated with heightened risk for self-directed aggression compared with the general population.
Self-directed aggressive behaviour describes any behaviour intended to harm oneself in active or passive ways (VandenBos, 2007). Self-directed active aggressive behaviour is de ned as an active engagement in self-harm (e.g., cutting oneself, self-punishment; Buss, 1961), whereas self-directed passive aggressive behaviour is de ned as harmful inactivity (e.g., omission of one´s own needs or reduced self-reward; Turp, 2007). The link between self-directed aggression and depression may be explained by the self-control model of depression (Rehm, 1977), which is based on Kanfer's (1971) behavioural self-control model.
According to the self-control model of depression, depressive symptoms are a result of a maladaptive feedback loop of dysfunctional self-monitoring and distorted self-evaluation, which leads to reduced selfreward (self-directed passive aggressive behaviour) and increased self-punishment (self-directed active aggressive behaviour).
The above mentioned cognitive factors are assumed to contribute to this feedback loop: Rumination represents a form of dysfunctional self-monitoring ( (Turp, 2007). However, a recent study found a moderate association between self-directed passive aggressive behaviour and depressive symptoms in an inpatient sample (Schanz et al., 2021). Thus, intensi ed research efforts about the role of self-directed passive aggressive behaviour in depression seem a promising avenue of identifying new prevention and treatment options for depressive disorders that are much needed considering the enormous burden of depression (Bromet et al., 2011;Steinert et al., 2014;Vittengl et al., 2007;Vos et al., 2004) Study aims Study 1 (preregistered at German Clinical Trials Register: DRKS000140051) aims to determine whether the correlation between self-directed passive aggressive behaviour and depressive symptoms found in inpatients (Schanz et al., 2021) holds-up in patients seeking outpatient psychotherapy. Furthermore, Study 1 investigates the hypothesis that depressed patients report higher levels of self-directed passive aggressive behaviour than patients with other mental disorders.
Based on the self-control model of depression, Study 2 (preregistered at German Clinical Trials Register: DRKS000190201) aims to test the assumptions that self-directed passive aggressive behaviour is associated with dysfunctional self-monitoring (rumination) and self-evaluation (dysfunctional attitudes and negative attributional style) processes and that it mediates their association with depressive symptoms. Additionally, Study 2 examines whether self-directed passive aggressive behaviour accounts for a unique amount of variance in depressive symptoms when controlling for the described cognitive factors.

Methods Of Study 1
Participants and procedure To examine the association between self-directed passive aggressive behaviour and depression in a sample of patients seeking outpatient treatment, Study 1 recruited patients from the Centre for Cognitive-Behaviour Therapy at the Saarland University and the Institute for Postgraduate Studies in Psychotherapy Saarbruecken. Adult patients (age ≥ 18 years) were asked for participation in the study after their rst consultation. The study design and all methods were approved by the Ethic Committee of Saarland University and was performed in accordance with the guidelines of the Declaration of Helsinki (General Assembly of the World Medical Association, 2014). All participants gave written informed consent. The study was approved by the local ethics committee. Diagnoses were based on the structured clinical interview for mental disorders for DSM-IV axis I (SCID-I; First, 1997). Additionally, patients completed the Beck-Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996), the Brief Symptom Inventory (BSI; place to analyze the interaction between symptom change and self-directed passive aggressive behaviour during psychotherapy. To date, the follow-up is not completed, and its results will be reported elsewhere. Initially, 251 patients agreed to participate. For 31 patients the SCID-I interview did not reveal an axis I disorder. Therefore, these patients were excluded from all subsequent analyses. For sample characteristics of the remaining 220 patients, see Table 1.  was not available at the start of the present study, diagnoses were obtained using the German version of the SCID-I for DSM IV (Wittchen, Zaudig, & Fydrich, 1997). SCID-I interviews were conducted by the rst author (clinical psychologist, master level) as well as by trained and supervised students of clinical psychology (bachelor level).
The BDI-II is a self-reporting questionnaire (Beck et al., 1996)  Statistical Analyses All analyses were performed using IBM SPSS Statistics version 25 (IBM Corp, 2017). Bivariate relationships between self-directed passive aggression (TPA-SD), depressive symptoms (BDI-II), and global symptom severity (BSI) were analysed using Pearson correlation coe cients (r). To control for the association between self-directed passive aggression with depression and general psychopathology, a multiple regression analyses including GSI score as predictor variable was conducted. An ANOVA with depression group versus control group as between-subject factor was performed to examine the hypothesis that patients with depressive disorders report more self-directed passive aggression than controls. Inclusion criteria for the depression group were unipolar affective disorders (including recurrent depression and dysthymia). To control for effects of comorbidity on self-directed passive aggression, the same analysis was rerun excluding all patients of the depression group with comorbid disorders (hereinafter referred to as 'depression only' group). Both analyses were repeated with age and gender being controlled for.

Results Of Study 1
Association between self-directed passive aggressive behaviour and depressive symptoms In line with our expectations, self-directed passive aggressive behaviour and depressive symptoms were signi cantly correlated, r = 56, p < .001. This correlation remained stable after controlling for global symptom severity (see Table 2).  (Antonovsky, 1993;Bachem & Maercker, 2016) were assessed. Findings on these constructs will be reported elsewhere. In total, 200 students completed the online survey (for descriptive statistics see Table 3).  Based on Beck's cognitive theory the DAS assesses dysfunctional attitudes using 40 items (Beck, 1963;Weissman & Beck, 1978 The BSCL is a revised version of the BSI used in Study 1 (Derogatis & Spencer, 1993;Franke, 2000Franke, , 2017. Both questionnaires differ with respect to their item order, only. In the current sample, the GSI showed excellent internal consistency (α = .96).
As in Study 1, depressive symptoms were assessed using the BDI-II and self-directed passive aggression using the TPA-SD, with good internal consistencies for the BDI-II (α = .90) and acceptable internal consistency for the TPA-SD (α = .78).

Results Of Study 2
Association between passive self-directed aggressive behaviour and depressive symptoms Self-directed passive aggressive behaviour was strongly associated with depressive symptoms, r = .54, p < .001. This association remained signi cant after controlling for global symptom severity (see Table 4). Association between self-directed passive aggressive behaviour and cognitive factors Self-directed passive aggression was positively associated with all cognitive factors of depression (see Table 5). In a joint multiple regression model, all cognitive factors accounted for a unique amount of variance in self-directed passive aggression (see Table 5). Unique association of self-directed passive aggressive behaviour with depressive symptoms Depressive symptoms were signi cantly associated with all cognitive factors (see Table 6). When accounted for the in uence of cognitive factors in a multiple regression model, self-directed passive aggression still explained an incremental proportion of variance in depressive symptoms (F(1, 195) = 15.12; ΔR 2 = .04; p < .001; see Table 6). Note

Discussion
Both studies con rmed a strong correlation between self-directed passive aggressive behaviour and depressive symptoms. In multiple analyses, this association remained robust when gender, age, general psychopathology, and cognitive factors of depression were controlled for. Furthermore, a mediation analysis showed that passive self-directed AB serves as a partial mediator for the relationship between cognitive factors and depression.

Bivariate association between depression and self-directed passive aggression
According to the self-control model of depression, insu cient self-reward (a component of self-directed passive aggression) is a major cause of the development of depressive disorders (Rehm, 1977). The results of Studies 1 and 2 validated the hypothesis of a speci c association between self-directed passive aggression and depressive symptoms. Furthermore, patients with depressive disorder had signi cantly higher scores on an inventory of self-directed passive aggression than patients with other mental disorders. In sum, Studies 1 and 2 extended the body of evidence demonstrating that not only self-directed active-AB (Bentley et al., 2015; Harford et al., 2018) but also self-directed passive aggression is robustly related to depressive symptoms.

Self-directed passive aggressive behaviour and cognitive factors of depression
Based on the self-control model of depression (Rehm, 1977) and previous studies demonstrating associations between self-directed active-aggressive behaviour and attributional style (Fox et al., 2015;Ribeiro et al., 2018), dysfunctional attitudes (Chioqueta & Stiles, 2007;Ranieri et al., 1987), as well as ruminative response style (Nicolai et al., 2016;Rogers & Joiner, 2017), we hypothesized that these factors are correlated with self-directed passive aggression. Findings of Study 2 found these associations, thereby supporting the notion that dysfunctional self-monitoring (rumination) as well as distorted selfevaluation (dysfunctional attitudes and attributional style) contribute to self-directed passive aggression. If results of future longitudinal studies establish them to be risk factors for self-directed passive aggression, they should become the focus of interventions of self-directed passive aggression. Self-directed passive aggression as mediator between cognitive factors and depressive symptoms Study 2 identi ed self-directed passive aggression as a mediator of the relationship between cognitive factors and depressive symptoms. Furthermore, the potential relevance of self-directed passive aggression for the development, onset and course of depression was supported by a unique amount of variance explained by self-directed passive aggression in depressive symptoms (even after controlling for cognitive factors of depression). Furthermore, these results also raise the question if self-directed passive aggression should be a potential target of psychotherapeutic interventions. Future studies should thus examine if treatment concepts that include speci c interventions aimed at lessening self-directed passive aggression in addition to cognitive therapy increase treatment e cacy and reduce recurrence of depression.

Limitations And Future Directions
Several limitations need to be taken into account when interpreting the ndings of Studies 1 and 2.
Study 2 supports the hypothesis of self-directed passive aggression being a (partial) mediator of the relationship between cognitive correlates of depression and depressive symptom severity. However, the sample consisted of undergraduate psychology students, with low levels of psychopathological symptoms [mean BDI-II < 14 (cut-off for minimal depression), mean BSCL < 0.68 (clinical cut-off for a mixed-gender student sample)]. Thus, the restriction of variance caused by the overall low symptom levels could have reduced external validity and generalizability of our results (Taylor & Asmundson, 2008). Therefore, a replication of this study in a clinical sample seems necessary.
Additionally, due to the cross-sectional study design the present studies cannot clarify if self-directed passive aggressive behaviour is a symptom of depressive disorders or a risk factor for its development.

Declarations
Authors Contribution Statement small empirical evidence in favour of the depression leads to aggression assumption. To examine whether self-directed passive aggression should be included in prevention and treatment strategies of depression, longitudinal studies in high-risk samples are be needed.
CGS developed the study design, prepared the materials, recruited the participants, analysed the data, and wrote the manuscript. EM, SKS and TM participated in the development of the study design, reviewed the materials, and reviewed the manuscript.
Ethics approval and consent to participate: Both studies were approved by the Ethic Committee of Saarland University and all participants gave written informed consent according to the Declaration of Helsinki and its latest revisions.
Consent for publication: Not applicable.
Availability of data and materials: The datasets generated during the current study are not publicly available due to the ongoing follow up of Study 1. The datasets supporting the conclusion of this article are available from the corresponding author on reasonable request.
Competing interests: The authors declare that they have no competing interests.
Funding: The authors declare that this research did not receive any external funding.
Authors' contributions: CGS developed the study design, prepared the materials, recruited the participants, analysed the data, and prepared the rst draft of the manuscript. EM, SKS, and TM participated in the development of the study design, reviewed the materials, and reviewed the manuscript.