Persistence and clinical consequences of post-traumatic and dissociative symptoms in people with depressive symptoms: a one-year follow-up study

ABSTRACT

Background:

Recent studies found that post-traumatic and dissociative symptoms are common in people with depressive symptoms. Although a trauma-related subtype of depression has been proposed, little is known about the persistence and clinical consequences of these symptoms.

Objective:

This one-year follow-up study investigated the persistence and clinical consequences of post-traumatic and dissociative symptoms in people with depressive symptoms.

Methods:

We analyzed longitudinal data from an international sample of people self-reporting depressive emotions (N = 152) (mean Patient Health Questionnaire-9 score = 17.27; SD = 6.31).

Results:

More than half (58.4%) of participants with baseline post-traumatic stress disorder (PTSD) still met the criteria for PTSD after one year. Participants with dissociative symptoms at baseline were significantly more likely to report lifetime psychiatric hospitalization (31.2% vs 14.7%), past-year use of psychiatric hospitalization (10.4% vs 0%) and emergency services (16.9% vs 4%) than those without dissociative symptoms. All post-traumatic and dissociative symptom clusters were cross-sectionally (r = .286 to .528, p < .001) and longitudinally (r = .181 to .462, p < .001) correlated with depressive symptoms. A sense of current threat (β = .146, p < .05) and negative self-concept (β = .173, p < .05) at baseline significantly predicted depressive symptoms after one year.

Conclusions:

These findings contribute to the increasing body of knowledge regarding the PTSD/dissociation-depression comorbidity. Given their persistence and clinical consequences, we recommend that post-traumatic and dissociative symptoms be regularly screened for in clinical settings. The existence of a possible trauma-related subtype of depression should receive more attention in both research and clinical practice.

HIGHLIGHTS

• Post-traumatic and dissociative symptoms are common in people with depressive symptoms.

• These symptoms generally persist over one year and predict more depressive symptoms at follow-up.

• Trauma-related symptoms should be regularly screened for in clinical settings.

Antecedentes: Estudios recientes encontraron que los síntomas postraumáticos y disociativos son comunes en las personas con síntomas depresivos. Aunque se ha propuesto un subtipo de depresión relacionado con el trauma, se sabe poco sobre la persistencia y las consecuencias clínicas de estos síntomas.

Objetivo: Este estudio de seguimiento de un año investigó la persistencia y las consecuencias clínicas de los síntomas postraumáticos y disociativos en personas con síntomas depresivos.

Métodos: Analizamos datos longitudinales de una muestra internacional de personas que reportaron emociones depresivas (N = 152) (Cuestionario de Salud del Paciente-9, puntuación media = 17.27; DS = 6.31).

Resultados: Más de la mitad (58.4%) de los participantes con trastorno de estrés postraumático (TEPT) al inicio seguían cumpliendo los criterios de TEPT después de un año. Los participantes con síntomas disociativos al inicio fueron significativamente más propensos a informar hospitalización psiquiátrica en algún momento de sus vidas (31.2% frente a 14.7%), uso de hospitalización psiquiátrica durante el último año (10.4% frente a 0%) y de servicios de emergencia (16.9% frente a 4%) que aquellos sin síntomas disociativos. Todos los grupos de síntomas postraumáticos y disociativos se correlacionaron transversalmente (r = .286 a .528, p < .001) y longitudinalmente (r = .181 a .462, p < .001) con síntomas depresivos. Una sensación de amenaza actual (β = .146, p < .05) y un autoconcepto negativo (β = .173, p < .05) al inicio predijeron significativamente los síntomas depresivos después de un año.

Conclusiones: Estos hallazgos contribuyen al creciente cuerpo de conocimiento sobre la comorbilidad TEPT/disociación-depresión. Dada su persistencia y consecuencias clínicas, recomendamos que los síntomas postraumáticos y disociativos se evalúen regularmente en entornos clínicos. La existencia de un posible subtipo de depresión relacionado con el trauma debería recibir más atención tanto en la investigación como en la práctica clínica.

背景：最近的研究发现，创伤后和解离症状在有抑郁症状患者中很常见。尽管已经提出了一种与创伤相关的抑郁亚型，但人们对这些症状的持续性和临床后果知之甚少。

目的：这项为期一年的随访研究调查了有抑郁症状患者创伤后和解离症状的持续性和临床后果。

方法：我们分析了来自自我报告抑郁情绪的国际样本的纵向数据 (N = 152)（患者健康问卷 9 平均得分 = 17.27；SD = 6.31）。

结果：超过一半 (58.4%) 患有基线创伤后应激障碍 (PTSD) 的参与者一年后仍符合 PTSD 标准。基线时出现解离症状的参与者与没有的参与者相比，更有可能报告终身精神科住院治疗（31.2% vs 14.7%）、过去一年使用精神科住院治疗（10.4% vs 0%）和紧急服务（16.9% vs 4%）。 所有创伤后和解离症状簇在横断面（r = .286 至 .528，p < .001）和纵向（r = .181 至 .462，p < .001）上均与抑郁症状相关。基线时的当前威胁感 (β = .146，p < .05) 和消极自我概念 (β = .173，p < .05) 可显著预测一年后的抑郁症状。

结论：这些发现有助于增加有关 PTSD/解离抑郁共病的知识体系。鉴于其持续性和临床后果，我们建议在临床环境中定期筛查创伤后和解离症状。可能存在的创伤相关抑郁症亚型应该在研究和临床实践中受到更多关注。

KEYWORDS:

• Depression
• trauma
• complex post-traumatic stress disorder (complex PTSD)
• dissociation
• early intervention

PALABRAS CLAVE:

• Depresión
• trauma
• trastorno de estrés postraumático complejo (TEPT complejo)
• disociación
• intervención temprana

关键词:

• 抑郁；创伤；复杂性创伤后应激障碍（复杂性PTSD）；解离；早期干预

1. Introduction

As a major public health problem, depression has been the largest factor contributing to global disability in the world (Ferrari et al., 2013; Smith, 2014). According to the World Health Organization (2017), approximately 4.4% of the global population is estimated to be suffering from a depressive disorder. The severity of depression wax and wane, with around 21–50% of individuals who achieved full remission from their depressive episode relapsed within one year (for a review, see Richards, 2011). Despite advances in treatment, a considerable number of people with depression do not respond well to standard medications and psychotherapy for depression (Gaynes et al., 2020; Walsh et al., 2002). A recent review of meta-analyses has shown that current psychotherapies and pharmacotherapies have limited effectiveness in the treatment of depression (Leichsenring et al., 2022).

Given the current challenges in the prevention and treatment of depression, personalized treatment, which considers individual characteristics and needs, is recommended to improve the care provided for people with depression (Maj et al., 2020; Winokur, 1997). It has been suggested that people with depression is a heterogeneous population (Goldberg, 2011) and there are different subtypes of depression, and each of them may have different etiologies and intervention needs (Fung, Chien, Lam, et al., 2022b; Rantala et al., 2018). In particular, a possible subtype of depression is closely related to trauma and stress (Rantala et al., 2018). In other words, it is reasonable to believe that a considerable subgroup of people with depression may develop these symptoms because of trauma and stress. Previous studies also revealed that people with depression who suffer from co-occurring trauma-related symptoms may have poor clinical outcomes (Fung, Chien, Lam, et al., 2022a; Şar et al., 2013), although there is a lack of longitudinal studies.

In fact, there is strong evidence showing that traumatic and stressful events, particularly when experienced during childhood, could increase the risk of developing depressive symptoms (Chapman et al., 2004; Maniglio, 2010; Vibhakar et al., 2019). Trauma could lead to a variety of biological and psychosocial problems, such as maladaptive cognitive processes, dysregulated stress response systems, interpersonal stress, problematic lifestyle or coping strategies, and heightened levels of inflammation and cortisol, which could in turn lead to depressive symptoms (Fung, Lam, et al., 2022; Kuzminskaite et al., 2021; Liu, 2017; Rantala et al., 2018). Trauma-and stress-related psychological problems, such as post-traumatic stress disorder (PTSD) symptoms, may mediate the relationship between stressful experiences and depressive symptoms (Fung, Chien, Ling, et al., 2022). The comorbidity of depression and PTSD is also well-documented in the clinical literature – a meta-analysis indicated that 52% of people with depression may suffer from comorbid PTSD (Rytwinski et al., 2013). A recent study in a sample of people with depressive symptoms has even found that up to 62.7% and 60% of participants suffered from co-occurring PTSD/complex PTSD and dissociative symptoms, respectively (Fung, Chien, Lam, et al., 2022a, 2022b). Moreover, dissociative symptoms, which involve disruption or discontinuity of normally integrated psychophysiological experiences (e.g. memories, emotions, identities) (American Psychiatric Association, 2013), often co-occur with depressive symptoms. Both PTSD and dissociative symptoms are theoretically associated with trauma and extreme stress, and are generally regarded as trauma-related symptoms that can be observed across cultures (Dalenberg et al., 2012; Fung, Chien, et al., 2023; Ross et al., 2008). Conceptually speaking, post-traumatic and dissociative symptoms are the core features of traumatization (Van der Hart et al., 2006). Over 80% of patients with a complex dissociative disorder experienced major depressive episodes (Ross et al., 1990; Şar et al., 1996). Using a cross-sectional design, dissociative symptoms are found to co-occur with depressive symptoms (Fung et al., 2020). A recent study has shown that dissociative symptoms are longitudinally related to self-harm among Japanese adolescents (Tanaka et al., 2023). Therefore, it is believed that post-traumatic/dissociative symptoms could be a contributing factor for depressive symptoms in some, if not many, cases. The hypothesis is also consistent with the demoralization model, which proposes that trauma-related symptoms may contribute to the development and maintenance of depressive symptoms (Schindel-Allon et al., 2010).

While trauma-related symptoms may contribute to depressive symptoms, the possibility that depressive symptoms may also lead to trauma-related symptoms should also be recognized. In a few studies, depressive symptoms were found to predict post-traumatic symptoms over time (Cheng et al., 2018). The reciprocal relationship between depressive and post-traumatic symptoms may explain the common co-occurrence of depressive and post-traumatic symptoms. Some scholars also speculate the existence of a depressive subtype of PTSD (Deen et al., 2022; Flory & Yehuda, 2022), although the clinical utility of the nosology is yet to be established. Therefore, we do not exclude the possibility of a depressive subtype of PTSD.

Despite the possibility that depressive symptoms might also lead to trauma-related symptoms, the present study focused on the potential role of trauma-related symptoms in the context of depression. A better understanding of the persistence and potential impacts of trauma-related symptoms in people with depression is important because there would be important implications for clinical practice, as will be further discussed below. Briefly speaking, such insights could call for a trauma-informed perspective when investigating and managing depression. More importantly, previous studies showed that trauma-focused interventions could be effective in reducing depressive symptoms (Dominguez et al., 2021) but depression-focused interventions are not effective in reducing trauma-related symptoms (Rosen et al., 2020). Therefore, with recognition of the role of trauma-related symptoms in people with depression, we can promote the timely identification and management of these symptoms, thereby improving the care provided for people with depression who also suffer from trauma-related symptoms.

Although understanding the role of trauma-related symptoms in the context of depression for clinical assessment and management could have significant implications for clinical assessment and treatment, as discussed above, further research is necessary since there are some major knowledge gaps in this regard.

1. To what extent can trauma-related symptoms predict depressive symptoms in individuals with depression? Although there is an increasing number of studies investigating the longitudinal relationship between post-traumatic and depressive symptoms, the findings are not consistent across studies. Some studies found that post-traumatic symptoms could predict depressive symptoms over time (e.g. An et al., 2019; Cheng et al., 2020; Fung, Hung, et al., 2023), but other studies did not support this hypothesis (e.g. Schindel-Allon et al., 2010). More importantly, most of these studies employed a specific trauma-exposed sample, which did not necessarily cover a range of severity of depressive symptoms. In addition, little is known about the long-term effects of dissociative symptoms on depressive symptoms, as cross-sectional designs were generally used in previous studies on this topic (Fung & Chan, 2019; Şar et al., 2013). Therefore, our understanding of the potential consequences of specific post-traumatic/dissociative symptom clusters in individuals with depressive symptoms is limited.

2. To what extent do trauma-related symptoms persist in people with depression? There is currently no available data on the persistence of post-traumatic symptoms and dissociation in people who suffer from depressive symptoms. Longitudinal studies revealed that some people with PTSD would spontaneously recover over time (Hiller et al., 2016; Smith et al., 2019). Nevertheless, it is currently unclear to what degree people with depressive symptoms may experience spontaneous reduction in PTSD symptoms over time. The same is true for dissociative symptoms in people with depression. A recent general population study showed that dissociative experiences were stable over a period of four years (Maraldi, 2022). In addition, dissociative disorders are typically chronic conditions involving substantial medical costs (Gonzalez Vazquez et al., 2017) and requiring long-term mental health care (Brand et al., 2012; Myrick, Webermann, Loewenstein, et al., 2017). Having said that, some studies found that dissociative symptoms may also change, or even decrease over time in the clinical populations (Schäfer et al., 2012). Dissociative symptoms may also be acute conditions in some situations, especially when there are influences from drugs or sleep conditions (Giesbrecht et al., 2013; van Heugten-Van der Kloet et al., 2015). Therefore, more research is required to understand the stability of post-traumatic and dissociative symptoms among people with depressive experiences. Such findings would be important for understanding the importance of early identification and management of trauma-related symptoms in people with depression.

3. Would trauma-related symptoms be associated with more needs for psychiatric services? Previous studies have shown that trauma-related symptoms such as dissociation are associated with more impairments (e.g. lower levels of social-occupational participation and higher rates of psychiatric service usages)(e.g. Fung, Wong, Lam, Chien, Hung, & Ross, 2022; Gonzalez Vazquez et al., 2017). However, less is known about the relationship between trauma-related symptoms and the needs for psychiatric services in people with depression.

Keeping these unaddressed research questions in mind, we conducted a one-year follow-up investigation on post-traumatic and dissociative symptoms in a sample of people who self-reported having experienced depressive emotions. In particular, there were three major research questions corresponding to the above-mentioned knowledge gaps:

First, we examined to what extent post-traumatic and dissociative symptoms would be persistent in this sample.

Second we examined the predictive value of these symptoms on subsequent depressive symptoms.

Third, informed by previous studies showing that post-traumatic and dissociative symptoms/disorders are associated with more impairments and/or healthcare costs in both psychiatric and nonpsychiatric samples (as these symptoms are less likely to respond well to standard, non-trauma-specific treatments) (e.g. Fung, Wong, Lam, Chien, Hung, Ross, et al., 2022; Gonzalez Vazquez et al., 2017; Myrick, Webermann, Langeland, et al., 2017; Ross & Dua, 1993), we also examined whether participants with PTSD or elevated levels of dissociation would be more likely to report use of psychiatric hospitalization and psychiatric emergency services.

2. Methods

2.1. Participants

This study analyzed data from an international survey project which was approved by the institutional review board at the Chinese University of Hong Kong. In 2021/2022, this project recruited English-speaking participants who self-reported having experienced depressive emotions. The inclusion criteria, which were checked at baseline, included the following: 1) was aged 18 or above; 2) can read and write English; 3) had access to the Internet; 4) agreed to provide informed consent and participate; and 5) endorsed the item ‘Have you ever suffered from any depressive emotions?’. The exclusion criteria included: 1) immediate need for professional help; 2) had a diagnosed reading disorder, dementia or intellectual disabilities; and 3) recurrent suicidal ideation, suicidal attempts or homicidal plans in the past 2 weeks. We recruited potential participants through social media platforms, and 468 participants gave online written informed consent and completed the online survey, but only 301 participants provided a valid email for a follow-up survey. Data collected at Time 1 had been reported elsewhere (Fung, Chien, Lam, et al., 2022b). A total of 152 participants completed a follow-up survey after one year (i.e. retention rate = 50.50%) (average number of days = 402.4; SD = 16.24).

Most participants were female (94.1%). Their age ranged from 18 to 62 (M = 26.18; SD = 8.78). Only few participants were married (15.8%) and were full-time employed (32.2%) as reported at baseline. About half of them had an undergraduate degree (46.7%). This was a regionally diverse sample as participants came from 13 different countries/regions, including: 25% from Canada, 19.7% from Singapore, 15.8% from the United States, and 15.1% from the United Kingdom. A clinical diagnosis of major depressive disorder was self-reported by 38.2% of participants; 23.7% of participants reported a clinical diagnosis of PTSD or complex PTSD; 5.3% reported a clinical diagnosis of any dissociative disorder.

At Time 1, participants reported high levels of depressive symptoms as measured with the Patient Health Questionnaire (PHQ-9) (M = 17.27; SD = 6.31) (85.5% scored 10 or above on PHQ-9, which indicated moderate to severe level of depressive symptoms; 42.1% scored 20 or above on the PHQ-9, which indicated severe depressive symptoms); 56.6% had seen a psychiatrist before. Most participants (94.08%) reported at least one traumatic event during lifetime on the Brief Betrayal Trauma Survey (BBTS) (Goldberg & Freyd, 2006); 50.7% met the ICD-11 criteria for PTSD on the International Trauma Questionnaire (ITQ) (Cloitre et al., 2018). In addition, 50.7% of participants scored ≥ 67 on the Multiscale Dissociation Inventory (MDI), which indicated clinically significant (or ‘elevated’) (versus normal) levels of dissociative symptoms (Briere, 2002; Briere et al., 2005; Mitchell, 2006). The sample characteristics are reported in Table 1.

Table 1. Descriptive analysis, and cross-sectional and longitudinal correlates of depressive symptoms (N = 152).

	Frequency	Correlations
Variables	Percentage / Mean (SD)	T1 Depressive symptoms	T2 Depressive symptoms
Age	26.18 (8.78)	−.190*	−.076
Gender (Female)	94.1%	.104	.049
T1 Number of traumatic experiences	5.61 (3.87)	.145	.186*
T1 Re-experiencing	69.7%	.301***	.181*
T1 Avoidance	82.2%	.387***	.246**
T1 Sense of current threat	80.9%	.335***	.330***
T1 Affective dysregulation	92.8%	.484***	.246**
T1 Negative self-concept	88.8%	.513***	.462***
T1 Disturbances in relationships	92.8%	.416***	.351***
T1 Dissociation	50.7%	.448***	.248**
T2 Re-experiencing	71.1%	.223**	.343***
T2 Avoidance	86.8%	.283***	.286***
T2 Sense of current threat	75.7%	.200*	.416***
T2 Affective dysregulation	90.8%	.440***	.448***
T2 Negative self-concept	77.0%	.448***	.528***
T2 Disturbances in relationships	88.2%	.346***	.421***
T2 Dissociation	40.8%	.435***	.511***

Notes: * p < .05 ** p < .01 *** p < .001. T1 = Time 1; T2 = Time 2.

All post-traumatic and dissociative symptoms were categorical variables. For each post-traumatic symptom cluster, which was assessed using the International Trauma Questionnaire (ITQ), a symptom was present if a score of 2 (‘moderately’) or higher was selected for at least one item within the symptom cluster. For dissociation, an elevated level of overall psychoform dissociative symptoms was present if the total Multiscale Dissociation Inventory (MDI) score was 67 or above. Depressive symptoms were continuous variables, which were assessed using the Patient Health Questionnaire-9 (PHQ-9).

Independent sample t and chi-square tests revealed no significant differences in any major variable at Time 1 between participants who completed the follow-up survey and those who did not, except for PTSD symptoms. Participants who completed the follow-up survey reported fewer classical PTSD symptoms on the ITQ than those who did not (M = 13.30; SD = 5.98 vs M = 14.59; SD = 6.20), t(467) = 2.124, p = .034.

2.2. Measures

In addition to demographic and health backgrounds (e.g. age, gender, location, use of psychiatric services), the online surveys included measures of various mental health symptoms. The present study focused on the following variables:

Depressive symptoms were measured using the Patient Health Questionnaire-9 (PHQ-9) (Kroenke & Spitzer, 2002), which is a 9-item self-report measure of depression with good reliability and validity (Kroenke et al., 2001; Kung et al., 2013). The total score of the PHQ-9 may range from 0 to 27. The PHQ-9 was internally consistent in our sample at baseline (α = .873).

Trauma exposure was assessed at baseline using the Brief Betrayal Trauma Survey (BBTS), which is a 24-item self-report questionnaire of 12 different types of traumatic events during childhood and adulthood (Goldberg & Freyd, 2006). The traumatic events asked on the BBTS can be divided into betrayal and non-betrayal trauma. Betrayal trauma refers to trauma perpetrated by a close person, such as a family member. A sample item for betrayal trauma is: ‘You were deliberately attacked that severely by someone with whom you were very close.’ A sample item for non-betrayal trauma is: ‘You were deliberately attacked that severely by someone with whom you were not close.’ The BBTS was reported to have good reliability in previous studies (Fung, Chien, Ling, et al., 2022; Goldberg & Freyd, 2006).

PTSD symptoms were measured using the International Trauma Questionnaire (ITQ) (Cloitre et al., 2018), which is a 18-item self-report measure of PTSD and complex PTSD (CPTSD) symptoms according to the ICD-11 framework. On the ITQ, both classical PTSD symptoms (i.e. re-experiencing, avoidance, a sense of current threat) and ‘disturbances in self-organization’ (DSO) symptoms (i.e. affect dysregulation, negative self-concept and disturbances in relationships) are assessed (0 = Not at all; 4 = Extremely). A symptom is considered to be endorsed if the individual selects a score of 2 (‘moderately’) or higher for at least one item within each of the six symptom clusters. The ITQ was found to have good to excellent reliability and validity (Cloitre et al., 2021). Moreover, the ITQ can be used to make a provisional PTSD/CPTSD diagnosis. Both the PTSD and the DSO subscales were internally consistent in our sample at baseline (α = .824 and .825, respectively). In the present study, a provisional PTSD/CPTSD diagnosis was made only when the participant reported having at least one lifetime traumatic event on the BBTS (i.e. N = 143). However, when calculating the levels of PTSD/DSO symptoms and counting whether a specific PTSD/DSO symptom cluster was present, all participants (with or without any BBTS traumatic event) were included for analysis (i.e. N = 152). This decision was made considering the fact that some individuals may present with PTSD symptoms even if they have not experienced any Criterion A stressor (Hyland et al., 2021).

Dissociative symptoms were measured using the Multiscale Dissociation Inventory (MDI), which is a 30-item self-report measure that has been validated and normed (Briere, 2002; Briere et al., 2005). The MDI measures six moderately intercorrelated dissociative symptom clusters (e.g. disengagement, depersonalization, identity dissociation) (1 = Never; 5 = Very often). As the MDI is a normed assessment tool, the raw scores can be converted to T-scores following the MDI Manual (Briere, 2002). Therefore, it is possible to determine ‘clinically significant’ (or ‘elevated’) (versus normal) levels of dissociative symptoms; a raw total score of 67 or above on the MDI indicates an elevated level of overall psychoform dissociative symptoms (Briere, 2006; Mitchell, 2006). Therefore, participants were defined as exhibiting a high level of dissociative symptoms if they scored 67 or above on the MDI. The MDI had excellent internal consistency in our sample at baseline (α = .959).

Psychiatric service usages were also measured at both timepoints. At baseline, we asked about use of psychiatric hospitalization and psychiatric emergency services during lifetime: ‘Have you ever been hospitalized because of emotional or mental health problems?’ and ‘Have you ever received emergency medical services because of emotional or mental health problems?’ At follow-up, we asked about use of psychiatric hospitalization and psychiatric emergency services in the past 12 months: ‘Have you ever been hospitalized because of emotional or mental health problems in the past 12 months?’ and ‘Have you ever received emergency medical services because of emotional or mental health problems in the past 12 months?’

2.3. Data analysis

Statistical analyses were conducted using SPSS 22.0. For the first research question, we first conducted descriptive analysis to report the persistence of PTSD and dissociative symptoms – in particular, we reported the rates of people meeting criteria for PTSD on the ITQ and scoring 67 or above on the MDI at each timepoint. We then conducted Pearson and point-biserial correlations to examine the cross-sectional and longitudinal relationships between trauma-related symptoms and depressive symptoms. For the second research question, we conducted a hierarchical multiple regression to examine whether the endorsement/presence of PTSD and dissociative symptom clusters (as categorical variables) at T1 would predict depressive symptoms (as a continuous variable) at T2, after controlling for depressive symptoms (as a continuous variable) at T1. An additional exploratory analysis was also conducted using independent sample t test to examine whether ‘persistors’ (i.e. meeting PTSD criteria or scoring 67 or above on the MDI at both timepoints) and ‘non-persistors’ (i.e. meeting PTSD criteria or scoring 67 or above on the MDI at baseline, but remitted at follow up) would have significant differences in the levels of trauma histories (i.e. the BBTS subscores).  Finally, for the third research question, we examined the differences in psychiatric service usage (as categorical variables) between participants with and without PTSD or an elevated level of dissociative symptoms using chi-square tests.

3. Results

3.1. PTSD and dissociative symptoms persisted over time

As shown in Table 1, all PTSD symptom clusters and dissociative symptoms were commonly endorsed by the participants at both timepoints. At T1, affective dysregulation (92.8%) and disturbances in relationships (92.8%) were the most commonly endorsed PTSD symptoms, while 50.7% of participants exhibited an elevated level of dissociative symptoms.

In addition, among participants with PTSD at T1, 58.4% still met the criteria for PTSD at T2; among participants with an elevated level of dissociative symptoms at T1, 64.9% still scored ≥ 67 on the MDI at T2 (see Table 3). These figures indicated that PTSD and dissociative symptoms were persistent over time in our sample.

Further analyses using independent sample t test showed that participants whose PTSD persisted (n = 45) reported more non-betrayal traumatic events on the BBTS at baseline (M = 2.93; SD = 2.29 vs M = 1.50; SD = 1.46) than those who remitted (n = 32), t(74) = 3.350, p = .001. Moreover, participants whose elevated MDI scores (i.e. ≥ 67) persisted (n = 50) reported more betrayal traumatic events on the BBTS at baseline (M = 3.82; SD = 2.08 vs M = 2.63; SD = 2.45) than those who remitted (n = 27), t(75) = 2.251, p = .027.

3.2. PTSD and dissociative symptoms predicted depressive symptoms

Correlation analyses showed that the presence of all PTSD and dissociative symptoms was cross-sectionally (r = .286 to .528, p < .001) and longitudinally (r = .181 to .462, p < .001) correlated with depressive symptoms (Table 1). In particular, among PTSD and dissociative symptoms, negative self-concept (r = .513, p < .001) and affective dysregulation (r = .484, p < .001) had the strongest cross-sectional correlations with depressive symptoms at T1. Moreover, negative self-concept (r = .462, p < .001) and disturbances in relationships (r = .351, p < .001) at T1 had the strongest longitudinal correlations with depressive symptoms at T2. The cross-sectional and longitudinal correlations of depressive symptoms with PTSD and dissociative symptoms are presented in Table 1.

The findings of the hierarchical multiple regression analysis predicting changes in depressive symptoms by PTSD and dissociative symptoms at T1 are presented in Table 2. A hierarchical multiple regression analysis further indicated that the presence of a sense of current threat (β = .146, p = .034) and negative self-concept (β = .173, p = .025) at T1 was significantly associated with T2 depressive symptoms, even after controlling for depressive symptoms at T1 (ΔR² = .054, ΔF = 2.074, p = .050).

Table 2. Hierarchical multiple regression predicting depressive symptoms at Time 2 (T2) (N = 152).

	Model 1		Model 2
	β	p	β	p
Constant		.017		.600
T1 Depressive symptoms	.640	.000	.568	.000
T1 Re-experiencing			−.034	.630
T1 Avoidance			−.029	.686
T1 Sense of current threat			.146	.034
T1 Affective dysregulation			−.106	.160
T1 Negative self-concept			.173	.025
T1 Disturbances in relationships			.065	.389
T1 Dissociation			−.043	.541
Adjusted R^2	.405		.434
F	103.846***		15.446***
ΔR^2	.409		.054
ΔF	103.846		2.074
p	< .001		.050

Notes: * p < .05 ** p < .01 *** p < .001.

Independent variables in Model 2 were categorical variables, while depressive symptoms were continuous variables..

3.3. Participants with high levels of dissociation reported greater use of psychiatric services

Chi-square tests showed that participants with an elevated level of dissociative symptoms at T1 were significantly more likely to report lifetime psychiatric hospitalization (p = .016) at T1, as well as past-year use of psychiatric hospitalization (p = .004) and emergency services (p = .010) at T2 (see Table 3). Nevertheless, participants with and without PTSD at T1 did not differ in these variables.

Table 3. Clinical differences between participants with and without PTSD/dissociation at Time 1 (T1)

	Participants without PTSD at T1 (n = 75)	Participants with PTSD at T1 (n = 77)
Variables	%	%	χ2 (df = 1)	p
Lifetime psychiatric hospitalization reported at T1	21.3%	24.7%	0.239	.625
Lifetime psychiatric emergency services reported at T1	26.7%	23.5%	0.613	.433
Past-year psychiatric hospitalization reported at T2	4.0%	6.5%	0.474	.491
Past-year psychiatric emergency services reported at T2	8.0%	13.0%	1.003	.317
Meeting criteria for PTSD at T2	36.0%	58.4%	7.675	.006
	Participants with MDI < 67 at T1 (n = 75)	Participants with MDI ≥ 67 at T1 (n = 77)
Variables	%	%	χ2 (df = 1)	p
Lifetime psychiatric hospitalization reported at T1	14.7%	31.2%	5.837	.016
Lifetime psychiatric emergency services reported at T1	25.3%	33.8%	1.296	.255
Past-year psychiatric hospitalization reported at T2	0.0%	10.4%	8.225	.004
Past-year psychiatric emergency services reported at T2	4.0%	16.9%	6.695	.010
MDI ≥ 67 at T2	16.0%	64.9%	37.671	.000

Note: MDI = The Multiscale Dissociation Inventory.

4. Discussion

This study contributes to the limited literature on the persistence and clinical consequences of trauma-related symptoms in people self-reporting depressive emotions. The major findings included the following: 1) all post-traumatic symptom clusters as well as dissociative symptoms were cross-sectionally (r = .286 to .528, p < .001) and longitudinally (r = .181 to .462, p < .05) correlated with depressive symptoms in our sample of participants with depressed symptoms; 2) baseline presence of a sense of current threat (β = .146, p = .034) and negative self-concept (β = .173, p = .025) significantly predicted depressive symptoms after one year, even after controlling for baseline depressive symptoms; 3) even after one year, more than a half of participants with baseline PTSD (58.4%) still met the criteria for PTSD; and, 4) participants with elevated dissociative symptoms at baseline were significantly more likely to report lifetime psychiatric hospitalization (31.2% vs 14.7%), as well as past-year use of psychiatric hospitalization (10.4% vs 0%) and emergency services (16.9% vs 4%) than those without elevated dissociative symptoms at baseline. The findings and their theoretical and clinical implications require further discussion.

First of all, the findings supported the hypothesis that post-traumatic symptoms could be a contributing factor for depressive symptoms, thus pointing to the potentially important role of trauma-related symptoms in the development or maintenance of depression. Although the heterogeneous nature of people with depression is often well recognized (Fung & Chan, 2019; Goldberg, 2011), the current approach to diagnose depression may have ignored its complex clinical reality (Rantala et al., 2018). In fact, when depressive symptoms are the presenting problems driving a client to receive mental health services, there could be many possible reasons behind the presenting depressive symptoms, such as infections, loneliness, trauma, other medical conditions, or something else, and different etiological factors may require different specific interventions (Rantala et al., 2018). In light of the growing body of evidence indicating that trauma and its consequences may contribute to the development of depressive symptoms (e.g. An et al., 2019; Fung, Lam, et al., 2022), as confirmed by the findings of the present study, we recommend that the potential existence of a trauma-related subtype of depression be further explored and given greater attention in both research and clinical practice. Theoretically and diagnostically, the underlying causes of depression should be recognized before we can provide personalized treatment and improve intervention effectiveness. As suggested by Rantala et al. (2018), treatment for depression should be informed by recognizing its subtype; if the depressive symptoms are related to loneliness, efforts should be made to reduce loneliness. Similarly, if the underlying problems are trauma-related symptoms, trauma-specific interventions should be considered. In fact, identifying subtypes of a psychiatric condition could facilitate researchers and practitioners to reveal etiological factors and improve interventions. Therefore, in the context of depression, a trauma-related subtype of depression should be further investigated and validated in the future (as suggested by Rantala et al., 2018), by employing diverse subtype validation strategies (e.g. conducting cross-method replication, and examining temporal stability and predictive validity) (see Agelink van Rentergem et al., 2021).

As post-traumatic and dissociative symptoms could be persistent and have considerable clinical consequences in people self-reporting depressive emotions, as demonstrated in our study, these trauma-related symptoms should be regularly screened for and treated in people seeking interventions for depression. As we observed, certain specific post-traumatic symptoms could predict later depressive symptoms; more importantly, these trauma-related symptoms generally did not remit after one year in most of our participants. In addition, dissociative symptoms were associated with more use of psychiatric hospitalization and emergency services. Therefore, early identification and management of trauma-related symptoms are important. It should be noted that trauma-related conditions are frequently overlooked in clinical settings, as evidenced by previous research (Cusack et al., 2006; Fung, Wong, Lam, Chien, Hung, Ross, et al., 2022). In our sample, only 36 participants (23.7%) self-reported a previous clinical diagnosis of PTSD/complex PTSD. In fact, treating depression without addressing the underlying post-traumatic symptoms may be ineffective; on the other hand, trauma-specific interventions could reduce depressive symptoms (Rosen et al., 2020). Thus, regular screening for trauma-related symptoms is recommended in mental health service settings. When trauma-related symptoms are recognized in people with depressive symptoms, evidence-based trauma therapies (e.g. cognitive processing therapy or phase-based interventions) should be considered in addition to standard depression treatment.

In addition, while PTSD and dissociative symptoms are persistent to a certain degree in people with depression, more research is needed to identify factors that contribute to the persistence of these trauma-related symptoms in this population. Our exploratory analyses showed that participants with persistent PTSD had more non-betrayal trauma, and participants with persistent dissociative symptoms had more betrayal trauma. It implies that a better understanding of the trauma histories of people with depression may have implications for interventions. Nevertheless, future studies should further explore what biopsychosocial factors would affect the trajectory of trauma-related symptoms in people with depression.

Our findings highlighted that a sense of current threat and negative self-concept may be particularly predictive of future depressive symptoms. According to the demoralization model, trauma-related symptoms such as hyperarousal and negative thoughts could lead to depressive symptoms like feeling sad and hopeless and having no motivation (Schindel-Allon et al., 2010). Our findings are consistent with a few previous studies in other populations showing that post-traumatic symptoms could predict subsequent depressive symptoms (e.g. An et al., 2019). People who perceive a sense of current threat may have a lower sense of safety, which could make them difficult to experience positive emotions or enjoy their lives. Trauma-related negative self-concept, such as self-blame, may be similar in phenomenology to feeling worthless in the context of depression. It is possible that trauma-related negative self-concept could contribute to the development of generalized depressive symptoms, although further research is needed in this regard. However, it is important to note that negative self-concept is not as specific to PTSD as other core PTSD symptoms (e.g. re-experiencing). The findings, on the one hand, suggest that some PTSD and depressive symptoms may share transdiagnostic and non-specific etiological factors or features in their expressions and development. On the other hand, our findings might also be partly due to the overlap of complex PTSD and depressive symptoms, highlighting the fuzziness of the traditional diagnostic boundaries that require future research. Regardless of the reasons behind the relationship between sense of current threat, trauma-related negative self-concept, and depressive symptoms, our results point to the importance of managing sense of current threat and trauma-related negative self-concept in people with depressive symptoms. Considering that a sense of current threat and negative self-concept may predict future depressive symptoms, interventions targeting these symptoms should be considered. For example, exposure-based therapy could be used to treat the sense of threat, while cognitive therapy may help reduce negative self-concept (Banz et al., 2022; Karatzias et al., 2019).

The present study has some strengths: we employed well-validated standardized measures to assess the mental health symptoms, we used an international and diverse sample, and we tested our hypotheses using longitudinal data. However, this study also has several limitations. First, we used online methods to recruit participants, and a number of participants who completed the baseline assessment did not respond to our follow-up survey, and therefore the sample may not be representative of people seeking interventions for depressive symptoms in clinical settings. Our findings would require replications in more representative clinical and nonclinical samples and in diverse language and sociocultural contexts. For example, we did not ask about ethnicity in the present study. The generalizability of our findings may also be limited because we excluded participants who reported immediate need for professional help or recurrent suicidal ideation, suicidal attempts or homicidal plans. Second, given self-selection bias, most participants were female in this sample. Third, we did not conduct structured interviews to assess the depression and trauma-related conditions in these participants; only self-report data could be analyzed, although these data were collected using reliable and valid assessment tools. Fourth, because we only measured traumatic experiences using the BBTS at baseline but not at follow-up, we did not know whether the participants had experienced new traumatic experiences during the follow-up period. Thus, we should acknowledge our limitation that we could not explain why there were 36% of participants who did not meet criteria for PTSD at baseline subsequently did at follow-up. Additionally, it should be noted that we adopted a continuum perspective on depression, and therefore we did not exclude participants with mild and minimal severity (i.e. scored below 10 on the PHQ-9). Having said that, the pattern of results from hierarchical multiple regression results remained the same even if we only included participants who reported moderate to severe levels (i.e. scored 10 or above on the PHQ-9) of depression. Finally, as emphasized above, the present study focused on the potential role of trauma-related symptoms in the context of depression. We acknowledge that the bidirectional relationship between trauma-related symptoms and depressive symptoms is a complex and important area of inquiry, but the present study focused primarily on the persistence, correlates, and outcomes of trauma-related symptoms in people reporting depressive emotions. Having said that, to provide additional information, a secondary exploratory analysis was done using the same dataset: hierarchical multiple regression analyses indicated that that baseline PHQ-9 scores also significantly predicted PTSD symptoms (β = .206, p = .012) and DSO symptoms (β = .2015, p = .031) at follow-up after controlling for baseline symptom scores. Thus, the bidirectional relationship between trauma-related symptoms and depressive symptoms in various clinical and nonclinical populations requires further studies, including in both people with depression and in people with trauma-related disorders.

4.1. Concluding remarks

This study is the first to show that post-traumatic and dissociative symptoms could be persistent and are associated with considerable clinical consequences in people self-reporting depressive emotions. We recommend that a trauma-related subtype of depression be given greater attention in both research and clinical practice. Trauma-related symptoms should be regularly screened for in people with depression. When there are comorbid trauma-related symptoms, timely trauma-specific interventions may be important to manage depressive symptoms.

Copyright statement

We acknowledge that we are submitting our original work, that we have the rights to the work, that we are submitting the work for first publication in the journal (that it is not being considered for publication elsewhere and has not already been published elsewhere), and that we did not include any works not owned by us.

Acknowledgment

The first author received the RGC Postdoctoral Fellowship Scheme 2022/23 offered by the Research Grants Council, Hong Kong.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The second author was supported by the Vice Chancellor’s Discretionary Fund of the Chinese University of Hong Kong.