Introduction

It is estimated that 7 out of 10 people worldwide will experience at least one traumatic event during their lifetime. Amongst those who experience a traumatic event the majority experience more than one, with an average number of three traumatic experiences (Kessler et al., 2017). Nevertheless, only about 6% of trauma survivors develop posttraumatic stress disorder (PTSD; Koenen et al., 2017), which is a mental illness with debilitating symptoms including re-experiencing (e.g., intrusive thoughts, flashbacks, or nightmares), avoidance (of thoughts and activities that remind of the event), and hyperarousal (Eilers & Rosner, 2021). These symptoms result in significant impairment of personal, social, educational, or occupational functioning (World Health Organization, 2022). Key risk factors for the development of PTSD include trauma severity and lack of social support following the traumatic event (Brewin et al., 2000).

Another central risk factor for developing and maintaining PTSD symptoms are negative posttraumatic cognitions (PTC; Brewin, 2020; Dalgleish, 2004; Ehlers & Clark, 2000; Foa et al., 1999; Meiser-Stedman et al., 2009), also called maladaptive cognitive appraisals (e.g., Agar et al., 2006). PTC include negative cognitions about the world, the self, and self-blame (Agar et al., 2006; Bryant & Guthrie, 2005; Carek et al., 2010). This is reflected in the current fifth edition of the diagnostic and statistical manual of mental disorders (American Psychiatric Association, 2013). In addition to the three PTSD symptom clusters—re-experiencing, avoidance, and hyperarousal—the DSM-5 includes negative alterations in cognitions and mood as a fourth PTSD symptom cluster. Thus, it is important to further examine the role of PTC in PTSD.

There are three theoretical models explaining the mental processes in PTSD, which are highly similar (Brewin & Holmes, 2003): the dual representation theory (Brewin & Burgess, 2014), the emotional processing theory (Rauch & Foa, 2006), and the cognitive model (Ehlers & Clark, 2000). The dual representation theory focuses on how traumatic memories are stored and integrated in the brain. The emotional processing theory emphasizes the importance of processing and integrating emotional responses to the traumatic event and the cognitive model emphasizes the role of negative beliefs and assumptions. Because of the strong overlap between the models, it is difficult to empirically differentiate between them. For illustration of the processes underlying PTSD, we are focusing Ehlers and Clark’s cognitive model, which puts a strong emphasis on the negative beliefs (i.e., PTC) most relevant to the current study.

Ehlers and Clark’s (2000) cognitive model postulates that fear is the result of appraisals related to imminent danger; even though the threat is a memory of a past event, the fear is generalized to the present and leads to avoidance, which in turn perpetuates the threat. PTC about the sequelae of traumatic events can elicit a sense of current threat, which can in turn cause and maintain the symptoms of PTSD (McNally & Woud, 2019). This sense of current threat can be the result of an overgeneralization of the perceived threat to the world (e.g., “The world is a dangerous place.”) or a threat to one’s view of oneself as a capable person (e.g., “I can´t rely on myself.”). PTC pre-existing before the traumatic event, especially about oneself, can promote the development and increase the severity of PTSD symptoms (Bryant & Guthrie, 2007).

The specific role of PTC in PTSD has been examined in a meta-analysis by De la Cuesta and colleagues (2019), which included 135 studies of patients exposed to a traumatic event—as defined in the DSM-5. The meta-analysis supports the association between PTC and PTSD symptoms in patients without treatment. The association had a large effect size (r = 0.53, 95% CI = 0.51–0.56). The largest effect size was found for negative cognitions about the self (r = 0.61), followed by negative cognitions about the world (r = 0.46), and self-blame (r = 0.28). However, the meta-analysis has excluded studies with a sample of patients with a sole PTSD diagnosis or studies of PTSD treatment trials. A review by Brown et al. (2019) investigated the association between changes in PTC and changes in PTSD symptoms in the course of a treatment. The review included 65 studies of trauma-exposed individuals. Most studies have found that a reduction in negative PTC during a treatment was associated with a reduction in PTSD symptoms. This association was further supported by a randomized controlled trial, which investigated the effects of a cognitive bias modification paradigm, an intervention specifically targeting PTC (Woud et al., 2021). In this study of inpatients with PTSD, the reduction of PTC during treatment was uniquely associated with a reduction in PTSD symptoms. This effect has also been shown in an experimental study that exposed healthy participants to distressing films as an analog trauma induction and received positive or negative cognitive bias modification via an app. They found that positive modification of cognitive bias elicited training-congruent judgments compared to negative modification (Würtz et al., 2022). Other studies investigating the direction of the relationship between PTC and PTSD found that changes in PTC preceded changes in PTSD symptoms (Bryant & Guthrie, 2007; Kleim et al., 2013; Kumpula et al., 2017; Schumm et al., 2015; Zalta et al., 2014).

Previous research supports the associations between PTC and the overall PTSD symptom severity. However, there are mixed results about distinct associations between PTC and the different PTSD symptom clusters. In a longitudinal study of sexually abused women, Carper et al. (2015) found that all types of PTC are associated with all three different PTSD symptom clusters. However, studies by Foa and Rauch (2004) and by Raab et al. (2015) discovered that only changes in negative cognitions about the self and the world are related to changes in all three PTSD symptom clusters. In the study by Foa and Rauch (2004), changes in self-blame are only associated with re-experiencing and hyperarousal. Raab et al. (2015) observed that self-blame is only associated with symptoms of re-experiencing and avoidance, but not hyperarousal. The results from another study showed that negative cognitions about the self are associated with all three PTSD symptom clusters but negative cognitions about the world are only associated with avoidance and hyperarousal and not re-experiencing symptoms (Hiskey et al., 2015). In contrast to the two other studies, self-blame is not associated with any of the three PTSD symptom clusters of this study.

PTSD symptoms are often accompanied by symptoms of anxiety, depression, substance abuse (Brady, 2000), and difficulties in emotion regulation (ER; Poole et al., 2018). Therefore, the associations between changes in PTC and changes in these accompanying symptoms should also be investigated. Several studies in the review by Brown et al. (2019) also assessed symptoms of depression (e.g., Kangas et al., 2013; Nacasch et al., 2015; Schumm et al., 2015) and anxiety (e.g., Beck et al., 2016; Kangas et al., 2013). Thus far, only one study by Barlow et al. (2017) has examined the association between PTC and ER strategies, suggesting that there might be a direct link between PTC and PTSD symptoms but also an indirect link via ER (Barlow et al., 2017). According to Gross and John (2003) ER can be divided into cognitive reappraisal and expressive suppression. Cognitive reappraisal refers to changing the way one thinks about the meaning of a stimulus, whereas expressive suppression refers to the active inhibition of external emotional expression. Previous studies found that expressive suppression is positively associated with PTSD symptoms (Boden et al., 2013; Khan et al., 2021; Lee et al., 2015; Seligowski et al., 2015), whereas cognitive reappraisal is negatively associated with PTSD symptoms (Boden et al., 2013; Khan et al., 2021). The suppression of thoughts and emotions related to the traumatic event corresponds to the avoidance symptom cluster (First, 2013). Furthermore, difficulties in regulating emotions in interpersonal situations can contribute to the development and maintenance of mood and anxiety disorders. The interpersonal model of emotion regulation emphasizes the processes by which interpersonal interactions influence emotional experiences and suggests that interventions that target interpersonal emotion regulation may be effective in treating mood and anxiety disorders (Hofmann, 2014). Thus, ER can be an important factor in the association between PTC and symptoms of depression and anxiety.

Moreover, only one previous study has investigated the association between PTC, ER and PTSD symptoms in a sample of participants who reported childhood abuse (Barlow et al., 2017). The results show that childhood abuse is linked to greater PTSD symptom severity and negative PTC (particularly self-blame) as well as more maladaptive ER.

The Present Study

The current study aimed at deepening our understanding by assessing the association between changes in PTC and changes in symptoms of PTSD, depression, anxiety, and difficulties in ER, during an inpatient PTSD treatment. First, we tested whether symptom severity improved during the psychotherapeutic treatment in a naturalistic clinical care setting. We hypothesized that PTSD-, depression and anxiety symptoms, as well as ER and PTC improve during treatment.

Secondly, we investigated whether the baseline levels of PTC were related to changes in symptoms of PTSD, depression, anxiety and ER. Thus, we tested whether stronger PTC were associated with less changes in symptoms.

Third, we hypothesised that changes in PTC were associated with changes in symptoms of PTSD, depression and anxiety symptoms; similarly to previous studies (Beck et al., 2016; Kangas et al., 2013; Schumm et al., 2015) and to changes in ER. Thus, we tested whether reductions in PTC after treatment were associated with less symptoms and improvements in ER. Further, we hypothesized that changes in PTC about the self were most strongly associated with changes in PTSD symptoms (De La Cuesta et al., 2019).

Method

Participants

We analyzed data from 339 patients (279 women, 60 men). All participants were inpatients in a trauma treatment unit of a psychiatric-psychosomatic hospital in Austria. The hospital requires patients to have a minimum age of eighteen years, a minimum level of conversational skills in German, and be motivated to actively engage in therapy practices. All individuals with acute psychotic symptoms, current suicidal behavior, or acute intoxication are excluded from treatment. Patients are typically referred to treatment by general practitioners, psychiatrists, or other health care professionals. Patients were included in the study if their self-reported PTSD symptoms at admission exceeded a cut-off value of ≥ 45 in the Post-traumatic Stress Disorder Checklist- Civilian Version (PCL-C), which is indicative of PTSD (Karstoft et al., 2014).

Design and Procedure

The current study retrospectively analyzed data collected during routine clinical care using a naturalistic pre-post-study design without a control group. The data was collected between June 2017 and April 2021. At first, the study aimed to examine whether symptom severity (PTSD, anxiety and depression symptoms, PTC, and ER) decreases between admission and discharge. Secondly, we calculated associations between baseline PTC and differences in symptom severity. Thirdly, we compared associations between changes in PTC and changes in symptom severity and ER.

Sociodemographic variables such as age and sex were collected in the hospital information system. All other measures were assessed using the Computer-based Health Evaluation System (Holzner et al., 2012). Participants completed the questionnaires electronically in a computer assessment room with up to seven other participants in separated cubicles. The assessment time at intake and discharge had a duration of two hours each, divided in two one-hour sessions. All participants consented to the processing of the data. The study was approved by the Ethics Committee of the Karl Landsteiner University of Health Sciences (No: 1003/2021) and complies with the ethical standards of the Declaration of Helsinki (World-Medical-Association, 2013).

Measures

The German version of the following questionnaires were used to assess PTC, ER and PTSD, depression, and anxiety symptoms, as well as childhood trauma and symptoms of borderline personality disorder (BPD).

Post-traumatic Cognitions Inventory (PTCI)

The PTCI assessed PTC regarding the traumatic event about the self, the world, and self-blame, within the last month (Foa et al., 1999). Next to a total scale score, three subscale scores for PTC about the self, the world, and self-blame were calculated. Previously, the questionnaire showed acceptable to excellent internal consistency as indicated by a Cronbach´s alpha of αtotal = 0.95, αself = 0.95, αworld = 0.83, and αblame = 0.77 (Müller et al., 2010). In the current study the internal consistency was equally good with αtotal = 0.93, αself = 0.92, αworld = 0.84 and αblame = 0.74.

Posttraumatic Stress Disorder Checklist—Civilian Version (PCL-C)

The PCL-C assessed symptoms of PTSD within the last month. It consists of three subscales: re-experiencing, avoidance, and hyperarousal (Höcker & Mehnert, 2012; Weathers et al., 1994). The total PCL-C score ranges from 17 to 85 (Weathers et al., 1994) and previous research has shown that a cut-off value of 45 or higher indicates a high level of PTSD symptoms (Karstoft et al., 2014).The PCL-C has previously shown excellent internal consistency for the total scale score with α = 0.94 (Conybeare et al., 2012) and acceptable to good internal consistency for the three subscales re-experiencing (α = 0.88), avoidance (α = 0.74) and hyperarousal (α = 0.88; Höcker & Mehnert, 2012). In the present study the internal consistency for the total PCL-C scale and for the re-experiencing subscale was equally good, with α = 0.78 respectively α = 0.81. The internal consistency for the two other subscales was questionable to poor with α = 0.65 for avoidance and α = 0.57 for hyperarousal. However, poor internal consistency is likely due to the fact that the present sample only includes patients with a total PCL-C score of 45 or higher, which limits variance.

Patient Health Questionnaire—Depression Module (PHQ-9)

Depression symptoms in the last two weeks were assessed with the German version of the PHQ-9 (PHQ-9; Kroenke et al., 2001a). In previous studies the PHQ-9 showed good internal consistency (α = 0.86; Beard et al., 2016) and test–retest reliability (r = 0.84; Kroenke et al., 2001b), good convergent and discriminant validity and good sensitivity to change. To detect major depressive disorder (MDD), a cut-off score of ≥ 10 has a sensitivity and specificity of 88% each (Manea et al., 2012). The German version showed excellent psychometric properties with α = 0.90 (Reich et al., 2018). In the current study internal consistency was also good with α = 0.80.

Patient Health Questionnaire—Anxiety module (GAD-7)

The GAD-7 was used to assess the level of symptoms of Generalized Anxiety in the last 2 weeks. Williams (2014) suggest a score of 10, with a sensitivity of 89% and a specificity of 82%, to detect moderate anxiety disorder. The internal consistency of the GAD-7 has previously been reported as excellent (α = 0.92) and the test–retest reliability as good (r = 0.83; Spitzer et al., 2006). The German version showed similarly good psychometric properties with α = 0.85 (Hinz et al., 2017). In the present study the internal consistency was also good with α = 0.80.

Emotion Regulation Questionnaire (ERQ)

The ERQ measured emotion regulation strategies and consists of two subscales: cognitive reappraisal and expressive suppression (Gross & John, 2003). Previous research showed good internal consistency for the cognitive reappraisal (α = 0.89–0.90) and acceptable internal consistency for the expressive suppression (α = 0.76–0.80) subscale (Preece et al., 2020). The German version showed similar psychometric properties for reappraisal (α = 0.82) and suppression (α = 0.76; Wiltink et al., 2011). In the current study only the reappraisal subscale showed good internal consistency (α = 0.86), the internal consistency of the suppression subscale was questionable (α = 0.66).

Childhood Trauma Questionnaire (CTQ)

The CTQ is a retrospective measure for childhood maltreatment in adults (Bernstein et al., 2003). The questionnaire uses 28 items to assess five types of childhood trauma: sexual abuse, physical abuse, emotional abuse, physical neglect, and emotional neglect. In addition to the scores for the five subscales a combined total score can be calculated. In previous studies the German version of the CTQ has shown good to excellent internal consistency (αtotal = 0.94, αemotional abuse = 0.89, αphysical abuse = 0.89, αsexual abuse = 0.96, αemotional neglect = 0.90, and αphysical neglect = 0.62) and good construct validity (Wingenfeld et al., 2010). In the current study the internal consistency ranged from excellent to questionable depending on the subscale, with αtotal = 0.91, αemotional abuse = 0.87, αphysical abuse = 0.9, αsexual abuse = 0.97, αemotional neglect = 0.60 and αphysical neglect = 0.75.

McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD)

The MSI-BPD assessed symptoms of BPD (Zanarini et al., 2003). The original English version showed acceptable internal consistency (α = 0.74), test–retest reliability (rs = 0.72), good validity, and diagnostic accuracy (Chanen et al., 2008; Noblin et al., 2013; Patel et al., 2011; Zanarini et al., 2003). The German version performed equally well with α = 0.96 (Kröger et al., 2010, 2011). The internal consistency in the current study was poor (α = 0.65).

Sociodemographic Variables

Sex (male, female) was defined in terms of the self-reported sex. Age refers to the chronological age (in years) at the time of admission. Education level was assessed based on the Austrian school system and later categorized into the following three education levels: (1) low education level (i.e., compulsory school education), (2) medium education level (i.e., secondary school education), and (3) high education level (i.e., college education).

Treatment

Participants received an eclectic mixture of trauma-focused therapy. Treatment was delivered in both individual and group settings by an interdisciplinary team and was heterogeneous between patients. Specific treatment protocols and elements included Dialectic Behavior Therapy for PTSD, Eye Movement Desensitization and Reprocessing (EMDR), Imagery Rescripting & Reprocessing Therapy (IRRT), Trauma-Focused Cognitive Behavioral Therapy (TF-CBT), Imagery Rehearsal Therapy (IRT), Narrative Exposure Therapy (NET), Concept of Structural Dissociation (Enactive Trauma Therapy), Somatic Experiencing, Neuroaffective Relational Model (NARM), and Schema Therapy.

Treatment was chosen based on therapist’s expertise and preferences. The mandatory group therapy sessions included open-topic groups, knowledge transfer, mindfulness training group, and skills training group, each lasting 50 min per week. Additional group therapies included art therapy, music therapy, group biofeedback and animal-assisted therapy. Patients participated in group therapy about 10 times per week and received two one-hour individual psychotherapy sessions per week. Most patients stayed for eight weeks, with a range of one to 13 weeks.

Statistical Analysis

All analyses used data with single imputations to account for missing values in questionnaires. Missing values were replaced with the mean score of the respective scale if less than 20% of the responses per scale were missing. Patients with more than 20% missing values were excluded from the analysis. Supplementary Fig. 1 displays the percentage of missing data before imputation at admission and discharge. Supplementary Tables 1 and 2 display correlations between PTC, symptoms of depression, anxiety and PTSD, and emotion regulation, at the times of admission and discharge, respectively.

First, we tested for changes between admission and discharge. To test these changes, we conducted paired-sample t-tests for all 13 measures and reported Cohen’s d effect sizes (with pooled standard deviations). To control for multiple testing, Bonferroni correction was used lowering the significance level to p ≤ 0.004. We conducted these t-Tests first, using addressing a completer sample and secondly, an intent-to-treat analysis (ITT). In the completer sample participants were only included if they provided data for both measurement points. We additionally applied the last-observation-carried-forward (LOCF) method in line with ITT replacing missing values at discharge with values from admission of that particular person (= the last observation is carried forward). This method is a commonly used statistical approach for analyzing longitudinal data with repeated measures when data are missing (Lachin, 2016). It is the most conservative method for participant attrition in an intent-to-treat analysis (ITT).

Secondly, we examined the associations between PTC at admission and changes in symptoms. We only included PTC subscales, not summed PTC data. Symptoms referred to PTSD, depression, anxiety, and ER strategies. We calculated eight multiple linear regressions, with the three PTC subscales at admission as predictors, controlling for gender and age, and the respective change scores of the symptoms as exploratory variables.

Thirdly, we calculated the associations between changes in PTC and changes in symptoms..Again, we only included PTC subscales, not summed PTC data. We calculated eight multiple linear regressions, with the change scores of the three PTC subscales as predictors, controlling for gender and age and change scores of symptoms of PTSD, depression and anxiety and ER as exploratory variables.

Results

First, we describe the sample characteristics. Then we present descriptive statistics of all measures at admission, discharge, and the intent-to-treat sample with the LOCF method. Subsequently, we compare changes in all measures between admission and discharge and between admission and intent-to-treat (LOCF) with paired sample t-tests. Finally, we analyze the connections between PTC at admission and changes in symptoms of PTSD, depression and anxiety and ER as well as changes in PTC and changes in symptoms of PTSD, depression and anxiety, and ER with multiple linear regression analyses (Table 1).

Table 1 Sociodemographic characteristics of the study sample—Admission

Sample Characteristics

Descriptive Statistics

Table 2 displays means and standard deviations of all measures at the times of admission and discharge as well as for the intent-to-treat analysis (LOCF method). The absence of discharge assessment data is either due to dropout or because of administrative and clinical reasons (e.g., they were discharged before the assessment was scheduled, holidays, scheduling conflict, sick leave of testing stuff, therapists felt that the patients were still too distressed to complete the discharge assessment).

Table 2 Means and standard deviations of all measures

Changes in the Course of Treatment

We used t-tests to assess changes between admission and discharge (see Table 3). There was a significant change in the PTCI total score and the two subscales negative cognitions about the world and the self with medium effect sizes. Self-blame did not show a significant change. PTSD symptoms had significantly improved in the course of treatment with medium effect sizes for the total PCL-C sum score, the PCL-C subscales avoidance and hyperarousal and low effect sizes for the re-experiencing subscale.

Table 3 T-tests of changes in all measures of treatment completers between admission and discharge

In the PCL-C sum score, a decrease of at least 5 points indicates a reliable improvement and a decrease of at least 10 points indicates a clinically significant improvement in PTSD symptoms. In our sample, only 121 (54%) participants had decreased at least 5 points and only 77 (34%) had decreased at least 10 points. Moreover, at the time of discharge 88.50% of patients still had a PCL-C score of 45 or higher indicating a high level of PTSD severity.

Depression and anxiety severity, ER strategies and BPD symptom severity all significantly improved in the course of treatment with high effect sizes for depression severity, medium effect sizes for anxiety severity and BPD symptoms, and low effect sizes for ER strategies. Table 3 presents patients who scored above the cut-off of 10 on the PHQ-9, indicating a moderate level of depression (Kroenke et al., 2001a). Further, the presented GAD-7 ≥ 10 indicated a moderate level of anxiety severity (Spitzer et al., 2006). Moreover, patients who had a MSI-BPD score of 7 or higher showed comorbid BPD symptomatology (Zanarini et al., 2003).

Regarding changes in depression severity, only 111 participants (45%) changed at least 5-points, which would indicate a reliable improvement (McMillan et al., 2010). Regarding anxiety symptoms, only 62 participants (26%) had reliably improved in the GAD-7 score, as indicated by a minimum change of 6-points (Bischoff et al., 2020). Concerning changes in ER strategies, the participants’ use of reappraisal increased, whereas the use of suppression of emotions decreased in the course of treatment.

The intent-to-treat data (LOCF) showed similar results for t-tests comparing the measures of PTC and symptoms of PTSD, depression and anxiety and ER between admission and discharge (see Table 4). Regarding the PTCI there was a significant change in the PTCI sum score as well as the subscales negative cognitions about the self and about the world (with small effect sizes), but no significant change for the PTCI subscale self-blame. PTSD symptoms significantly improved with medium effect sizes for the PCL-C sum score and low effect sizes for the three PCL-C subscales (re-experiencing, avoidance, and hyperarousal). Symptom severity of depression and anxiety, ER strategies, and BPD symptom severity all significantly improved in the intent-to-treat analysis with medium effect sizes for depression and anxiety severity and low effect sizes for ER strategies and BPD symptom severity.

Table 4 T-tests of changes in all measures between admission and the last observation carried forward analysis (Last observation carried forward)

Relationship Between PTC and Symptoms of PTSD, Depression and Anxiety, and ER

To assess the relationship between PTC and symptoms of PTSD, depression, anxiety, and ER, we calculated two sets of eight multiple linear regression analyses. Table 5 displays the results of our first set with baseline levels of PTCI as predictors and the changes in PCL-C-, PHQ-9-, GAD-7- and ERQ scores as explanatory variables. The PTCI subscale self-blame at admission was significantly associated with changes in depression severity. However, there were no other significant associations between the PTCI scale scores at admission and changes in any of the above measures.

Table 5 Regression of post-traumatic cognitions at admission and changes in symptoms of posttraumatic stress disorder, depression, anxiety and emotion regulation in the course of treatment

Table 6 shows the results of the second set of regression analyses with changes in the PTCI scores as predictor variables and changes in PCL-C, PHQ-9, GAD-7 and ERQ scores as explanatory variables. Changes in the PTCI subscale negative cognitions about the self were significantly associated with changes in all the above measures. Changes in the PTCI subscale negative cognitions about the world were significantly associated with all PCL-C scores, except re-experiencing. The PTCI subscale self-blame was only significantly associated with the PCL-C subscale re-experiencing but not with any of the other measures. However, the change in the PTCI subscale self-blame was not significant (see Tables 2, 3, 4).

Table 6 Regression of changes in post-traumatic cognitions and changes in symptoms of PTSD, depression and anxiety and ER in the course of treatment

Discussion

This study has investigated the role of PTC on treatment outcomes of inpatients with PTSD. We were especially interested in the distinct association between PTC subscales and different PTSD symptom clusters. Additionally, we focused on the role of PTC for symptoms of depression and anxiety as well as ER strategies during treatment.

First, analyses of the treatment completers show that, except for self-blame, the other PTC (i.e., negative cognitions about the world and about the self) as well as symptoms of PTSD, depression and anxiety and ER strategies have improved between admission and discharge. The analyses using the last observation carried-forward method show the same results. Thus, regardless of selective drop-out, symptom severity improves in the course of the inpatient treatment. Hence, the eclectic trauma-focused treatment is likely to have resulted in an improvement of symptoms. This is in line with previous evidence for the effectiveness of trauma focused treatments regarding PTSD (e.g., Ehring et al., 2014; Schnurr, 2017), symptoms of depression (e.g., Ehring et al., 2008; Schumm et al., 2015), anxiety (e.g., Beck et al., 2016; Littleton et al., 2012) and ER (van Toorenburg et al., 2020). The interpretation of the current data is limited by the absence of a control group.

Although PTSD symptoms as well as symptoms of depression have improved significantly in the current study, these improvements are not reliably or clinically significant for the majority of the patients. This is indicated by the small percentage of patients for whom the improvement is above the level of the reliable change index (RCI). Unfortunately, this also suggests that there may not have been sustained improvement in PTSD symptoms in the patient group. Thus, although the eclectic inpatient trauma-focused treatment in the present study has led to a statistically significant symptom reduction, it is unclear if symptoms have been reduced sufficiently to induce sustained improvements. The relatively small effect sizes stand in contrast to the resource heavy eight-week inpatient treatment, which included two individual and 10 group therapy sessions per week. However, the effect sizes are comparable to the results of a meta-analysis on the improvement of symptom severity of inpatient psychiatric-psychosomatic rehabilitation (Sprung et al., 2018). Secondly, we have investigated associations between baseline levels of PTC on symptom changes. Only the level of negative cognitions about the self at admission is related to changes in symptoms of depression. This is unsurprising given that negative cognitions about the self include feelings of hopelessness, sadness, and the inability to cope with situations, as well as self-doubt, which are akin to a depressive thinking style. The absence of significant associations between the PTC subscales negative cognitions about the world and self-blame and symptom change might have resulted from sample characteristics. We have only included patients with PTSD, therefore the variance in PTSD symptoms may be limited, which could have influenced the analyses. In fact, the meta-analysis by De la Cuesta et al. (2019) excluded studies with an all PTSD patient sample due to the limited variance in PTSD symptoms.

Thirdly, we have assessed the association between changes in PTC and changes in symptoms of PTSD, depression, anxiety, and ER. The changes during treatment are significantly associated with changes in PTC. The results are in line with previous findings that show that PTC change with PTSD symptoms (e.g., Brown et al., 2019; De La Cuesta et al., 2019). Previous studies found that changes in PTC are associated with changes in PTSD (e.g., Ehring et al., 2014; Schnurr, 2017) changes in symptoms of depression (e.g., Ehring et al., 2008; Schumm et al., 2015) and anxiety (e.g., Beck et al., 2016; Kangas et al., 2013; Littleton et al., 2012); the current data replicates these findings. Additionally, changes in PTC are associated with changes in ER. Notably, the results support previous evidence that changes in negative cognitions about the self are most strongly related to reductions in PTSD symptoms (De La Cuesta et al., 2019; Foa & Rauch, 2004; Nacasch et al., 2015).

Negative cognitions about the self are also associated with changes in symptoms of depression and anxiety and are related to changes in ER. This suggests that the influence of negative cognitions about the self may go beyond symptom reduction and may also foster skill development—possibly because reducing negative cognitions about the self may in turn foster self-competence. In contrast, changes in negative cognitions about the world are only related to changes in symptoms of PTSD. This suggests that there might be a specific link between negative cognitions about the world and symptoms of PTSD. Changes in self-blame are only associated with changes in re-experiencing. However, self-blame has not changed significantly during treatment, thus this association is unlikely to be the result of a successful treatment of self-blame. There is contradictory evidence for the role of self-blame in PTSD treatment, even though some studies found an improvement during treatment (Beck et al., 2016; Schumm et al., 2015), the majority of studies have not (e.g., Carper et al., 2015; Foa & Rauch, 2004; Foa et al., 2005; Gray et al., 2012; Kumpula et al., 2017; McLean et al., 2019). These contradictions might have resulted from differences in the intervention programs. Since there is a great variability in treatment and treatment dose in the present sample, this might as well contribute to the finding that self-blame has not changed significantly during treatment.

A limitation of the study is that no structured diagnostic interviews were used to determine diagnoses. However, using the scores from the PCL-C allows to reliably distinguish between traumatised and non-traumatised individuals (Gelaye et al., 2017). The study is limited to measuring different types of childhood abuse but has not captured traumatic experiences in adulthood. Thus, we were not able to address the effects of different types of traumatic experiences in adulthood. Future studies should include structured assessments of the type of traumatic experiences to determine if there are differences in the associations between PTC and PTSD symptoms depending on the type of traumatic experience. Another limitation of the current study is that we have not included a control group. It is, thus, possible that changes over time in PTC may occur even without an intervention, however, the study is highly relevant to evaluate effects in a naturalistic clinical care setting. Notably, the patients in the current study have not received a uniform treatment program. Depending on the decisions of their respective therapists, they have received different treatments and therefore showed great variability in treatment and treatment dose. Some symptoms may improve with higher session frequency or different treatment. Moreover, the treatment may have been too eclectic; a solely trauma-focused CBT approach might have produced better results. Another study investigating the effects of a treatment that specifically targeted PTC using a cognitive bias modification paradigm has found that such a training can significantly improve PTC (Woud et al., 2021). However, treatment effects have not lasted until follow-up. In contrast, another study showed that adding specific cognitive interventions has not increased changes in PTC (Foa et al., 2005). Hence there might not be a need to specifically target PTC in the treatment. Further studies need to be conducted regarding these mixed results. Moreover, future studies should investigate changes in interpersonal ER (Hofmann, 2014) in response to PTSD treatment.

In summary, the present findings confirm the decrease in symptom severity and PTC during inpatient trauma-focused treatment, except for the PTC self-blame. The data show that only baseline level of PTC about the self are related to changes in depression severity. Baseline level of PTC about the world and self-blame are not associated with changes in symptom severity. The data also supports that PTC and symptoms of PTSD, depression and anxiety, and ER change with trauma-focused treatments. Particularly, reductions in negative cognitions about the self change with all symptoms and ER. Reductions in negative cognitions about the world are only linked to changes in symptoms of PTSD and Changes in self-blame are only associated with symptoms of re-experiencing. These general reductions of symptoms may be caused by the same processes. The model by Ehlers and Clark (2000) assumes that PTSD is caused by generalizing past fears about the traumatic event to the present. A treatment that successfully reduces PTC might imply that these fears (i.e., negative thoughts) are no longer generalized into the present and as a result symptoms of depression and anxiety might cease.

Given the mixed findings on the effectiveness of PTSD interventions that target PTC (Foa & Rauch, 2004; Woud et al., 2021) future studies should compare the effects of different treatments on PTSD. Further, the role of self-blame in PTSD treatment should be clarified (e.g., Carper et al., 2015; Raab et al., 2015). Specifically targeting self-blame during treatment and assessing PTC as well as PTSD symptoms on three different measurement points (pre and post-appraisal training and follow-up) would help to further investigate the causality and the direction of this underlying mechanism.