8-day intensive treatment programme for PTSD and complex PTSD vs treatment as usual: a clinical trial

ABSTRACT

Background: Intensive treatment programmes (ITP) have emerged as a treatment option for Posttraumatic Stress Disorder (PTSD) and Complex PTSD (CPTSD) during the last decade. Results are promising, but further studies with control group are needed.

Objective: To compare the effect of an 8-day ITP with traditionally spaced treatments for PTSD.

Method: Participants with PTSD or CPTSD (n = 101) in a Swedish psychiatric outpatient clinic received either 8-day ITP, including prolonged exposure, eye movement desensitisation and reprocessing therapy, physical activity and psychoeducation, or traditionally spaced treatment (TAU). Participants were allocated by shared decision-making.

Results: A significant reduction in PTSD symptoms was observed at posttreatment, with large effect sizes in both conditions. Symptom reduction was maintained at follow-up. There was no significant difference between treatment groups in degree of symptom reduction at posttreatment and follow-up. In the ITP, 73.3% did not meet criteria for PTSD at follow-up, and 74.4% in TAU. There was a significant difference in dropout rates between treatment groups: 4.3% in ITP and 24.1% in TAU.

Conclusions: Results indicate that ITP can be considered an effective treatment in healthcare settings where multiple treatment options are available. Randomised control studies to isolate treatment effects and finding underlying factors explaining the difference in dropout rates are important directions for future research.

HIGHLIGHTS

• An 8-day intensive treatment programme is equally effective as a 16-week treatment as usual.

• Dropout levels were significantly lower in the intensive treatment programme (4.3%) than in the PTSD treatment delivered once a week (24.1%).

• ITP is an effective treatment option for patients with Complex PTSD and previously treatment-resistant PTSD; 73.3% no longer met criteria for PTSD at 3-month follow-up.

There are several effective and recommended treatments for posttraumatic stress disorder (PTSD) (Lewis et al., 2020; NICE, 2018). However, while the effectiveness of PTSD treatment can be considered well-established, the mean dropout from PTSD treatments in a systematic review was 20.9% (Varker et al., 2021). One strategy to reduce dropout, especially among clients with complex symptom profiles or extensive trauma histories, is to adapt evidence-based protocols into flexible formats tailored to individual needs (Karatzias & Cloitre, 2019). Another approach to develop effective treatments with a shorter time frame that has also demonstrated low dropout rates is intensive treatment programmes (ITP) for PTSD (Burback et al., 2024). While evidence-based PTSD treatments are typically delivered by one session per week, ITP sessions are massed into treatment schedules ranging from 1-3 weeks (Foa et al., 2018; Zepeda Méndez et al., 2018). Apart from showing levels of symptom reduction comparable to traditionally spaced therapy, several studies have reported lower dropout rates than week-by-week treatment (Ragsdale et al., 2020; Sciarrino et al., 2020). Given the psychologically painful and disturbing nature of PTSD, the prospect of significantly shortening the treatment period is positive from a patient's perspective. From a socioeconomic perspective, fewer dropouts and less time consumed could be important in light of the substantial costs associated with PTSD (Davis et al., 2022). However, to be considered a favourable choice of treatment for both individuals and healthcare providers, ITP should be effective in both the short term and the long term. ITP formats have shown maintained treatment effects over 12 months (Klaeth et al., 2024), also in direct comparison to traditional PTSD treatment (Dell et al., 2023). Thus, the intensified schedule does not necessarily compromise the long-term effects of PTSD treatment.

Since the introduction of the formal diagnosis of Complex PTSD (CPTSD) in the International Classification of Diseases (ICD-11) (WHO, 2019), the need to find treatment options for more severe forms of PTSD has been highlighted (ISTSS, 2019). According to ICD-11, the core PTSD symptoms are intrusive re-experiencing of a traumatic event, avoidance of reminders and memories of the event, and perception of the heightened current threat. In addition to PTSD, a diagnosis of CPTSD requires severe and persistent emotional dysregulation, persistent negative beliefs about oneself characterised by shame and guilt and persistent difficulties in maintaining relationships and feeling close to other people; together, these symptoms are labelled disturbance in self-organisation (DSO). An 8-day ITP has been designed in the Netherlands (Van Woudenberg et al., 2018) to be used not only in the treatment of PTSD but also for patients with additional difficulties such as suicide ideation, extensive psychiatric comorbidity and previous but unsuccessful PTSD treatment. The programme includes prolonged exposure (PE) (Foa et al., 2021), eye movement desensitisation and reprocessing therapy (EMDR) (Shapiro, 2018), physical activity and psychoeducation. The 8-day ITP has shown effect not only for PTSD with severe comorbidity but also for patients with CPTSD (Voorendonk et al., 2020).

To our knowledge, the 8-day ITP has yet to be directly compared to traditionally spaced PTSD treatments. While results are promising, also in patients with severe psychiatric comorbidity, public healthcare needs to be well informed when making decisions on which treatments should be made available. Thus, comparing the 8-day ITP to evidence-based, traditionally spaced PTSD treatments is an essential next step.

This study aimed to compare treatment outcomes of an 8-day ITP and traditionally spaced PTSD treatment (TAU) in a Swedish psychiatric outpatient clinic. Specifically, we wanted to examine whether a treatment effect in PTSD and CPTSD could be observed after the 8-day ITP and TAU and whether a difference in levels of symptom reduction and functional impairment could be observed between treatment groups. In addition to treatment outcomes, potential adverse or unwanted effects of either treatment option were to be examined: dropout rates and increased suicidal ideation. The main hypothesis was that the effect of ITP would not be inferior to TAU in terms of reduction in PTSD and CPTSD symptoms. Based on previous research findings (Sciarrino et al., 2020), we also hypothesised that dropout levels in the IPT condition would be lower than in TAU.

1. Method

1.1. Participants

All patients referred to a psychiatric outpatient clinic specialising in PTSD treatment at Södra Älvsborgs Sjukhus (SÄS) between August 2021 and December 2023 were assessed for PTSD and CPTSD. Patients were referred from either primary care centres or other psychiatric clinics. Inclusion criteria: (1) age 18 and above; 2a) CPTSD according to ICD-11; 2b) PTSD with at least one comorbid disorder requiring psychiatric outpatient treatment; 2c) PTSD with previous unsuccessful trauma-focused treatment. Exclusion criteria: 1) ongoing substance or alcohol abuse; 2) use of benzodiazepines; 3) acute suicidality; 3) dissociative identity disorder (DID); 4) insufficient knowledge of the Swedish language to complete questionnaires. See Figure 1 for a record of the inclusion and exclusion process. Written consent was obtained from all patients after they had received a complete description of the study. Procedures involving patients were approved by the Swedish Ethics Committee in Lund (DNR 2021-018006).

Figure 1.

Study flowchart.

1.2. Procedure

This study was designed as an open clinical trial with control group. Upon referral, patients were assessed regarding their history of traumatic events with the Life Events Checklist for DSM-5 (LEC-5) (Weathers et al., 2013a). For the diagnostic assessment of PTSD and CPTSD, the semi-structured International Trauma Interview (ITI) 3.2 (test version) (Roberts et al., 2019) was used. The clinicians received training in applying the ITI manual prior to the study's commencement. All eligible participants were screened for high levels of dissociative symptoms using the Dissociative Experiences Scale II (DES II) (Körlin et al., 2007). A cut-off level of >20 (Chu, 2011) on the subscale DES-T was used to determine whether further assessment regarding pathological dissociation should be initiated with the SCID-D (Steinberg, 1994). Beck’s scale for suicide ideation (SSI) (Beck et al., 1979) was used to screen high levels of suicidal intentions at pretreatment; the sum of scores on items 4 and 5 was used to assess suicide risk during treatment. If the risk of suicide attempts was deemed acute, participants were excluded and referred to appropriate treatment. The DSM-5 Self-Rated Level 1 Cross-Cutting Symptom Measure (DSM-5 CCSM) Adult version (American Psychiatric Association [APA], 2013) was used as an indication of psychiatric comorbidity but not for diagnostic purposes (Bastiaens & Galus, 2018). Eligible patients who accepted the study's conditions were allocated to either 8-day ITP or TAU, according to principles of shared decision-making (Joosten et al., 2008). Participants were presented with written and verbal information about the treatment options. In discussion and subsequent decisions regarding treatment choice, both participant and clinician were active, while the participant was allowed to make their own judgement on suitable treatment options.

Participants completed pretreatment measures between 1 and 2 weeks prior to treatment; see Table 1 for a summary of the measures. SSI (Beck et al., 1979) was administered mid-treatment. Posttreatment assessment was scheduled a week after treatment completion. Follow-up assessment was scheduled 3 months after treatment completion. Self-reported data were collected via REDCap (Research Electronic Data Capture) (Harris et al., 2019), an electronic data capture tool hosted at Gothia Forum REDCap (RRID: SCR_003445). Participants entered symptom ratings via a QR code or web link. Participants who were uncomfortable using electronic devices were offered paper versions of self-report measures.

Table 1.

Measures and time points.

Pretreatment	Post-psychoed.	Posttreatment	3-month follow-up
LEC-5 ITI SSI DES-II SCID-Da PCL-5 ITQ DSM-5 -CCM WHODAS	PCL-5 ITQ WHODAS	ITI SSI PCL-5 ITQ	ITI SSI PCL-5 ITQ WHODAS

^aIn case of elevated DES II score and suspected dissociative syndrome.

^bTAU participating in pretreatment psychoeducation.

Abbreviations: LEC-5 = Life Events Checklist, ITI = International Trauma Interview (test version 3.2), SSI = Beck’s Scale Suicide Ideation, DES-II = Dissociative Experience Scale II, SCID-D = Structured Clinical Interview for DSM-IV (R) Dissociative Disorders, PCL-5 = Post Traumatic Checklist for DSM -5, ITQ = International Trauma Questionnaire, DSM-5 CC = DSM-5 Self-Rated Level 1 Cross-Cutting Symptom Measure, WHODAS = World Health Organisation Disability Assessment Schedule 2.0.

1.3. Treatment

The 8-day intensive treatment programme was performed as described by Van Woudenberg et al. (2018). Treatment planning and decisions on trauma memories to be targeted were scheduled a week before the start of the ITP. Typically, a new trauma memory was targeted each day, but the procedure was adapted to individual needs. Each day of treatment included PE, EMDR, physical activity and psychoeducation (see Table 2). Both PE (Foa et al., 2021) and EMDR (Shapiro, 2018) protocols were slightly modified due to the treatment format. In-vivo exposure between sessions was removed from the PE protocol but still emphasised in sessions, while stabilisation before trauma processing was removed from the EMDR protocol. Psychical activity was included in the treatment programme to regulate physical tension, prevent rumination and passive re-experience of trauma memories, and increase body awareness. The activities included both high-pulse exercise and relaxation. Psychoeducation about PTSD, coping and regulation skills was given in the evening. During the ITP, participants stayed at the hospital hotel at SÄS and had access to telephone coaching outside of treatment hours. Therapist rotation was used for PE and EMDR sessions based on the assumption that less focus on the relationship between patient and individual therapists would increase focus on processing trauma memories (Van Minnen et al., 2018). Treatment progress and planning were reviewed in a daily 30-minute meeting.

Table 2.

Schedule 8-day ITP.

Morning
	Physical activity 30 min PE 90 min Physical activity 30 min Lunch
Afternoon
	Physical activity 60 min EMDR 90 min Physical activity 30 min Dinner
Evening
	Psychoeducation 90 mina

^aExcept days 4 and 8.

Treatment as usual (TAU) included 16 sessions of PE or EMDR, each lasting 90 min and scheduled once a week. TAU also included optional pretreatment group psychoeducation about PTSD, coping and regulation skills, 6 sessions, scheduled once a week.

A total of 9 therapists from different healthcare professions conducted treatment in both ITP and TAU conditions. All therapists were trained in PE, and 6 therapists were trained in EMDR. Therapists rated protocol adherence for each session on a 4-point Likert scale developed by Blomdahl et al. (2018). The scale was modified for PE and EMDR, and possible responses ranged from 1 (‘did not follow at all) to 4 (‘followed the manual completely’).

1.4. Outcome measures

For the diagnostic assessment of PTSD and CPTSD, the International Trauma Interview (ITI) 3.2 (test version) (Roberts et al., 2019) was used. ITI 3.2 is a semi-structured interview divided into two parts. Part one has 7 items to help the interviewer assess the three symptom domains of PTSD and 2 items to assess functional disability. Part 2 has 6 items to assess the three DSO domains and 2 items to assess functional disability. All items are evaluated on a 5-point Likert scale based on symptom intensity and frequency during the last month (0 = ‘absent’ to 4 = ‘extremely). ITI 3.2 contains items to assess symptom duration and overall validity. Ratings are summed in a matrix where at least one item per symptom domain must be rated ≥2, and at least one item for functional disability must be rated ≥ 2 to make up a diagnosis of PTSD and CPTSD (PTSD + DSO), respectively. ITI 3.2 has shown satisfying interrater reliability (α = .76) (Bondjers et al., 2019).

To make results interpretable for clinicians using both DSM-5 and ICD-11 for diagnosis and evaluation, two primary measures were used. The first primary outcome measure was the Post Traumatic Checklist for DSM-5 (PCL-5) (Weathers et al., 2013b), a self-report measure for PTSD symptoms according to DSM-5 criteria. PCL-5 contains 20 items, rated on a 5-point Likert scale (0 = ‘not at all’ to 4 = ‘extremely’). It has shown good internal reliability (.96) and test-retest reliability (.84) (Bovin et al., 2016). PCL-5 has also demonstrated sensitivity to clinical change during treatment (Marx et al., 2022). The baseline internal consistency in this sample was good (Cronbach’s α = .84).

The second primary outcome measure was the International Trauma Questionnaire (ITQ) (Cloitre et al., 2018), which is a self-report measure developed to measure PTSD symptoms and CPTSD symptoms according to ICD-11. ITQ contains 18 items in total, 12 items to assess symptoms (6 for PTSD and 6 for DSO) and six items to assess functional disability associated with PTSD and CPTSD. In this study, only items measuring symptoms were included. Items are rated on a 5-point Likert scale (0 = ‘not at all’ to 4 ‘extremely’). ITQ has demonstrated good internal reliability, ranging from .89 to .94 (Cloitre et al., 2021). The baseline internal consistency in this sample was acceptable (Cronbach’s α = .76).

The World Health Organisation Disability Assessment Schedule 2.0 (WHODAS 2.0), a 12-item self-report version was used to assess levels of functional impairment according to the International Classification of Functioning, Disability and Health (ICF) (World Health Organisation [WHO], 2001), p. 12 items measure functioning in everyday situations, including cognition, mobility, self-care, relationships, activities, and participation in society. Items are rated on a five-level scale from 0 ‘no difficulty’ to 4 ‘extreme difficulty’. The simple sum of all scores was used to express levels of functional disability from 0 to 100% (Ustün et al., 2010). The 12-item version of WHODAS 2.0 has demonstrated high internal consistency (ranging from .83 to .92) (Axelsson et al., 2017). The baseline internal consistency in this sample was acceptable (Cronbach’s α = .87).

1.4.1. Data analysis

Data analyses were conducted in IBM SPSS statistics version 29. Data was scanned for entry errors, and assumptions underlying planned statistical analysis were checked. Differences between treatment groups in demographics, traumatic exposure and baseline clinical data were conducted using independent t-tests for scale data and χ² tests for nominal data. Dropouts were compared to treatment completers using the same procedures as those for the demographic and baseline data comparisons. Treatment effects of psychoeducation within TAU were assessed using paired samples t-test. 14% of the data was missing at posttreatment and 23% at follow-up. Missing data was primarily due to dropouts, with 3% of data missing among participants who completed treatment. In the primary and secondary analyses, multiple imputation (Fully Conditional Specification) was used to ensure that the sample adequately reflected the population (see appendix for full specification). Baseline clinical variables, including age, sex, and treatment group, were included in the imputation model. Sensitivity analyses were conducted using complete cases to assess the model's robustness (see Appendix 2). We performed a longitudinal data analysis (LDA) using a linear mixed-effects model (LMM) to evaluate effects of time and group x time. The restricted maximum likelihood (REML) estimator was used, with the model specified to include random intercepts to account for individual differences in baseline levels while slopes were fixed. Unstructured and autoregressive heterogeneous residual structures were tested but did not converge. A variance components covariance structure was used for the repeated effects, and random intercepts were included to ensure that within-subject correlation was still modelled. Hedges’ g was calculated as a measure of between-group effect size using pooled standard deviation of pretreatment scores. Hedges’ g was calculated using the baseline standard deviation for the within-group effect of time.

SSI scores were severely skewed and were thus transformed into median scores. Friedman’s test was used to analyse progress over time within each group. Dunn-Bonferroni post hoc tests were conducted to assess change between time points.

Differences between dropouts and completers were analysed with two-tailed independent t-tests for numerical data and χ² test for nominal data. Reliable change index (RCI) was calculated according to Jacobson and Truax (1991), using standard deviation and test-retest reliability from Sloan et al. (2018) and Cloitre et al. (2021) for PCL-5 and ITQ respectively. An RCI greater than 1.96 was considered a reliable change.

2. Results

2.1. Baseline characteristics

No significant differences were observed between treatment groups in demographics and baseline clinical measures, except for a higher rate of CPTSD diagnosis in the ITP condition (see Table 3). Trauma exposure is summarised in Table 4.

Table 3.

Sample characteristics.

Characteristic	Total (n = 101)	8-day IT (n = 47)	TAU (n = 54)
Demographics
Age mean (SD)	36.0 (11.3)	34.9 (10.3)	37.1 (12.0)
Gender female n (%)	81 (80.1%)	42 (89%)	39 (72%)
Married or partner n (%)	39 (38.6%)	19 (42.3%)	20 (44.4%)
Children n (%)	34 (33.6%)	20 (42.6%)	14 (25.9%)
Education
Jr High school n (%)	11 (10.8%)	4 (8.5%)	7 (13.2%)
High school n (%)	51 (50.4%)	25 (53.2%)	26 (49.1%)
Vocational school n (%)	10 (9.9%)	3 (6.4%)	7 (13.2%)
University n (%)	27 (26.7%)	15 (31.9%)	12 (22.6%)
Occupational status
Employed n (%)	49 (48.5%)	22 (46.8%)	27 (50%)
Student n (%)	7 (6.9%)	3 (6.4%)	4 (7.4%)
Unemployed n (%)	8 (7.9%)	6 (12.8%)	2 (3.7%)
Work-incapacity n (%)	48 (47.6%)	19 (40.4%)	29 (53.4%)
Treatment current and history
Previous PTSD treatment n (%)	48 (47.6%)	25 (53.2%)	23 (42.6%)
Pharmacological treatment n (%)	43 (42.5%)	25 (53.2%)	18 (33.3%)
Comorbidity
ITI – CPTSD n (%)	43 (42.5%)	26 (55.3%)*	17 (31.5%)*
DES-II mean (SD)	32.2 (15.6)	32.1 (15.4)	32.4 (16.2)
DSM-5 CCM (above threshold**)
Depression n (%)	96 (95%)	45 (95.7%)	51 (94.4%)
Hypomania n (%)	39 (38.6%)	18 (38.3%)	21 (38.9%)
Anxiety n (%)	99 (98.0%)	47 (100%)	52 (96.3%)
Psychosis n (%)	27 (26.7%)	14 (29.8%)	13 (24.1%)
Somatic n (%)	87 (86.1%)	41 (87.2%)	46 (85.2%)
Personality disorder n (%)	80 (79.2%)	38 (80.9%)	42 (77.8%)
WHODAS mean (SD)	47.2 (19.6)	50.5 (17.7)	44.3 (20.8)

Note *Sign difference (p = .016) between groups. **According to Bastiaens and Galus (2018).

Abbreviations: ITI = International Trauma Interview (test version 3.2). DES-II = Dissociative Experience Scale II. DSM-5 CC = DSM-5 Self-Rated Level 1 Cross-Cutting Symptom Measure. WHODAS = World Health Organisation Disability Assessment Schedule 2.0

Table 4.

Trauma exposure.

Traumatic experience, age span and frequency		Total (n = 101)	8-day ITP (n = 47)	TAU (n = 54)
Physical violence (direct exposure). n (% within-group)		82 (81.2%)	39 (83.0%)	43 (79.8%)
<13 years	1	2 (1.9%)	1 (2.1%)	1 (1.8%)
	2–5	16 (15.8%)	7 (14.9%)	9 (16.7%)
	>5	31 (30.1%)	16 (34.0%)	15 (27.8%)
13–18	1	6 (5.9%)	2 (4.2%)	4 (7.4%)
	2–5	11 (10.9%)	2 (4.2%)*	9 (16.7%)*
	>5	12 (11.8%)	9 (19.2%)*	3 (5.6%)*
>18	1	9 (8.9%)	3 (6.3%)	6 (11.1%)
	2–5	18 (17.8%)	10 (2.3%)	8 (14.8%)
	>5	12 (11.8%)	7 (14.9%)	5 (9.2%)
Sexual violence (direct exposure). n (% within-group)		87 (86.1%)	43 (91.5%)	44 (81.5%)
<13 years	1	8 (7.9%)	5 (10.6%)	3 (5.6%)
	2–5	18 (17.8%)	8 (17.0%)	10 (18.5%)
	>5	19 (18.8%)	12 (25.5%)	7 (12.9%)
13–18	1	12 (11.8%)	6 (12.7%)	6 (11.1%)
	2–5	14 (13.7%)	8 (17.0%)	6 (11.1%)
	>5	7 (6.9%)	5 (10.6%)	2 (3.7%)
>18	1	13 (12.8%)	4 (8.5%)	9 (16.7%)
	2–5	17 (16.8%)	11 (23.4%)	6 (16.7%)
	>5	12 (11.8%)	7 (14.9%)	5 (9.2%)
Other sexual abuse		42 (41.5%)	20 (42.6%)	22 (40.7%)
Weapon assault		40 (39.6%)	18 (38.3%)	22 (40.7%)
War (witness/exposure)		4 (3.9%)	1 (2.1%)	3 (5.6%)
Captivity		21 (20.8%)	11 (23.4%)	10 (18.5%)
Accident (car, fire)		61 (60.4%)	29 (61.7%)	32 (59.3%)
Life-threatening disease		17 (16.8%)	7 (14.9%)	10 (18.5%)
Other trauma		40 (39.6%)	16 (34.0%)	24 (44.4%)
Other sexual abuse		42 (41.5%)	20 (42.6%)	22 (40.7%)

Note * Sign difference (p = .019) between groups.

2.2. Preliminary effects of pretreatment psychoeducation

No difference was observed in baseline clinical measures between TAU participants who took part in pretreatment psychoeducation. No significant treatment effects could be observed on PCL-5 and ITQ due to the pretreatment psychoeducation in the TAU group. However, a small but significant increase in WHODAS score could be observed between pre- (M = 45.2, SD = 21.5) and post-psychoeducation (M = 50.2, SD = 22.2), t(31) = −2.67, p = .012, g = −.47.

2.3. Dropout analysis

Dropouts (n = 15) did not differ significantly from treatment completers on baseline clinical measures, frequency or age of exposure to physical or sexual violence. The reasons cited by participants were diverse: different treatments needed, somatic disease, acute stressors, private life or family problems and lack of motivation.

2.4. Within-group treatment effects

Both treatment groups showed a significant reduction in PTSD symptoms in both primary measures between pre- and posttreatment (see Table 5), with large within-group effect sizes. No significant within-group changes in symptoms were observed between posttreatment and 3-month follow-up; effect sizes were minor to small. On the measure of functional impairment, both treatment groups showed a significant reduction between pretreatment and 3-month follow-up, with a large within-group effect size for the ITP and a medium effect for TAU.

Table 5.

Estimated marginal means, effect sizes, and mixed linear model with fixed effects of time and time × treatment comparisons.

Measure		Estimated marginal means (SD)			Linear mixed model				Effect size Hedges' g
		Pre	Post	3 m FU	Pre-post  β (SE)	p	Post-3 m FU β (SE)	p	Pre-Post (95% CI low; high)	Post- 3mFU (95% CI low; high)
PCL-5
	ITP	57.8 (11.18)	27.5 (17.64)	26.3 (17.44)	−30.3 (2.59)	.000*	−1.23 (3.48)	.723	−2.67 (−3.28;−2.06)	−0.07 (−0.52; 0.38)
	TAU	53.2 (11.25)	25.9 (18.67)	23.2 (19.47)	−27.3 (2.49)	.000*	−2.65 (2.14)	.215	−2.43 (−3.02;−1.84)	−0.14 (−0.59; 0.31)
	ITP vs TAU	NA	NA	NA	3.04 (3.53)	.390	−1.41 (4.34)	.745	0.26 (−0.19; 0.71)	−0.07 (−0.52; 0.38)
ITQ
	ITP	34.6 (7.96)	20.4 (11.59)	16.9 (11.59)	−14.2 (1.76)	.001*	−3,56 (2.27)	.117	−1.75 (−2.28;−1.22)	−0.29 (−0.74;0.16)
	TAU	31.7 (7.94)	17.2 (12.72)	14.6 (13.52)	−14.5 (1.65)	.000*	−2.67 (1.61)	.100	−1.80 (−2.33;−1.27)	−0.20 (−065;0.25)
	ITP vs TAU	NA	NA	NA	0.34 (2.28)	.880	2.66 (2.05)	.812	0.03 (−0.42; 0.48)	0.26 (−0.19; 0.71)
		Pre		3 m FU			Pre-3 m FU β (SE)		Pre-3mFU
WHODAS
	ITP	0.50 (0.19)	NA	0.30 (0.22)	NA	NA	−0.20 (0.03)	.001**	−1.05 (−0.63;−1.47)
	TAU	0.44 (0.19)	NA	0.31 (0.24)	NA	NA	−0.01 (0.03)	.001**	−0.64 (−0.23;−1.05)
	ITP vs TAU	NA	NA	NA	NA	NA	0.08 (0.04)	.061	0.05 (−0.35; 0.45)

Note. * Significant at α = .025 (Bonferroni correction); **Significant at α = .05.

Abbreviations: Pre = pretreatment; Post = end of treatment (16 weeks for TAU, and 8 days for ITP); 3 m FU  = 3-month follow-up; ITP = Intensive treatment programme; TAU = Treatment as usual; NA, not applicable; SD = Standard deviation; β = unstandardised beta-coefficient; SE = Standard error; p = p-value; Hedges’ g = bias-corrected Hedges’ g; PCL-5 = Posttraumatic Stress Disorder Checklist for DSM-5; ITQ = International Trauma Questionnaire; WHODAS = World Health Organization Disability Assessment Schedule 2.0 12 item Self-Report (simple score).

2.5. Between-group treatment effects

No significant differences in baseline symptom scores were observed on either measure of PTSD symptoms or functional impairment (see Table 5). No significant between-group differences in the magnitude of change were observed at pre- to posttreatment or posttreatment to follow-up on PTSD symptoms. Effect sizes were minor to small. No significant between-group differences in the magnitude of change of functional impairment were observed from pretreatment to follow-up, and the effect size was minor.

2.6. Adverse effects

For the ITP, there was a small but statistically significant overall reduction in SSI scores, χ²(3) = 9.35, p = .025. Post hoc tests showed no significant differences in pairwise comparisons between time points. For TAU, there was no overall change, χ²(3) = 2.65, p = .448 and no significant differences in pairwise comparisons between time points (see appendix for descriptive data).

No adverse effects, such as suicidal behaviour or suicide attempts, occurred during the ITP or TAU.

2.7. Comparison of dropout rates between treatment and TAU groups

The proportion of dropouts in the ITP was 4.3%; in the TAU group, the proportion was 24.1%. A significant association was found between treatment group and dropout rates, χ²(1, n = 101) = 7.08, p = .005.

2.8. Reliable change and clinical significance

At the pretreatment assessment, a significant association was observed between the treatment group and the proportion of participants meeting CPTSD criteria based on ITI 3.2, χ²(1, n = 101) = 5.85, p = .026. In the ITP condition, 55.3% (n = 26) met the criteria and 31.5% (n = 17) in the TAU condition. At posttreatment, no significant association between treatment group and the proportion of participants meeting criteria of CPTSD, PTSD or no diagnosis was observed χ²(2, n = 85) = .295, p = .863. The total loss of PTSD diagnosis in ITP was 68.9% (n = 31) and 65% (n = 26) in the TAU. At follow-up, no significant association between treatment group and the proportion of participants meeting criteria of CPTSD, PTSD or no diagnosis was observed, χ²(2, n = 84) = .636, p = .728. The total loss of diagnosis in ITP was 73.3% (n = 33) and 74.4% (n = 29) in the TAU. Reliable change indexes (RCI) are presented in Table 6.

Table 6.

Reliable Change Index (RCI) PCL-5 and ITQ.

Posttreatment RCI				Follow-up RCI
	No change (%)	Improved (%)	Worsened (%)	No change (%)	Improved (%)	Worsened (%)
PCL-5a
ITPbc	22.2 (n = 10)	77.8 (n = 35)	0	0	100 (n = 43)	0
TAUde	24.4 (n = 10)	75.6 (n = 31)	0	8.6 (n = 3)	91.4 (n = 32)	0
ITQf
ITPgh	21.3 (n = 10)	74.5 (n = 35)	4.3 (n = 2)	16.7 (n = 7)	83.3 (n = 35)	2.1 (n = 1)
TAUij	17.5 (n = 7)	82.5 (n = 33)	0	17.6 (n = 6)	82.5 (n = 28)	0

^aBased on test-retest reliability and SD values in Sloan et al. (2018).

^bPosttreatment n = 45.

^cFollow-up n = 43.

^dPosttreatment n = 41.

^eFollow-up n = 35.

^fBased on test-retest reliability (alpha-based) and SD values in Cloitre et al. (2021).

^gPosttreatment n = 45.

^hFollow-up n = 42

ⁱPosttreatment n = 40.

^jFollow-up n = 34.

2.9. Protocol adherence

89% of all sessions were rated. The overall mean rating was 2.71 (SD 0.62). 70% of all rated sessions were reported as fully adherent to the treatment protocol. Sessions in the ITP group were rated as more adherent to the protocol than those in TAU, t(83) = 2.19, p = .033, Hedges’ g = 0.49, 95% CI [0.06, 0.91].

3. Discussion

The primary purpose of this study was to evaluate the effect of an 8-day ITP on PTSD symptoms in a population with CPTSD diagnosis, psychiatric comorbidity and previous unsuccessful treatment. The results suggest that both ITP and standard treatment are effective treatments in this sample, both in terms of PTSD symptom reduction and functional impairment. Furthermore, no meaningful differences were observed between the treatment groups. Thus, the results were consistent with the main hypothesis. We did not observe any increase in suicide ideation during and after either ITP or traditionally spaced treatment compared to baseline, and there were no adverse events, suggesting no difference regarding the safety of the treatments. Two individuals in the ITP group showed worsened ITQ scores at posttreatment and one at follow-up, whereas no such cases were observed with the PCL-5. Follow-up analysis of ITQ-PTSD and ITQ-DSO subscales indicated that this was not attributable to a specific subscale. As the ITQ performs comparably to the PCL-5 at group level (Cloitre et al., 2021), these discrepant cases likely reflect minor differences in scale properties and in the samples used to derive RCI thresholds. In our view, the most salient finding in our results was the difference in dropout levels between treatment groups, which raises an important question: how come two treatments with the same amount of trauma-focused psychotherapy but with different spacing of sessions show such a wide discrepancy in dropout rates? These findings are consistent with previous research (Sciarrino et al., 2020) but add one important piece of information: the therapist factor appears to be an implausible explanation, given that the same therapists were involved in both treatment options. Other differences between treatment formats remain besides the spacing of sessions. The ITP has an all-day schedule, including the patient hotel and physical activity between sessions, which could support emotion regulation and perhaps counteract session avoidance, although positive results of ITP have been shown without a patient hotel (Klaeth et al., 2024), and the specific role of the physical activity within an ITP is unclear (Voorendonk et al., 2023). A more daily life-related hypothesis is that during a 16-week treatment, there are more instances where life events can hinder the participants’ efforts to achieve symptom relief and improved quality of life. Our analyses could not explain the difference in dropout rates between the treatment groups, and the possibility that unmeasured variables could account for these results must be considered due to the lack of randomisation. Still, we can conclude that our findings are consistent with previous research findings regarding the positive effect of ITP on PTSD symptoms, that the intensity does not seem to provoke adverse effects and that dropout levels improve with ITP compared to traditionally spaced treatment despite equal amounts of trauma-focused therapy in both conditions (Ragsdale et al., 2020; Van Woudenberg et al., 2018). The 8-day ITP appears to be a viable alternative for psychiatric patients with PTSD or CPTSD, and it reduces duration considerably compared to a 16-week treatment. The study also provides insight into treatment effects in a real-world setting, where patients are engaged and supported in making their own choices about treatment methods.

We would also like to reflect on two aspects of the interventions and the study setup. First, while the effect of therapist rotation as opposed to receiving treatment from one single therapist was not explicitly measured, the results ask questions about what therapeutic alliance means, given that a treatment where the participant processes painful, traumatic memories with several different therapists is not inferior to when the same process is done with a single therapist. Qualitative studies have shown that some participants find the therapist rotation in ITP to be overall positive, while others may form a closer bond with specific therapists (Vaage-Kowalzik et al., 2024). While we can conclude that the therapist rotation does not necessarily compromise treatment effects on group level, there is still much to learn about the therapeutic alliance within the ITP and whether it is qualitatively different from traditional treatment. Additionally, while the small effect on protocol adherence ratings should not be over-interpreted, it aligns with one of the main hypothesised advantages of therapist rotation: to increase the focus on processing traumatic memories (Van Minnen et al., 2018). Secondly, we note that there are more participants diagnosed with CPTSD who chose ITP compared to TAU. Thus, having a presumably worse condition did not necessarily deter an individual from choosing the ITP in the shared decision-making process. While we cannot explain the imbalance within the scope of this study, we suggest that the role of participants being assigned a preferred treatment choice should be further explored within the context of ITP and traditionally spaced treatment, especially given the disparities between these options. Meta-analyses have demonstrated that receiving a preferred treatment impacts outcomes and dropout rates across various treatment types (Swift et al., 2018). However, due to the design of this study, we cannot conclude whether the observed differences – particularly dropout rates – are actual treatment effects and whether dropout is moderated differently by preferred treatment assignment in TAU and ITP.

3.1. Limitations

First and foremost, the study design lacks randomisation, which limits the possibility of concluding treatment effects. Unmeasured variables that introduce bias must be considered a possible confounding effect. Also, despite efforts to match the ‘dose’ of trauma-focused treatment between conditions, there are still additional differences that could contribute to treatment effects. Furthermore, while symptom levels, suicidal ideation and functional impairment were assessed, other psychological aspects of well-being or negative experiences were not measured and thus not fully explored. While the demography and clinical features of our sample are similar to other ITP studies (Voorendonk et al., 2020), it should be emphasised that the sample may not be generalisable to populations with a different composition of age, gender, and traumatic experiences. Finally, the categorical nature of the data we collected on trauma exposure does not allow for an indication of overall trauma burden in terms of mean and standard deviation, which limits the possibilities to compare with other studies and further analysis.

4. Conclusions and directions for future research

While the lack of randomisation is a limitation, the study gives ecologically valid insight into the effect of ITP and TAU in the treatment of PTSD and CPTSD in psychiatric outpatient settings, where patients have the option of either intensive or traditionally spaced treatment. The results support previous research in suggesting that ITP is a viable treatment option also for patients with severe traumatic exposure and psychiatric comorbidity. Additionally, the results align with previous studies reporting lower dropout rates in ITP compared to traditionally spaced treatment. Studies with random allocation of treatments are recommended to bring further knowledge on treatment effects. The role of the therapist rotation and the treatment preferences of clients should be explored, and qualitative studies, along with an in-depth understanding of the client experience, are of great importance. Finally, there is a need to explore treatment mechanisms and individual characteristics among clients who experience little or no improvement from either ITP or traditionally spaced treatment.

Appendices.

Appendix 1

Table A1.

SSI items 4 and 5.

	Pretreatment			Mid-treatment			Posttreatment			Follow-up
	Mean Rank	M	SD	Mean Rank	M	SD	Mean Rank	M	SD	Mean Rank	M	SD
ITP	2.68	0.4	.99	2,44	0.1	0.37	2.48	0.1	0.61	2.40	0.0	0.3
TAU	2.51	0.2	0.71	2,46	0.1	0.34	2.59	0.2	.053	2.44	0.0	0.16

Abbreviations: SSI = Beck’s Scale Suicide Ideation.

Table A2.

Missing data outcome variables (Pretreatment sample n = 101).

Variables	Posttreatment % (n)	Follow-up % (n)
PCL-5	14.9% (15)	22.8% (23)
ITQ	15.8% (16)	23.8% (24)
WHODAS	a	24.8% (25)
ITI	15.8% (16)	16.8% (17)
SSI	15.8% (16)	16.8% (17)

^a Not measured at Posttreatment.

Abbreviations: ITI = International Trauma Interview (test version 3.2), SSI = Beck’s Scale Suicide Ideation, PCL-5 = Post Traumatic Checklist for DSM −5, ITQ = International Trauma Questionnaire, WHODAS = World Health Organization Disability Assessment Schedule 2.0.

Table A3.

Multiple imputation (SPSS ver. 29.0).

Method	Fully Conditional Specification (FCS)
Number of imputations	50
Imputation model	Predictors		Imputed
		PCL-5a		PCL-5bc
		ITQa		ITQbc
		DES II		WHODASc
		CPTSD diagnosisa
		Age
		Sex
		Group

^a Pretreatment, ^bPosttreatment, ^cFollow-up.

Abbreviations: PCL-5 = Post Traumatic Checklist for DSM -5, ITQ = International Trauma Questionnaire, DES-II = Dissociative Experience Scale II, WHODAS = World Health Organization Disability Assessment Schedule 2.0

Table A4.

Reliable change index (RCI) ITQ-PTSD and DSO.

Posttreatment RCI					Follow-up RCI
		No change %	Improved %	Worsened %		No change %	Improved %	Worsened %
ITQ-PTSDa
	ITPbc	24.4 (n = 11)	71.1 (n = 32)	4.3 (n = 2)		6.9 (n = 3)	90.4 (n = 38)	2.1 (n = 1)
	TAUde	22.5 (n = 9)	77.5 (n = 31)	0		20.6 (n = 7)	79.4 (n = 27)	0
ITQ-DSOa
	ITPbc	35.3 (n = 15)	66.6 (n = 28)	4.3 (n = 2)		33.3 (n = 14)	64.2 (n = 27)	2.1 (n = 1)
	TAUde	37.5 (n = 15)	62.5 (n = 25)	0		35.2 (n = 12)	64.7 (n = 22)	0

^a Based on test-retest reliability (alpha-based) and SD values in Cloitre et al. (2021)

^b Posttreatment n = 45.

^c Follow-up n = 42.

^d Posttreatment n = 40.

^e Follow-up n = 34.

Appendix 2

Sensitivity analysis

A complete cases-only analysis was performed on both primary and secondary measures (Table A4). At posttreatment ITP8 n = 45 and TAU n = 41. At follow-up, ITP8 n = 42 and TAU n = 35. Results were adjudged comparable with only minor differences in beta-coefficients and effect sizes and no observed reversal in direction of effects.

Table A5.

Estimated marginal means, effect sizes, and mixed linear model with fixed effects of time and time × treatment comparisons. Complete cases only.

Measure		Estimated marginal means (SD)			Linear mixed model				Effect size Hedges' g
		Pre	Post	3 m FU	Pre-post β (SE)	p	Post-3 m FU β (SE)	p	Pre-Post (95% CI low; high)	Post- 3mFU (95% CI low; high)
PCL-5
	ITP	57.7 (10.92)	27.8 (17.29)	26.3 (17.17)	−31.3 (2.69)	.001 *	−1.56 (3.55)	.661	−1.70 (−2.26;−1.14)	−0.06 (−0.56;0.44)
	TAU	51.5 (10.91)	24.7 (17.30)	22.4 (17.70)	−29.0 (2.4)	.001*	−2.24 (1.62)	.182	−1.92 (−2.50;−1.34)	−0.22 (−0.72;0.28)
	ITP vs TAU	NA	NA	NA	3.05 (3.70)	.407	−0.01 (4.72)	.999	0.18 (−0.30; 0.66)	−0.01 (−0.51; 0.49)
ITQ
	ITP	34.7 (7.89)	20.7 (11.32)	16.9 (11.48)	−17.8 (1.77)	.001*	−3.82 (2.36)	.110	−1.47 (−2.01;−0.93)	−0.24 (−0.74;0.26)
	TAU	30.4 (8.00)	15.6 (11.46)	13.8 (11.96)	−16.6 (1.48)	.001*	−1.93 (1.15)	.115	−1.78 (−2.35;.1.21)	−0.26 (−076;0.24)
	ITP vs TAU	NA	NA	NA	0.71 (2.31)	.757	1.99 (2.98)	.506	0.07 (−0.41; 0.55)	0.15 (−0.35; 0.65)
		Pre		3 m FU			Pre-3 m FU β (SE)		Pre - 3mFU
WHODAS
	ITP	0.51 (0.18)	NA	0.31 (0.22)	NA	NA	−0.20 (0.03)	.001 **	−0.98 (−0.51;−1.45)
	TAU	0.42 (0.17)	NA	0.31 (0.21)	NA	NA	−0.11 (0.28)	.001 **	−0.62 (−1.08;−0.16)
	ITP vs TAU	NA	NA	NA	NA	NA	0.09 (0.04)	.02 **	0.50 (0.04; 0.96)

Note: *Significant at alpha level p .025 (Bonferroni correction); **Significant at alpha level .05.

Abbreviations: Pre = pretreatment; Post = end of treatment (16 weeks for TAU, and 8 days for); 3 m FU = three month follow-up; ITP = Intensive treatment programme; TAU = Treatment as usual; NA, not applicable; SD = Standard deviation; β = unstandardized beta-coefficient; SE = Standard error; p = p-value; Hedges’ g = bias-corrected Hedges’ g; PCL-5 = Posttraumatic Stress Disorder Checklist for DSM-5; ITQ = International Trauma Questionnaire; WHODAS = World Health Organization Disability Assessment Schedule 2.0 12 item Self-Report (simple score).

To check for the influence of baseline pretreatment CPTSD diagnosis imbalance on the results, we tested the interaction of CPTSD by time within the LMM model for the primary analysis (Table A6). We found no interactions between CPTSD diagnosis and time at any measure of PTSD, indicating that the baseline imbalance in CPTSD diagnosis did not significantly influence the primary findings.

Table A6.

Estimated mixed linear model with fixed effects of CPTSD × time.

	Pretreatment			Pre-post difference			Post- 3mFU difference
	β (SE)	p	95% CI low; high	β (SE)	p	95% CI low; high	β (SE)	p	95% CI low; high
PCL-5	2.05 (3.29)	.53	−4.40 to 8.51	4.01 (5.45)	.45	−6.58 to 14.7	2.82 (5.51)	.61	−13.6 to 7.99
ITQ	2.58 (2.24)	.25	−1.88 to 7.05	2.55 (3.61)	.39	−4.62 to 9.72	−1.91 (4.16)	.65	−6.37 to 10.2

Abbreviations: Pre = pretreatment; Post = end of treatment (16 weeks for TAU, and 8 days for); 3 m FU = three month follow-up; β = unstandardized beta-coefficient; SE = Standard error; p = p-value.

PCL-5 = Posttraumatic Stress Disorder Checklist for DSM-5; ITQ = International Trauma Questionnaire; CPTSD = Complex Posttraumatic Stress Disorder.

Funding Statement

This work was supported by the Region Västra Götaland, Södra Älvsborg Hospital, Department of Research, Education and Innovation, Borås, Sweden; under Grant sas-995842, sas-1010858; Research and development unit Södra Älvsborg [grant VGFOUSA-963589, VGFOUSA-1003831].

Author contributions

All authors – HG, PFGC, SW and CB – equally contributed to the conception and design of the project. CB was the project leader. HG performed the statistical analysis. HG wrote the first draft of the manuscript. All authors contributed to the revision and read and approved the submitted version.

Disclosure statement

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

Data Availability Statement

The ethics approval of the current project does not support sharing of data. Upon request, the possibility of sharing data with ethics approval will be reviewed.

Generative AI statement

ChatGPT version o3 has been used solely for these purposes: to improve coding and identify coding errors and to discuss and expand knowledge in statistical methods. The sources behind the generated information have been verified, and decisions on methods have been made by the authors.