More is Not Always Better: 2 Weeks of Intensive Cognitive Processing Therapy-Based Treatment Are Non-Inferior to 3 Weeks

Abstract

Objective:

Although there is mounting evidence that massed treatment for PTSD is both feasible and effective, many questions remain about the optimal length of intensive treatment programs (ITPs), as well as the role of adjunctive services, such as psychoeducation, mindfulness, and yoga. Our setting recently transitioned from a 3-week ITP to a 2-week program. Adjunctive services were reduced, but the amount of individual CPT between programs remained similar. The present study examined the effectiveness of a 2-week ITP based on twice daily individual CPT sessions and evaluated the program’s non-inferiority to an established 3-week ITP using a Bayesian analytical approach.

Method:

Bayesian linear mixed regression models were used to explore PTSD and depression changes over time, as well as predictors of change. Non-inferiority of the 2-week ITP to a 3-week ITP was also established using a Bayes factor approach.

Results:

Results indicate that program participants change meaningfully in both PTSD and depression severity over the course of treatment, and that changes in post-traumatic cognitions predict subsequent changes in these outcomes. Further, the 2-week ITP can be considered non-inferior to the 3-week ITP in both clinical outcomes and overall satisfaction.

Conclusions:

In the context of intensive PTSD treatment, the content of the ITP appears to matter more than its overall length. Shorter programs have the potential to increase access and treatment capacity. Our findings demonstrate the importance of continuous and rigorous program evaluation. Limitations as well as future directions for research, such as identifying the most effective treatment components, are discussed.

In intensive (also referred to as massed) treatments, which generally involve one or more sessions of evidence-based PTSD treatment per day at least three days per week, entire courses of evidence-based interventions for PTSD are delivered in a matter of days or weeks (Held et al., 2019; Sciarrino et al., 2020). In these formats, evidence-based treatments can be delivered intensively as standalone interventions or as part of intensive PTSD treatment programs (ITPs), which generally combine evidence-based PTSD treatment with adjunctive services such as psychoeducation, skills groups, and wellness services (Harvey et al., 2017; Held et al., 2019; Sciarrino et al., 2020). Whether delivered standalone or as ITPs, intensive treatments have been shown to be associated with large PTSD and depression symptom reductions in civilians and military veterans, that are comparable to evidence-based treatments delivered in the traditional weekly format and can be maintained long-term (Beidel et al., 2017; Bryan et al., 2018; Ehlers et al., 2014; Foa et al., 2018; Held, Zalta, et al., 2020; Oprel et al., 2021; Zalta et al., 2018). Intensive PTSD treatments have also been shown to be robust and effective for a wide range of traumatic experiences, including military sexual trauma (MST; Lofgreen et al., 2020) and complex trauma (Tirone et al., 2020; Voorendonk et al., 2020), among others. Intensive treatments are generally associated with lower dropout (~5–10%) compared to weekly treatments (~40–60%). By increasing the number of individuals who receive adequate doses of treatment, intensive treatments have the potential to substantially reduce the risk for negative long-term mental and physical health outcomes that have been associated with untreated PTSD.

Intensive PTSD treatments appear to produce PTSD and depression symptom change via similar mechanisms as their weekly counterparts. Studies on intensive PTSD treatments based on Cognitive Processing Therapy (CPT; Resick, Monson, et al., 2017), a well-established evidence-based trauma-focused cognitive behavioral therapy that helps individuals identify and subsequently restructure beliefs about themselves, others, and the world, have shown that reductions in negative posttrauma cognitions predict greater symptom reduction over the course of treatment (Zalta et al., 2018). Moreover, changes in negative posttrauma cognitions have also been shown to be predictive of maintained treatment gains, with larger reductions in negative cognitions over the course of the ITP predicting lower PTSD and depression symptoms following treatment (Held, Zalta, et al., 2020). Thus, given their established feasibility, efficacy, and effectiveness, as well as the ability for rapid change in negative posttrauma cognitions, intensive treatments are increasingly considered as a viable treatment option for individuals with PTSD (Held et al., 2019; Sciarrino et al., 2020).

We have previously demonstrated the feasibility and effectiveness of our 3-week CPT-based ITP for veterans with PTSD (Held, Steigerwald, et al., 2020; Held, Zalta, et al., 2020; Zalta et al., 2018). Based on mounting research showing that shorter programming can be effective (Bryan et al., 2018; Ehlers et al., 2014; Foa et al., 2018; Oprel et al., 2021), our group systematically changed the established 3-week ITP into a 2-week ITP in two ways: (1) the amount of group-based CPT was reduced relative to individual CPT, and (2) time was given in the daily clinical schedule for completion of assigned CPT homework during the day, which did not exist in the 3-week ITP. Evidence from trials showing favorable outcomes for individual CPT relative to group CPT (Resick, Wachen, et al., 2017), the importance of homework (Stirman et al., 2018), as well as preliminary insights from an ongoing clinical trial of twice daily individual CPT (Held, Klassen, Small, et al., 2020) provided rationale for programmatic changes. These changes translated to a 64% reduced total clinical programming time (from 104 hours in the 3-week ITP to 67 hours in the 2-week ITP), while keeping the total time spent on individual CPT nearly identical between the 3- and 2-week ITPs (13 hours and 16 hours, respectively).

Shortening the program duration can have important feasibility implications, such as increased access. Shorter programs may allow individuals to attend who would otherwise not be able to due to other obligations, including work and family responsibilities. Moreover, in a cohort-based model wait times may be reduced, making treatment timelier. By shortening the program, more individuals can be treated over the course of a year. Shorter programs can also be associated with decreased cost for payors, making treatment more financially feasible. An often-held assumption is that more programming (i.e., combining multiple interventions and thereby increasing the total treatment duration) may lead to better outcomes. Longer programs may take an eclectic approach towards interventions, offering several modalities because it is often unclear what intervention will best help the patient. Furthermore, longer programs may include separate interventions to address other comorbidities or areas of concern such as communication, substance use, or physical health. As a result, programs may offer all-day programming over the course of several weeks, often without knowing which components are the most important drivers of improved clinical outcomes. In the case of our 3-week ITP for veterans with PTSD, this involved combining a full course of daily individual CPT with daily group CPT sessions, daily mindfulness and yoga groups, and various daily psychoeducational groups (Zalta et al., 2018). Although veterans were highly satisfied with the 3-week ITP, questions remained whether a shortened 2-week ITP based primarily on individual CPT with reduced adjunctive services could be equally as effective, well-liked, and ultimately increase treatment accessibility and financial feasibility.

Given the previously cited evidence that briefer PTSD treatments are feasible and effective, the present study had four primary goals: 1) determine the effectiveness of a novel 2-week ITP centered around twice daily individual CPT sessions for reducing PTSD and depression symptoms, 2) examine whether the 2-week ITP was equally effective (i.e., non-inferior) to our previously established 3-week ITP centered around daily individual and group CPT sessions, 3) evaluate whether reductions in negative posttrauma cognitions over the course of treatment, which are a well-established predictor of outcomes in the existing 3-week ITP, remained an important predictor of PTSD and depression symptoms in the 2-week ITP, and finally (4) examine whether satisfaction with the briefer 2-week ITP was comparable (i.e. non-inferior) to the established 3-week ITP. Based on prior research on exposure-based and CPT-based intensive PTSD treatments delivered over the course of two weeks (Bryan et al., 2018; Foa et al., 2018; Oprel et al., 2021), our own case series showing large effects associated with a similarly structured virtual 2-week ITP (Held, Klassen, Coleman, et al., 2020), as well as the fact that participants would receive a similar amount of individual PTSD-focused treatment sessions (3-week: 13 total CPT sessions; 2-week: 16 total CPT sessions), we expected that the 2-week ITP would produce large effects and be non-inferior at endpoint to the 3-week ITP despite its shorter duration. Further, based on prior research (Lofgreen et al., 2020), we also did not expect to find differences by cohort type (MST vs. combat trauma) in the 2-week ITP. Given the robust literature supporting reductions of negative posttrauma cognitions as a predictor of improved treatment outcomes (Brown et al., 2019; Zalta, 2015), we anticipated that changes in negative posttrauma cognitions would emerge as a predictor of outcomes in the 2-week ITP similar to what our team has previously shown for the 3-week ITP (Zalta et al., 2018). Finally, we also anticipated patient satisfaction to remain equally high for each program. We utilized a Bayesian approach to test the aforementioned research questions, which allowed us to incorporate existing knowledge about the outcomes and patient satisfaction of our 3-week ITP to set priors for the present analyses. Moreover, a recently developed Bayes Factor approach lends itself to non-inferiority testing as it makes results generally more interpretable compared to frequentist approaches.

Method

Participants

Data for the present study were collected from 201 veterans who participated in a 2-week in-person ITP at the Road Home Program at Rush University Medical Center in Chicago, Illinois. The 2-week ITP was started in June 2020 and data for this study were collected until August 2021. Individuals participated in cohorts of up to 14 individuals at a time. Cohorts were offered consecutively every two weeks. Co-ed combat trauma (n = 99) and MST cohorts (n = 102) were alternated. On average, participants in the 2-week ITP sample were 42.64 years old (SD = 9.12, range = 26 – 71 years), 50.00% male, 18.00% identified as Latinx, and 65.00% identified as White. A total of 85.65% (n = 172) of those who started the ITP completed treatment. Non-completers who started the program completed an average number of 5.07 treatment days (SD = 1.98).

The previously established 3-week ITP, to which the 2-week ITP was compared, was offered from April 2016 to March 2020. A total of 502 veterans were treated via this program and were included in the analyses. A new cohort of up to 24 individuals started approximately once per month. Most cohorts were combat trauma-focused (n = 330); approximately every fourth co-ed cohort was dedicated to MST (n = 172). On average, participants in the 3-week ITP sample were 41.35 years old (SD = 9.43, range = 24 – 74), 65.94% male, 20.12% identified as Latinx, and 67.33% identified as White. A total of 93.14% (n = 468) of those who started the 3-week ITP completed treatment. Non-completers who started the program completed an average number of 7.64 treatment days (SD = 2.87). Demographic characteristics can be found in Table 1.

Table 1.

Demographic characteristics by program and cohort type.

	Overall (N = 201)		2-Week ITP Combat (n = 99)		MST (n = 102)		Overall (N = 502)		3-Week ITP Combat (n = 330)		MST (n = 172)
	N	%	n	%	n	%	N	%	n	%	n	%
Sex
Male	101	50.25	88	88.89	11	10.78	331	65.94	306	92.72	25	14.53
Female	100	49.75	8	8.00	92	92.00	171	34.06	24	7.27	147	85.47
Ethnicity
Not Latinx	165	82.09	76	76.77	89	87.25	401	79.88	259	78.48	142	82.56
Race
American Indian/	3	1.49	0	0.00	3	2.94	10	1.99	4	1.12	6	3.49
Alaskan Native
Asian	5	2.49	1	1.01	4	3.92	6	1.20	4	1.12	2	1.16
Black or African	44	21.89	12	12.12	32	31.37	101	20.12	52	15.76	49	28.49
American
Native Hawaiian/	3	1.49	1	1.01	2	1.96	3	0.60	1	0.30	2	1.16
Pacific Islander
Other	16	7.96	9	9.09	7	6.86	42	8.37	25	7.58	17	9.88
Refusal	1	0.50	1	1.01	0	0.00	1	0.20	1	0.30	0	0.00
Unknown	1	0.50	0	0.00	1	0.98	1	0.20	1	0.30	0	0.00
White	128	63.68	75	75.76	53	51.96	388	67.33	242	73.33	96	55.81
Military Service Branch
Air Force	21	10.45	5	5.05	16	15.69	42	8.37	15	4.55	27	15.70
Army	110	54.73	57	57.58	53	51.96	333	66.33	239	72.42	94	54.65
Coast Guard	2	1.00	2	2.02	0	0.00	4	0.80	1	0.30	3	1.74
Marines	36	17.91	24	24.24	12	11.76	75	14.94	57	17.27	18	10.47
Navy	32	15.92	11	11.11	21	20.59	48	9.56	18	5.45	30	17.44
Service Era
Post September 11, 2001	147	73.13	93	93.94	54	52.94	451	89.84	317	69.06	134	77.91
Deployed
Yes	150	74.63	97	97.98	53	51.96	395	78.69	320	96.97	75	43.60
	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD
Age	42.64	9.12	41.07	8.11	44.14	9.79	41.35	9.43	40.85	10.72	42.32	8.66
PCL-5 Baseline1	53.70	14.21	53.25	14.44	53.87	14.00	55.76	12.22	55.55	11.84	56.18	12.96
PCL-5 Post-Treatment	35.05	16.19	33.43	14.57	36.50	17.57	34.12	19.15	32.99	18.88	36.37	19.55
PHQ-9 Baseline1	16.15	5.46	16.91	5.23	15.31	5.62	17.58	5.02	17.69	4.87	17.35	5.33
PHQ-9 Post-Treatment	11.66	5.35	11.51	5.43	11.83	5.36	11.97	6.16	11.77	6.10	12.37	6.28
PTCI Baseline	144.31	37.10	137.13	38.35	151.76	34.89	146.04	39.39	142.3	34.37	153.29	39.13
PTCI Post-Treatment	121.21	39.82	116.61	40.72	125.67	39.14	115.3	48.42	113.10	46.25	119.40	52.27

Note:

¹Baseline differences in both outcomes between the 2- and 3-week programs constituted small (d < 0.3) effect sizes. ITP: Intensive PTSD Treatment Program. MST: Military Sexual Trauma. PCL-5: PTSD Checklist for DSM-5. PHQ-9: Patient Health Questionnaire. PTCI: Posttraumatic Cognitions Inventory.

Procedures

The study procedures were approved by the Institutional Review Board at Rush University Medical Center with a waiver of consent as all assessments were collected as a part of routine care. Interested veterans completed an intake with a clinician which involved a biopsychosocial interview, a formal assessment/confirmation of PTSD via the Clinician-Administered PTSD Scale for DSM-5 (Weathers et al., 2018), and the completion of self-report measures to determine program eligibility and appropriateness. A confirmed PTSD diagnosis was required for program acceptance. Veterans were not accepted into the ITPs if they had unstable housing, inability to independently complete activities of daily living, a suicide attempt in the last 30-days, untreated psychosis or mania, or alcohol or other drug dependence. The intake process and exclusion criteria were identical for both ITPs.

The ITPs

The ITPs were designed around CPT. Whereas the 3-week ITP combined daily individual CPT with daily group CPT, the 2-week ITP focused on twice daily individual CPT. In both programs CPT was delivered by licensed masters or doctoral-level clinicians who completed the official 2-day CPT training and the associated consultation requirements or advanced trainees under the supervision of a licensed clinician. In each ITP, the standard CPT protocol was followed. In line with the flexible delivery approach, additional sessions beyond the standard 12-session protocol continued the format of themed sessions and were used to address remaining stuck points. CPT homework was modified given the intensive format. Individuals were asked to complete three worksheets after each session instead of the standard one worksheet per day until the next session in weekly CPT. CPT was complemented by adjunctive interventions, such as yoga and mindfulness, art therapy, skill building groups, and psychoeducation (see more detailed description below) delivered by certified providers. Assignment to cohort type (combat versus MST) was collaboratively discussed with the veteran, and made based on the index trauma that the veteran chose to work on during CPT. The ITPs are considered outpatient programs; although veterans are housed near the ITP, the housing facilities are not associated with the ITP and veterans are not monitored outside of treatment hours.

2-Week ITP.

The novel 2-week ITP was designed around 50-minute individual CPT sessions delivered twice per day (16 total CPT sessions), each of which was followed by 60 minutes dedicated to complete assigned CPT homework. This schedule was based on promising insights from on ongoing clinical trial of twice-daily individual CPT with dedicated homework time (Held, Klassen, Small, et al., 2020). Participants received thirteen hours of combined mindfulness and yoga, as well as five and a half total hours of art therapy, two hours of group psychoeducation, seven-and-a-half hours of group CPT or Dialectical Behavior Therapy skill-based groups, as well as five hours of case management services. Participants also had the opportunity, if needed, to receive up to six individual sessions focused on coping strategies, spiritual care, or substance use support. Furthermore, veterans could participate in six sessions of acupuncture, or two hours of consultation with physicians for concerns related to medication, traumatic brain injury or pelvic floor functioning. In total, considering all available optional or as-needed services, veterans were able to receive up to 67 hours of clinical programming in the 2-week ITP. Compared to the 3-week ITP, group CPT was reduced by 14 hours (82%), mindfulness by 8 hours (48%), yoga by 6 hours (60%), psychoeducation by 17 hours (89%). An additional six hours of programming was eliminated through reducing other non-core elements such as a lengthy program orientation, case management, or presentations from outside veteran service organizations. A sample schedule for the 2-week ITPs is shown in Supplemental Figure 1.

3-Week ITP.

The 3-week ITP included one session of individual CPT lasting 50 minutes each day. Participants had 13 120-minute sessions of group CPT, four 60-minute sessions of art therapy, 13 75-minute sessions of mindfulness, and 12 50-minute sessions of yoga. In addition, veterans were offered 19 hours of educational classes on a variety of topics focused on PTSD symptomology or general health, such as healthy cooking, communication, substance use education and sleep habits. In total, veterans received up to 104 hours of clinical programming in the 3-week ITP. A sample schedule for the 3-week ITP is shown in Supplemental Figure 2.

Measures

Demographic Characteristics.

At intake, participants reported various demographic and military specific characteristics. These included age, sex, ethnicity, race, military service, branch, whether they served before or after September 11, 2001, and whether they deployed.

PTSD Checklist for DSM-5 (Bovin et al., 2016).

The PCL-5 is a 20-item self-report measure of PTSD symptom severity based on the DSM-5 diagnostic criteria. Total scores range from 0 to 80 with higher scores indicating more severe PTSD symptoms. At intake, veterans reported their symptoms for the past month. During treatment, past week severity was reported. In the 2-week ITP, the PCL-5 was administered at intake and on days 1, 3, 5, 6, 8, and 10. In the 3-week ITP, the PCL-5 was administered at intake and on days 2, 3, 5, 6, 8, 10, 11, 13, and 14. Cronbach’s alpha ranged from .92 - .94 in the 2-week and from .89 - .96 in the 3-week ITP.

Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001).

The PHQ-9 is a 9-item self-report measure which assesses depression symptoms during the past two weeks. Scores range from 0–27, with higher scores indicating more severe symptoms of depression. In the 2-week ITP, the PHQ-9 was administered at intake and on days 1, 3, 5, 6, 8, and 10. During the 3-week ITP, the PHQ-9 was administered at intake and days 2, 3, 5, 6, 8, 10, 11, 13, and 15. Cronbach’s alpha ranged from .79 - .84 in the 2-week and from .81 - .89 in the 3-week ITP.

Posttraumatic Cognitions Inventory (PTCI; Foa et al., 1999).

The PTCI is a 33-item self-report measure that assesses trauma-related cognitions. Total scores range from 33 to 231 with higher scores indicating stronger negative beliefs. During the 2-week ITP, this measure was given at intake and on days 2, 4, 7, and 9. In the 3-week ITP, the PTCI was given at intake and on days 2, 4, 7, 9, 12, 14, and 15. Cronbach’s alpha ranged from .95 - .97 in the 2-week and from .95 - .98 in the 3-week ITP.

Satisfaction Surveys.

Following treatment completion, veterans were asked to complete an anonymous satisfaction survey to evaluate the ITP. This survey asked “Overall, I feel satisfied with the clinical care I received” and “The care I received has improved the problem(s) I needed help with” using a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree).

Statistical Approach

A Bayesian longitudinal mixed models was utilized to examine change in the 2-week ITP. Mixed effects models are particularly appropriate for longitudinal data analysis due to their ability to model individual change over time and to account for the inherent clustering across individuals’ measurements. The Bayesian approach to mixed effects models allows for incorporation of prior knowledge in modeling, which may vary depending on the strength and quality of prior evidence. Additionally, Bayesian modeling also has the benefit of generating a full simulated posterior distributions of estimated parameters, as well as intuitive and substantive assessment of uncertainty in models. Models examined linear and quadratic time, as well as age, sex, race, ethnicity, cohort type, and PTCI. Because PTCI is a time-varying covariate that is not available at all program timepoints, it was included in models following initial examination of time trends and baseline predictors. The use of a Bayesian approach requires that priors were chosen, representing existing knowledge about parameters of interest. Weakly informative priors were obtained via simulation based on earlier reports from 3-week ITP, with normal priors for fixed and random effects parameters and truncated and half normal priors for random effects variance and overall variance parameters. Utilization of results from the 3-week ITP informed priors for PTCI and time effects, given their past significance in modeling. Other baseline covariates were given more vague priors based on their scale (e.g., ~ N(0, 5) for age) given past research illustrating their lack of relevance to outcomes in the program (Zalta et al., 2018). Sensitivity analysis examining model results when using different priors, including non-informative priors was also conducted to ensure results were robust to specification of priors.¹ This included leave-one-out cross validation to assess predictive accuracy and likelihood of various priors for both time PTSD and depression time trends and PTCI. Prior predictive checks across 1000 simulated datasets were explored to ensure alignment of priors with appropriate values. Hamiltonian Markov Chain Monte Carlo (MCMC) was utilized for sampling from the posterior via the No-U-Turn Sampler algorithm (NUTS). Five chains were used with 1000 burn in iterations and 5000 inference iterations. Posterior distributions were summarized via means, medians, and highest density intervals (HDI), which indicate the most credible values along the posterior distribution. Trace, autocorrelation, and divergence plots were examined to assess the posterior, and visual posterior predictive checks of mean, median, variance, and skewness parameters as well as Leave One Out Probability Integral Transform (LOO-PIT) were utilized in model checking as generally recommended (e.g., Gabry et al., 2019).

To further explore whether the 2-week ITP could be probabilistically determined to be non-inferior to the 3-week ITP we explored non-inferiority tests of endpoint PCL-5 and PHQ-9 total scores, as well as program satisfaction, between the two programs using Bayes Factors as recently proposed (van Ravenzwaaij et al., 2019). Unlike frequentist or confidence interval approaches that may reach different conclusions regarding equivalence of groups based on whether superiority or non-inferiority is being tested, this approach quantifies the probability of a non-inferiority hypothesis in comparison to an inferiority hypothesis based on existing data. Fairly conservative 5-point non-inferiority margins were set for PCL-5 (PTSD severity), 3-point non-inferiority margins for PHQ-9 (depression severity) at program endpoint, and 0.5-point margins for program satisfaction scores. Completion rates were not compared inferentially due to substantive program differences regarding handling of dropout between the 3- and 2-week ITP. Because the 2-week model began and has exclusively operated during the COVID-19 pandemic, program leadership has taken a more lenient approach to participant drop out, in part to accommodate pandemic related stressors (e.g., personal and family health concerns) as well as participant frustration regarding the strict enforcement of COVID protocols (e.g., masks, distancing, etc.). Bayesian linear mixed models and model checks were explored using PYMC3 in Python version 3.9.0, and non-inferiority tests and figures were conducted using R version 4.1.1.

Results

Deviance Information Criterion (DIC) and Widely-Applicable Information Criterion (WAIC) analysis supported the superiority of including intercepts and random slopes for time for all mixed models, so results reported here reflect these parameterizations. Maximum Gelman-Rubin Rc for models was below 1.02 for all models, indicating effective convergence of MCMC chains, and prior and posterior predictive checks and diagnostics indicated appropriate model and posterior sampling with lack of divergences. Examination of effective sample size and autocorrelation also indicated lack of problematic autocorrelation. Sensitivity analyses supported the predictive accuracy of using weakly informative priors, but indicated that all results were generally robust to choice of priors, including non-informative priors (see Supplementary Table 1).

Overall symptom change over the course of the 2-week ITP reflected large effect sizes for both PCL-5 (d = 1.16) and PHQ-9 (d = 0.80) during the program that were comparable to those previously seen in the 3-week program (see Table 2). These improvements during the program were also consistent across those in combat trauma and MST cohorts (see Figures 1 and 2). Based on posterior samples in the current models, the probability of differences over time between MST and combat trauma cohorts constituting a moderate (d = 0.5) or larger effect size was less than 1% for both PTSD and depression severity (see Figures 1 and 2). Additionally, non-inferiority tests of endpoint differences suggested that the hypothesis of non-inferiority of the MST cohort is 2.62 times more probable than that of superiority of the combat trauma cohort for PTSD severity, and 3e⁺⁷ times greater than the superiority hypothesis for depression severity.

Table 2.

Effect sizes for pre-post program differences.

		PCL-5			PHQ-9
Program	Overall	Combat	MST	Overall	Combat	MST
3-week	1.30	1.39	1.13	0.98	1.06	0.83
2-week	1.16	1.27	1.08	0.80	0.98	0.66

Note: Effect size estimates represent Cohen’s d variant for repeated measures (Gibbons et al., 1993). PCL-5: PTSD Checklist for DSM-5. PHQ-9: Patient Health Questionnaire.

Figure 1.

2-week ITP PTSD severity trends by cohort type.

Note: Error bars represent +/−1 standard error. ITP: Intensive PTSD Treatment Program. MST: military sexual trauma. PCL-5: PTSD Checklist for DSM-5. PTSD: Posttraumatic Stress Disorder.

Figure 2.

2-week ITP depression severity trends by cohort type.

Note: Error bars represent +/−1 standard error. ITP: Intensive PTSD Treatment Program. MST: military sexual trauma. PHQ-9: Patient Health Questionnaire.

Existing model evidence also does not suggest program outcomes differ based on age, sex, race, or ethnicity. Model posterior samples indicated that the probabilities of moderate-sized (d = 0.5) or larger differences in PCL-5 or PHQ-9 over time based on sex, race, ethnicity, or a 10-year age difference were below 0.01%, indicating a lack of program outcome differences based on these variables that aligns with prior results from the 3-week ITP (Zalta et al., 2018). Negative posttrauma cognitions were the only meaningful predictor of PTSD or depression severity over time. The median slope estimate for this time-varying predictor represents approximately a 2.3-point PTSD severity reduction and a 0.7-point depression severity reduction for each 10-point reduction in PTCI over the course of the 2-week ITP (see Table 3). This indicates that a one standard deviation reduction in PTCI from baseline would equate to over half a standard deviation reduction in PCL-5 (b = 0.60) and slightly less than a half a standard deviation decrease in PHQ-9 (b = 0.45). Based on the current model, when negative posttrauma cognitions decreased by one standard deviation, the probability of a moderate effect size (d = 0.5) or greater reduction in PTSD severity occurring were 99%, though the probability of Depression severity decreasing by this same moderate amount was approximately 9%.

Table 3.

Median slopes for PTSD and depression severity.

	PCL-5		PHQ-9
Predictor	Median slope	94% HDI	Median slope	94% HDI
Time	−1.41	−1.79, −1.03	−0.22	−0.36, −0.08
Time2	−0.04	−0.07, −0.01	−0.03	−0.04, −0.01
Cohort type1	−1.79	−8.17,4.06	0.33	−2.72, 1.98
Cohort type x time	0.17	−0.27, 0.59	0.11	−0.04, 0.24
Age	0.03	−0.18, 0.24	−0.03	−0.11, 0.04
Race	0.13	−4.19, 3.76	−0.09	−1.57, 1.31
Sex	−2.61	−8.63, 3.78	−1.39	−3.67, 0.96
Ethnicity	−1.93	−6.79, 2.92	−0.60	−2.37, 1.11
PTCI	0.23	0.20, 0.26	0.07	0.05, 0.08

Notes:

¹Cohort type parameter estimates are pre-addition of cohort type x time interaction due to inherent change in the interpretation of this parameter following addition of the interaction term. PCL-5: PTSD Checklist for DSM-5. PHQ-9: Patient Health Questionnaire. PTCI: Posttraumatic Cognitions Inventory

Results of non-inferiority tests of final PCL-5 scores between the 2-week (M = 35.05, SD = 16.19) and 3-week ITP (M = 34.12, SD = 19.15), indicated that the 2-week ITP can be considered non-inferior at the program’s endpoint with regards to PTSD severity at a 5-point non-inferiority margin. The hypothesis of non-inferiority of endpoint PTSD severity is approximately 115.19 times more likely than that of the hypothesis that the 3-week program is superior. Similarly, the hypothesis of non-inferiority in endpoint depression severity is approximately 2e⁺⁹ times more likely than the hypothesis that the 3-week program (M = 11.97, SD = 6.16) is superior to the 2-week program (M = 11.66, SD = 5.35) in depression symptom severity at program endpoint, using a 3-point non-inferiority margin.

Finally, satisfaction ratings for both ITPs were equally high. A total of 95.65% and 94.91% agreed or strongly agreed that they felt “satisfied with the clinical care they received at the Road Home Program in the 2-week and 3-week ITPs, respectively. Similarly, 85.00% and 89.42% agreed or strongly agreed that the 2-week and 3-week ITPs improved the problems they needed help with, respectively, on the satisfaction survey. Program satisfaction with clinical care in the 2-week ITP (M = 4.70, SD = 0.72) can be considered non-inferior to that of the 3-week ITP (M = 4.74, SD = 0.72). Similarly, reporting of the program improving problems they needed help with can also be considered non-inferior for the 2-week ITP (M = 4.43, SD = 0.82) relative to the 3-week ITP (M = 4.54, SD = .79). For both satisfaction items the hypothesis of non-inferiority was 1e⁺¹¹ more likely than the superiority of the 3-week ITP.

Discussion

The present study examined the effectiveness of a 2-week ITP based on twice daily individual CPT sessions and evaluated the program’s non-inferiority to an established 3-week ITP using a Bayesian analytical approach. As has been shown with other ITPs (Bryan et al., 2018; Foa et al., 2018; Oprel et al., 2021), PTSD treatment can be effectively delivered within two weeks. Over the course of treatment, individuals reported large reductions in their PTSD (d = 1.16) and depression symptoms (d = 0.98), which is comparable to ITPs of similar timeframes (Bryan et al., 2018; Foa et al., 2018; Oprel et al., 2021). The present study adds to the existing research on intensive CPT by showing that two individual CPT sessions can be feasibly delivered in one day. Importantly, although the overall effects were slightly larger for individuals who participated in the combat trauma compared to MST tracks (PTSD d = 0.19; depression d = 0.32), these differences could potentially be an artifact of MST track participants starting with slightly lower PTSD and depression symptoms than individuals in the combat trauma track. These findings align with our team’s previous research, which has shown the equivalence of outcomes for these two cohort types in the 3-week ITP (Lofgreen et al., 2020).

In addition to examining the effectiveness of the 2-week ITP, a key goal of this research was to utilize a Bayesian analytical approach to evaluate the non-inferiority to the 3-week ITP. Our results clearly demonstrate that the 2-week ITP is equally as effective as the 3-week ITP. Although it may be somewhat unexpected that a shorter program was able to produce equivalent satisfaction and clinical outcomes, the amount of individual CPT was similar between the 3- and 2-week ITPs. Given the extensive prior research demonstrating the effectiveness of CPT (Asmundson et al., 2019), our findings appear to suggest that the amount of time spent in individual evidence-based PTSD treatment may be the primary driver of symptom reductions rather than the total number of clinical intervention hours. Our findings align with previous research which has demonstrated that the time focused on the trauma in session was a predictor of outcomes (Farmer et al., 2017). Clinician workload was comparable in the 2- and 3-week ITPs (i.e. either 2 or 4 hours of CPT per day) because the primary responsibility of clinicians in this program was the ITP. Importantly, veterans were as satisfied with the 2-week ITP and believe that the program helped with the problems they needed help with, which is another indicator to suggest that program length does not necessarily impact perceived quality of care.

As expected, changes in negative posttrauma cognitions emerged as a predictor of treatment outcomes. Specifically, for every self-reported reduction of 10 points on the PTCI, individuals experienced a reduction of approximately 2.3 points on the PCL-5 and 0.7 points on the PHQ-9. These findings are nearly identical to what we previously reported for the 3-week ITP where a 10-point PTCI change results in a reduction of 2.2 points and 0.8 points on the PCL-5 and PHQ-9, respectively². Negative posttrauma cognitions are an important treatment target and marker of success for PTSD treatment (Brown et al., 2019; Zalta, 2015). Given that the total amount of individual CPT, which has been repeatedly shown to reduce posttrauma cognitions (Monson et al., 2006; Owens et al., 2001; Resick et al., 2008), remained mostly unchanged when transitioning from the 3-week to the 2-week ITP, it is likely that CPT drives changes in negative posttrauma cognitions in the ITP, despite the presence of adjunctive services.

Finally, it is important to note that various demographic characteristics, including age, sex, race, and ethnicity did not predict treatment outcomes. Although this finding contradicts other PTSD treatment research (Asmundson et al., 2019; Resick et al., 2020; Rizvi et al., 2009; van Minnen et al., 2002), it mirrors findings from our 3-week ITP where demographic characteristics have also been shown to not impact treatment outcomes (Zalta et al., 2018).

A number of limitations need to be considered. As this was a clinical program rather than a research trial, individuals were not randomized and the two different ITPs were offered at separate times. Thus, we are unable to control for history effects. Compared to a clinical trial, there was also much less oversight over the delivery of various programming components which makes findings more generalizable to other clinical settings but may have impacted outcomes. Because we did not dismantle the ITP, but rather made multiple program changes at once, we are not able to determine which program components are the main predictors of outcomes. We only examined symptom changes using self-report measures over the course of treatment and have not yet examined follow-up data. Follow-up analyses would have been insufficiently powered due to limited follow-up data for the 2-week ITP at this time. We are therefore unable to determine whether individuals are equally as likely to maintain their treatment gains as a result of shorter treatment. Finally, the 2-week ITP was launched in the early phases of the COVID-19 pandemic, which may have impacted who selected to attend treatment and the ITP outcomes. It is possible that certain veterans selected not to apply for the program given that they would have needed to travel to the ITPs, as the treatment was delivered in-person, and mask wearing was required.

In summary, the present study demonstrated that a significantly shorter 2-week ITP was as effective as a 3-week ITP. Although the primary delivery of CPT was changed from a combination of daily individual and group sessions in the 3-week ITP to twice daily individual CPT sessions in the 2-week ITP, the total amount of individual CPT remained similar between programs. Thus, we can cautiously conclude that, in the context of intensive programs for PTSD, the content of the interventions, rather than the length of the program may be most relevant to outcomes. The notion that individual CPT is the likely driver of change is further supported by our findings that negative posttrauma cognitions are equally engaged in the 2-week ITP as we have previously shown for the 3-week ITP, despite reductions the overall number of program hours, as well as specific reductions in adjunctive services. Thus, when designing clinical programs, it may be helpful to start with smaller, leaner programs, evaluate their effectiveness, and then systematically add components to see whether doing so improve outcomes. Such a constructivist approach may be easier to facilitate than scaling down a large program. For clinicians and patients alike, it is important to note that more treatment hours and components, as impressive as it may appear, may not always be better, or may at least not be better for everyone.

Based on these findings, we strongly recommend clinical programs to rigorously evaluate their general effectiveness on an ongoing basis and to examine the impact changes to the content have on outcomes and known treatment mechanisms. Dismanteling programs systematically by removing one component at a time may aid with understanding what is driving the outcomes. Going forward, it will also be important to examine the extent to which intensive PTSD treatment can be further shortened or streamlined before clinical outcomes or patient satisfaction are negatively impacted. Finally, future research should identify for whom a shorter treatment duration may be preferable and who may benefit from longer programming, as well as which types of individuals, and under what circumstances, they may respond more favorably to combining evidence-based PTSD treatment with adjunctive services versus standalone evidence-based PTSD treatment delivered intensively.

Clinical Impact Statement.

The present study compared two intensive treatment programs for veterans with posttraumatic stress disorder (PTSD). The 2- and 3-week-long programs were based on daily Cognitive Processing Therapy (CPT) as well as adjunctive services. The 2-week program had fewer overall service hours but offered a comparable amount of CPT. Both programs produced large and equivalent PTSD and depression severity reductions. The findings suggest that the more/longer treatment is not necessarily better and that content of the intensive PTSD treatment programs appears to matter more than its overall length. Shorter PTSD treatment programs have the potential to increase access and capacity.