The purpose of the secondary machine learning analyses conducted in our study was to determine whether patients with PTSD stemming from childhood abuse would benefit more from a phase-based treatment approach or direct trauma-focused approach. The researchers aimed to develop a personalized advantage index (PAI) to identify the optimal treatment option for each patient and to calculate the potential advantage of offering the most appropriate treatment based on this index score.
This study used data from a randomized controlled trial comparing a direct trauma-focused treatment approach (EMDR therapy) with a phase-based treatment approach (Van Vliet et al., 2021). The phase-based approach consisted of EMDR preceded by Skills Training in Affective and Interpersonal Regulation (STAIR; Cloitre et al., 2002). A phase-based approach is based on the notion that it is important to increase safety, establish a therapeutic relationship, teach patients stress management, and affect regulation skills in the first phase of treatment, that is, prior to the phase during which patients are exposed to traumatic memories (Cloitre et al., 2002). The researchers noted that this was only one of the few PAI analysis in the area of PTSD research and treatment.
The primary outcome variable was PTSD symptom severity, measured using the Clinician-Administered PTSD Scale-5 (CAPS-5; Weathers et al., 2013). The pre-treatment variables that were included in the selection of potential predictors and moderators were patient demographic information, PTSD self-report (the PTSD Symptom Scale-Self Report; Foa et al., 1993), suicidality (determined by item 9 of the Beck Depression Inventory-II; Beck et al., 1996), presence of a borderline personality disorder (determined by the Structured Clinical Interview for DSM-IV; First et al., 1997) with severity of self-injury, severity of dissociative symptoms (by the Dissociative Experiences Scale-II; Carlson & Putman, 1993), Complex PTSD diagnosis ( measured by the Structured Interview for Disorders of Extreme Stress; Scoboria et al., 2008), emotion regulation difficulties (by the Difficulties in Emotion Regulation Scale; Gratz & Roemer, 2004), interpersonal problems (by the Inventory of Interpersonal Problems; Horowitz et al., 2000), and posttraumatic cognitions (subscales of the Posttraumatic Cognitions Inventory; Foa et al., 1999), and general psychopathology (as measured by the subscales of the Dutch Brief Symptom Inventory; De Beurs, 2006).
The following moderators were identified through machine learning: borderline personality disorder, suicidal thoughts, high scores on the cold/distant subscale of the Inventory of Interpersonal Problems, high scores on the hostility symptoms subscale of the Brief Symptom Inventory, and high scores on the Dissociative Experiences Scale showed lower post-treatment PTSD symptoms in the EMDR condition than in the STAIR-EMDR condition. Additionally, it was estimated that not being married was associated with higher posttreatment PTSD severity scores in the STAIR-EMDR condition than in the EMDR condition. The subscale self-sacrificing and the subscale vindictive/self-centered of the Inventory of Interpersonal Problems were ordinal moderators. This means that all scores are linked to different levels of the superior effect of one treatment over the other. In these cases, STAIR-EMDR was superior, indicating lower post-treatment PTSD symptom severity scores, and the degree of superiority was more pronounced when scores on the Inventory of Interpersonal Problems subscales were lower. After identifying the moderators, the selected variables were combined to construct the PAI.
The main finding of this study was that there were no significant differences in post-treatment severity scores between individuals receiving the treatment indicated by the PAI and those receiving the treatment not indicated by the PAI. The effect sizes, which provided an estimate of the magnitude of the differences between the two groups, were reported as small-to-medium (Cohen’s d between indicated and non-indicated treatment = 0.25; Cohen’s d between indicated and non-indicated treatment of the highest 60.0 % PAI scores = 0.41). These effect sizes were similar to those observed in other PAI studies with significant results (e.g. Cohen et al, 2019; Schwartz et al., 2021). Small to moderate matching effect sizes may be relevant for optimizing treatment in clinical practice. It is important to note that these conclusions are based on a small sample size and that further research with larger sample sizes and different populations is needed (Luedtke et al., 2019).
Target Article
Bremer, S., van Vliet, N. I., Van Bronswijk, S., Huntjens, R., de Jongh, A. d., & van Dijk, M. K. (2023). Predicting optimal treatment outcomes in phase-based treatment and direct trauma-focused treatment among patients with posttraumatic stress disorder stemming from childhood abuse. Journal of Traumatic Stress, 00, 1–12. https://doi.org/10.1002/jts.22980
Discussion Questions
- In what cases would you still use stabilization programs and why?
- What do therapists fear about direct trauma-focused treatments?
- What do patients fear about direct trauma-focused treatments?
About the Authors
Noortje van Vliet (first author) : Psychologist and researcher (MSc.) at Dimence Mental Health Group, The Netherlands. Noortje van Vliet can be contacted at: n.vanvliet@dimence.nl
Susanne Bremer (first author): Data analyst (MSc.) Dimence Mental Helath Group, The Netherlands.
Suzanne van Bronswijk: Psychiatrist and researcher (PhD), Maastricht University, The Netherlands.
Rafaele Huntjens: Professor at the University of Groningen, The Netherlands.
Ad de Jongh: Professor at the University of Amsterdam, The Netherlands; Salford University and University of Worcester, United Kingdom. Professor De Jongh can be contacted at: a.d.jongh@acta.nl
Maarten van Dijk: psychologist and researcher (PhD) at Dimence Mental Health Group, The Netherlands.
References Cited
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory II. Psychological Corporation.
Cloitre, M., Koenen, K. C., Cohen, L. R., & Han, H. (2002). Skills training in affective and interpersonal regulation followed by exposure: A phase-based treatment for PTSD related to childhood abuse. Journal of Consulting and Clinical Psychology, 70(5), 1067–1074. https://doi.org/10.1037/0022-006X.70.5.1067
Carlson, E. B., & Putnam, F. W. (1993). An update on the Dissociative Experiences Scale. Dissociation: Progress in the Dissociative Disorders, 6(1), 16-27.
Cohen, Z. D., Kim, T. T., Van, H. L., Dekker, J., & Driessen, E. (2020). A demonstration of a multi-method variable selection approach for treatment selection: Recommending cognitive-behavioral versus psychodynamic therapy for mild to moderate adult depression. Psychotherapy research: journal of the Society for Psychotherapy Research, 30(2), 137–150. https://doi.org/10.1080/10503307.2018.1563312
De Beurs, E. (2006). Brief Symptom Inventory. Handleiding. PITS.
First, M. B., Gibbon, M., Spitzer, R. L., Williams, J. B. W., & Benjamin, L. S. (1997). Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). American Psychiatric Press, Inc.
Foa, E. B., Riggs, D. S., Dancu, C. V., & Rothbaum, B. O. (1993). Reliability and validity of a brief instrument for assessing posttraumatic stress disorder. Journal of Traumatic Stress, 6, 459-473. https://doi.org/10.1007/BF00974317
Foa, E. B., Ehlers, A., Clark, D. M., Tolin, D. F., & Orsillo, S. M. (1999). The Posttraumatic Cognitions Inventory (PTCI). Development and validation. Psychological Assessment, 11(3), 303-314. https://doi.org/10.1037/1040-3590.11.3.303
Gratz, K. L., & Roemer, L. (2004). Multidimensional assessment of emotion regulation and dysregulation: Development, factor structure, and initial validation of the Difficulties in Emotion Regulation Scale. Journal of Psychopathology Behavioral Assessment, 26, 41-54. https://doi.org/10.1023/B:JOBA.0000007455.08539.94
Horowitz, L. M., Alden, L. E., Wiggins, J. S., & Pincus, A. L. (2000). Inventory of Interpersonal Problems. Psychological Corporation.
Luedtke, A., Sadikova, E., & Kessler, R. C. (2019). Sample Size Requirements for Multivariate Models to Predict Between-Patient Differences in Best Treatments of Major Depressive Disorder. Clinical Psychological Science, 7(3), 445-461. https://doi.org/10.1177/2167702618815
Schwartz, B., Cohen, Z. D., Rubel, J. A., Zimmermann, D., Wittmann, W. W., & Lutz, W.
(2021). Personalized treatment selection in routine care: Integrating machine learning and statistical algorithms to recommend cognitive behavioral or psychodynamic therapy. Psychotherapy Research, 31(1), 33-51. https://doi.org/10.1080/10503307.2020.1769219
Scoboria, A., Ford, J., Lin, H., & Frisman, L.(2008). Revision of the Structured Interview for
Disorders of Extreme Stress (SIDES): An exploratory and confirmatory factor analytic approach. Assessment, 15, 404-425. https://doi.org/10.1177/10731911083190
Van Vliet, N. I., Huntjens, R. J. C., Van Dijk, M. K., Bachrach, N., Meewisse, M. L., & De Jongh, A. (2021). Phase-based treatment versus immediate trauma-focused treatment for post-traumatic stress disorder due to childhood abuse: Randomised clinical trial. British Journal of Psychiatry Open, 7(6), E211. https://doi.org/10.1192/bjo.2021.1057
Weathers, F. W., Blake, D. D., Schnurr, P. P., Kaloupek, D. G., Marx, B. P., & Keane, T. M. (2013). The Clinician-Administered PTSD Scale for DSM-5 (CAPS-5). National Center for Posttraumatic Stress Disorder, 2013 (www.ptsd.va.gov [2013]).