Using Patient Writings in Psychotherapy: Review of Evidence for Expressive Writing and Cognitive-Behavioral Writing Therapy

In Pennebaker’s theory of emotional inhibition (3), he hypothesized that individuals who have experienced traumatic events want to share their experiences with others for social support. However, feared or actual negative responses from others inhibit them from sharing their feelings and thoughts. Pennebaker theorized that this inhibition resulted in chronic low levels of stress, which placed individuals at increased risk of both physical and mental disorders. Pennebaker designed the expressive writing technique as a means for testing whether disclosing previous trauma was associated with decreased stress and subsequently improved physical and mental outcomes.

The initial study by Pennebaker and Beall (2) involved college students writing about either a previous trauma or a trivial topic. Students were asked to write for 15 minutes for 4 consecutive days. Assessments were completed before and 4 months after the intervention, including self-report of mood and different physical symptoms, such as headache, racing heart, and tense stomach. There was a significant decrease in the number of physical symptoms between the individuals who wrote about their trauma versus those who wrote about a trivial topic (F=3.05, df=3, 38, p=0.04); however, there were no significant differences between the groups in terms of mood.

Since this initial trial, a large number of studies examining expressive writing for a variety of different populations and settings have been completed. The majority of studies examining psychiatric symptoms involve samples of individuals who have been exposed to either a major adversity, such as cancer (4, 5), or to significant trauma, such as sexual abuse (6, 7). There are a limited number of studies looking specifically at individuals with preexisting psychiatric diagnoses (Table 1). The studies on expressive writing have generally used interventions similar to Pennebaker’s initial study, with participants writing for 20–30 minutes for three to five sessions (4). Participants in these studies are usually given a prompt such as: “I would like for you to write about your very deepest thoughts and feelings about the most traumatic experience of your entire life. In your writing, I’d like you to really let go and explore your very deepest emotions and thoughts” (8).

There have been multiple meta-analyses examining the efficacy of expressive writing (9–11). An initial meta-analysis by Smyth (9) found a moderate effect size of expressive writing on physical health (Cohen’s d=0.421, p<0.0001), psychological well-being (Cohen’s d=0.661, p<0.0001), and biophysical measures (Cohen’s d=0.681, p<0.0001) but not on health behaviors, such as sleep, drug use, and exercise.

Several years later, Mogk et al. (10) completed a meta-analysis that included a larger number of studies than the number included in the Smyth study (N=30 versus N=13), which were completed in a larger array of settings by a more diverse group of researchers (Pennebaker was listed as an author in 23% of the studies versus 62% in the Smyth meta-analysis). Both meta-analyses only included randomized controlled studies in which the control group was asked to write about a neutral topic. However, Mogk et al. only included studies that examined the effect of expressive writing at least 4 weeks after the last writing session. Four of the 13 studies included in the Smyth meta-analysis were excluded from the meta-analysis by Mogk et al.

In their meta-analysis, Mogk et al. found no significant effects of expressive writing on either physical or psychological health. It is difficult to apply these findings to individuals with psychiatric disorders because the majority of these studies used nonclinical samples, largely students. Mogk et al. completed an exploratory analysis that examined the six studies in the meta-analysis that included posttraumatic symptoms as an outcome. They found no significant effect on arousal, intrusion, or avoidance. Although these six studies were conducted with individuals who had been exposed to some type of trauma, only one of the studies included individuals diagnosed with posttraumatic stress disorder (PTSD) (11).

Frisina et al. (12) conducted a meta-analysis examining the effect of expressive writing on clinical populations, unlike previous meta-analyses that combined studies of clinical and nonclinical populations. Of the nine studies included, four examined patients with psychiatric diagnoses and the remaining five examined patients with diagnoses of cancer, asthma, and rheumatoid arthritis. They found that expressive writing had a significant effect on physical health (Cohen’s d=0.21, p=0.01) but not on psychological health (Cohen’s d=0.07, p=0.17). No exploratory analyses were completed to examine whether the effect size differed for psychiatric versus nonpsychiatric samples.

Lange et al. (13, 14) incorporated this research on expressive writing to develop a psychotherapy that integrated writing and cognitive-behavioral therapy (CBT). The theoretical basis for their therapy was that the three components crucial to the efficacy of CBT are exposure, cognitive restructuring, and social sharing and support (13). They utilized 1) structured writing assignments as a means for exposure, 2) therapist feedback and rewriting of assignments to facilitate cognitive restructuring, and 3) sharing of the final writing for garnering social support and connectedness.

Their therapy, called interapy, is completed entirely on the Internet. It includes three phases with a total of 10 writing assignments that are 45 minutes in length and occur over a 5-week period. Therapists read the assignments and then give feedback through the website between assignments. The first phase is self-confrontation, in which patients write four assignments about their previous trauma. The second phase is cognitive restructuring, comprised of four assignments with the goal of assisting patients in challenging automatic thoughts and gaining a new perspective on the trauma. The last phase is social sharing, with psychoeducation about the benefits of sharing and having patients write a letter to someone involved in the trauma or a person who is significant in their life.

Lange et al. (14) completed a randomized controlled study examining the efficacy of interapy. Participants were individuals who went to the interapy website after learning about the trial through media coverage and completed online screening for participation. Of the 437 individuals who passed the screening, 184 committed themselves to the study. Participants were assigned to either a control group (being on a waitlist) (N=62) or to therapy (N=122). There was a high dropout rate in both arms of the study (N=30 in the control group; N=44 in the treatment group). However, there were no significant differences between people who completed the study and those who dropped out of the study in terms of measures of trauma or psychological functioning prior to starting the therapy.

The results demonstrated that the improvement in the treatment group was significantly larger than that in the control group, with large effect sizes for intrusion (Cohen’s d=1.28, p<0.0001) and avoidance (Cohen’s d=1.39, p<0.001). These effects decreased slightly, although not significantly, at the 6-week follow-up. There were also significant decreases in anxiety, depression, somatization, and sleep problems, with large effect sizes for depression (Cohen’s d=1.04, p=0.0001) and anxiety (Cohen’s d=0.76, p=0.001).

Cognitive-behavioral writing therapy was adapted for use in children with PTSD, with therapists completing the writing assignments in person with the children. An uncontrolled pilot study of this therapy found a significant decrease in PTSD and depression symptoms, in addition to a significant decrease in internalizing and externalizing behavior problems (15).

Van Emmerik et al. (16) completed a study examining the efficacy of CBT versus a structured writing therapy, which had been adapted from the interapy program for PTSD. Unlike the study by Lange et al. (14), patients were recruited from a psychiatric clinic, and diagnosis was based on structured clinical interview versus online screening. In addition, writing assignments were reviewed during in-person therapy sessions rather than over the Internet. Similar to the Lange et al. study, there was a high dropout rate, although this was not found to be related to which treatment a participant was receiving. Participants in structured writing therapy and CBT groups did better than those in a control group, with significant reductions in intrusion, avoidance, depression, anxiety, and dissociative symptoms. No significant differences were found between structured writing therapy and CBT in terms of efficacy.