Effects of expressive writing on sexual dysfunction, depression, and PTSD in women with a history of childhood sexual abuse: Results from a randomized clinical trial (Meston, Lorenz, & Stephenson, 2013)
Assessment Session: Sexual Schema Prompt
For the next 30 minutes, I would like for you to write about your personal thoughts and feelings associated with sex and sexuality. In your writing, I’d like you to link your thoughts about sex to past, current, or future sexual experiences or relationships. You might also address more broadly how you view yourself as a sexual person. Please try to be as detailed as possible in your description. I’d like you to really let go and explore your very deepest emotions and thoughts.
Schema, Session 1 Prompt
Today I would like you to focus on how your sexual abuse may have affected your sexual beliefs about any of the following: yourself, sexual partners, and sex in general. Write about how things may be different if you had a different sexual history. Do not be afraid to tap into negative feelings such as sadness, loss, bereavement or anger as you write and make sure you explore these emotions and the reasons behind them.
Schema, Session 2 Prompt
Today I would like you to focus on how your sexual abuse may have affected your sexual beliefs about any of the following: yourself, sexual partners, and sex in general. What are your beliefs, and what reasons do you have for maintaining them? Think also of reasons why those beliefs might not be valid (for example, think of evidence from your own experiences that does not support your beliefs about sexuality. Do not be afraid to tap into negative feelings such as sadness, loss, bereavement or anger as you write and make sure you explore these emotions and the reasons behind them.
Schema, Session 3-4 Prompt
Reflect on your work identifying how sexual abuse may have affected your beliefs about sexuality. Sometimes we are willing to maintain a belief or a position, even if this belief causes pain or discomfort, because of the potential negative effect of changing our views. What would it be like for you if you were to change some of your opinions regarding sex? What are some of the forces that keep you tied to beliefs or views that perpetuate sexual problems or distress about sex?
Schema, Session 5 Prompt
The purpose of today’s session is to put together the thoughts and ideas you have been exploring in the previous sessions. Spend today trying to envision some goals for yourself and planning how you might use your insights to help you reach those goals. Write about how you would like your sexual life to change in the future and what you need to do to get there. How might the thoughts and ideas you have been working on help you? Focus on the strengths you have learned about yourself and put any weakness you have encountered into perspective.
In-session symptom measure
- How much have you been bothered by unwanted or intrusive memories about past abusive experiences since your last session?
0 1 2 3 4 5 6
Much less About the Much more
than usual same as usual than usual
- How often have you experienced negative feelings such as anger, sadness, or loss since your last session?
0 1 2 3 4 5 6
Much less About the Much more
than usual same as usual than usual
- How much contact have you had with important people in your life (friends, close family members, significant others) since your last session?
0 1 2 3 4 5 6
Much less About the Much more
than usual same as usual than usual
- Have you, at any time, had any thoughts or urges to harm yourself since your last session?
YES NO
Information regarding statistical models for treatment outcome analyses
Growth Curve Modeling (GCM) is typically completed using two steps (Raudenbush & Bryk, 2002) . First, we created Random Coefficients Models to establish whether our outcomes exhibited change over time, the average strength and shape of that change, and whether there was variability in the rates of change. In the Random Coefficients Model, our outcome Y for person I at time J (Y ij ) was modeled as a function of time. To allow for non-linear change, we used quadratic and cubic iterations of the time factor. Thus, our initial Random Coefficients Model took the following form:
Y ij = π 0j + (π 1j)*TIME + (π 2j)*TIME 2 + e ij
π 0j = β 00 + r 0j
π 1j = β 10 + r 1j
π 2j = β 20 + r 2j
Y ij is modeled as a function of an intercept (person I’s outcome score at time zero: π 0j), Time (multiplied by person I’s linear growth rate: π 1j), Time squared (multiplied by person I’s quadratic growth rate: π 2j), and a residual (e ij). Additionally, person I’s intercept, as well as her linear and quadratic rates of change, were modeled as a function of the average intercept or growth rate within the sample (β 00, β 10, β 20), plus some residual (r). We tested the statistical significance of each of these parameters; specifically, whether the average growth rates were significantly different from zero, and whether there was significant inter-individual variation in the strength of each growth rate (π).
Assuming there was significant inter-individual variation, we then included person-level (Level-2) explanatory variables (e.g., condition) to attempt to explain this variation. This Conditional Model took the following form:
Y ij = π 0j + (π 1j)*TIME + (π 2j)*TIME 2 + e ij
π 0j = β 00 + (β 01)*Condition + r 0j
π 1j = β 10 + (β 11)*Condition + r 1j
π 2j = β 20 + ( β 21 )*Condition + r 2j
Again we tested the statistical significance of each parameter, to see if A) the person-level variable (in this case, “Condition”) predicted the strength of each growth rate (e.g., β 11 is the effect of condition on the linear growth rate), and B) whether there was significant inter-individual variation in growth rates remaining after accounting for our explanatory variable (π).
Retention and dropout
A total of 215 women contacted the laboratory and were assessed for possible participation. The most common reasons for exclusion were not experiencing any sexual problems (n = 17), abuse within the last two years (n = 10), concomitant psychotherapy (n = 8), reported suicidal ideation or other indication of high risk of suicidality (n = 6) and other (e.g., too far away to attend laboratory sessions, n = 9). Thirty-nine women were evaluated as eligible for the study and were scheduled for an initial assessment, but did not come to their appointment and could not be reached to reschedule. Forty-five women were randomized to an expressive writing condition not presented here; this condition was designed as a present-focused condition but participants reported inconsistent interpretations of the prompts, leading to significant variability in what they wrote about. The remaining 91 participants were randomized into the two experimental conditions. Dropout following initiation of treatment was approximately equal across conditions; in all cases participants chose to discontinue or were discontinued by the therapist due to increased risk of self-harm. To ensure true randomization (i.e., non-blocked randomization), we used a pre-defined randomly generated list of potential participant numbers with assignment to one of the three conditions. Each new participant who was determined eligible was matched to the next number on this list.
Staff and treatment fidelity checks
There were 4 therapists and 3 assessors, all women with master’s degrees in psychology. Therapists saw approximately equal numbers of patients in each condition. Assessors were blinded to participant’s treatment condition and therapists were blinded to the results of the assessor’s diagnostic interviews. Participants were assigned their randomization status by their therapist following their pre-treatment assessment; although eligibility was to be confirmed at intake, no participant who came to their pre-treatment assessment session was found to be ineligible.
All study therapists were supervised by the same person, a licensed clinical psychologist trained in the study procedures by the principal investigator. Study therapists were trained on the treatment procedures by the principal investigator, and then shadowed the principal investigator or a trained study therapist for 3 – 5 sessions. They were then directly observed for 5 – 7 sessions to ensure fidelity to treatment procedures. All study therapists met weekly with the supervisor, and were re-assessed at random intervals for fidelity to the core procedures (presentation of treatment rationale, administration of the check-in form, safety assessment following writing, providing non-specific, unstructured supportive contact following writing as needed).
Because the number of women who elected to leave following risk assessment was so small, we were unable to compare them to women who elected to stay). As per other studies of this nature, therapists were permitted to make basic empathic or supportive comments, reflect participant’s statements on her writing, validate the participants’ emotions or praise her strength for engaging in treatment, and discuss appropriate methods of coping with distress brought up by treatment (e.g., help plan a pleasant activity for later in the day).
Some participants chose to take a break of several weeks between sessions (generally around winter holidays). The flexibility of treatment scheduling (and choice to stay or leave following writing) was designed to increase participant’s sense of control of the therapy process, increase fidelity to the core components of the treatment (namely, expressive writing), and more closely mimic real-world therapy conditions. Although most participants elected the same schedules (one-hour appointments weekly), many participants reported in exit interviews that this flexibility helped them “buy-in” to the treatment (e.g., by allowing them the option of not disclosing what they wrote, otherwise avoidant participants were able to engage with treatment).
