Online Writing Intervention for Major Depressive Disorder

1. Introduction and Background

In a 2012 Statistics Canada Survey, 1.6 million Canadians reported that their needs for mental health services were unmet. In Ontario, almost 60% of mental health concerns are related to Major Depressive Disorder (MDD), representing an estimated excess of 210 million dollars in healthcare costs. Psychotherapy is an effective and enduring option for treating depression, but in Canada, its availability is limited. Increasing the number of psychotherapists does not seem to be an adequate solution to this issue; despite an increase in psychotherapy provision in the past decade, many Canadians are still in need of psychotherapy.

One solution is to develop innovative therapies that can be delivered reliably at low costs to large populations, for example, using the internet. Expressive writing (EW) is a simple intervention involving a series of journaling tasks where individuals write their deepest thoughts and feelings about a negative issue or problem. There is a large literature examining the various physical and psychological benefits associated with EW among different populations. EW appears to be a promising intervention for MDD. In one study of clinically depressed patients, EW completed in person for 20 minutes over three consecutive days resulted in less depressive symptoms relative to control writing, and this effect persisted after one month. Importantly, EW does not require manuals or highly trained clinicians, making it cost-effective and amenable to being administered remotely. Preliminary studies of its online administration with depressed individuals have produced encouraging findings.

II. Rationale and Study Objectives This study determines whether it is feasible to administer EW online to patients with MDD, and whether patients find it useful.

Feasibility concerns. Although the open-ended nature of EW makes it easy for patients to complete remotely (i.e., without the guidance or presence of a clinician), it presents a risk of drop-out particularly in severely depressed patients, due to a lack of motivation and other factors. The investigators assess the feasibility of administering EW remotely (i.e., online or by mail) to participants seeking treatment at CAMH for Major Depression. They also examine participants' perceptions of EW (i.e., its potential benefits and limitations), as well as their motivation and engagement with the intervention.

Efficacy concerns in a target patient population. In various studies, EW has reduced symptoms of depression. However, other studies have reported no benefits of EW or even adverse effects for individuals with severe levels of depression. These inconsistent findings have prompted researchers to caution against using EW in severely depressed samples. They suggest that writing facilitates depressive rumination (i.e., persistent and recursive thinking about the depressive episode), leading to a worsening of symptoms. Thus, it is also possible that unstructured treatments like EW are not helpful for individuals with depression, particularly if administered online. For this reason, the investigators also assess the efficacy of online EW in a sample of CAMH patients who are diagnosed with depression. Specifically, they examine its impact on symptoms of depression and anxiety, personal functioning, and perceived complexity of personal problems.

III. Aims and Hypotheses The investigators conduct a multi-methods study to assess the feasibility and efficacy of administering EW online to patients with MDD. They also examine perceptions of its impact on emotional and cognitive wellbeing. To achieve these aims, the investigators randomize patients into one of two groups: one group is prompted to complete four sessions of EW online, whereas the other group is not asked to complete this task. Although control conditions in studies of EW often involve sessions of writing about non-emotional topics, the investigators do not administer a control writing task due to their focus on feasibility. The aim is to examine whether a group of patients with MDD would complete a writing intervention online, which requires a comparison of their study completion rates and outcomes to a group not asked to complete any writing at all. Nevertheless, both groups complete self-reported measures of outcomes. The primary outcome related to efficacy is reduction in MDD symptoms; secondary outcomes include symptoms of Generalized Anxiety Disorder (GAD), functional impairment, and the perceived complexity of personal problems. Participants complete these measures before and after the EW task, and at a one-month follow up. The investigators examine the rate of EW completion and adherence to the suggested writing duration, as well as its impact on outcomes over time. They also conduct qualitative interviews which ask EW participants about their experiences with the intervention, with a focus on perceived benefits and limitations, and its emotional and cognitive effects.

The investigators propose that online administration is feasible if 70% of participants randomized to the intervention complete at least three 20-minute sessions of writing. EW is useful if there is a significant reduction in MDD symptoms over time, as compared to the control group, which either emerges after the intervention or at the one-month follow-up. Based on research finding temporary increases in emotional distress related to EW, it is also possible EW participants might report increased symptoms immediately following the intervention. However, changes in functional impairment or perceived problem complexity could take some time to emerge, so changes related to these outcomes are expected to emerge at the one month follow-up. In qualitative interviews, participants are expected to report EW to be beneficial, particularly as related to its emotional or cognitive effects.

IV. Methods

Participants The investigators recruit outpatients from the Centre for Addiction and Mental Health. Patients are referred to the study by psychiatrists during initial or follow-up appointments, or by psychotherapists either before the start or after the end of group psychotherapy. Inclusion criteria are adults aged 18 or older, who were assessed by a psychiatrist and received a diagnosis of MDD with or without another co-morbid condition. Participants also require access to the internet and an email address at which they could receive study materials. Patients are excluded if they cannot understand, speak, or write in English. Clinicians are also asked not to refer patients into the study if they had cognitive impairments that could prevent them from being able to complete a writing task. Clinicians are also asked not to refer patients reporting suicidal thoughts and behaviours or other mental health symptoms which could put them at risk for negative outcomes when completing the intervention remotely. Patients are not excluded if they are receiving treatment for MDD (e.g., antidepressants, group CBT).

Measures Demographics: Participants self-report their demographic characteristics, which included their age, gender, racial or ethnic background, and postal code, which was used to derive socioeconomic status.

Patient Health Questionnaire (PHQ-9): The PHQ-9 is a self-reported measure of MDD symptom severity; it measures the frequency of 9 symptoms on a scale from 0 (not at all) to 3 (nearly everyday). Scores are summed to reflect MDD symptom severity, and they range from 0-27, with higher scores indicating greater severity.

Generalized Anxiety Disorder (GAD-7): The GAD-7 is a self-reported measure of GAD symptom severity; it measures the frequency of 7 symptoms on a scale from 0 (not at all) to 3 (nearly everyday). Summed scores range from 0-21, with higher scores indicating greater severity.

WHO-Disability Assessment Schedule (WHODAS): WHODAS is a self-reported measure of functional impairment. It measures the degree of difficulty from the past month related to 12 activities in domains of cognition, mobility, self-care, and social participation. Activities are measured on a scale from 0 (none) to 4 (extreme or cannot do), and scores are summed to reflect overall functional impairment. They range from 0-48, with higher scores indicating greater impairment.

Problem Complexity Questionnaire (PCQ): PCQ measures the perceived complexity of personal problems. It consists of 8 statements reflecting the complexity of problems, and participants rate each item on a scale from 1 (completely disagree) to 4 (completely agree). Scores range from 8 to 32, with higher scores indicating higher perceptions of problem complexity.

Procedures The study is approved by CAMH's Research Ethics Board. Patients referred to the study are provided with a high-level description of study aims (i.e., "evaluating the extent to which completing questionnaires or a writing task is useful for people with depression"). Prior to enrolment, all participants are given information about study procedures and risks, and they provide their written informed consent. Once enrolled, participants are allocated to EW or control with 4-block randomization and they are sent the first study questionnaire consisting of the demographics form, PHQ-9, GAD-7, WHODAS, and PCQ. Their electronic health records are also accessed to extract information about their diagnoses and current treatment. After completing the first questionnaire, participants assigned to EW are sent instructions to complete the first EW session on the next day. Instructions are accompanied by a text box, in which participants can write freely. Although participants are instructed to write for 20 minutes, they are also told during that they could pause or stop the writing task at any time (e.g., if they felt unwell). EW participants are provided with the same instructions on the next three days. Participants who are allocated to control are not provided with the EW instructions. After one week (for the control group) or on the next day that participants complete their fourth EW session (for the EW group), participants receive a second questionnaire containing the PHQ-9, GAD-7, WHODAS, and PCQ. One month after completing the second questionnaire, they receive a third questionnaire containing the same measures. All study materials (i.e., questionnaires, EW sessions) are administered via RedCap, a web-based application for data collection available at CAMH, and study instructions are sent to participants by email. If participants do not complete an EW session or follow-up questionnaire within three days of receiving it (or after one week for the first questionnaire), participants who agree to receive reminders by text message or email are sent up to two reminders. Participants who do not respond or complete the relevant study components after two reminders are assumed to have dropped out of the study and are no longer contacted. However, participants who respond to reminders or complete the questionnaires or EW sessions within any period of time are included.

After completing the third questionnaire, all EW participants are invited to participate in a brief qualitative interview. Qualitative interviewing is guided by the question of how individuals with MDD experience online EW, barriers to its completion, perceived benefits and limitations, emotional and cognitive effects, and suggestions for improvement. At the end of the study, control participants are also given an opportunity to complete the EW task, if they wish.

Statistical Analysis The investigators examine the baseline demographic and clinical characteristics of the sample and compare them between the control and EW groups. Differences in categorical variables are compared with Chi-squared tests. Due to small samples in some racial or ethnic categories, this variable is recoded into a binary representation when testing for differences between groups. Continuous variables that are normally distributed are compared with t-tests, whereas Wilcoxon-signed rank tests are used for continuous variables with non-parametric distributions.

The investigators examine how many EW participants completed at least three sessions, how many wrote for at least 20 minutes, the median length of time spent writing, the median word count of EW tasks, and average time intervals between writing sessions. Intention-to-treat analysis involves generating mixed effect models with maximum likelihood estimation using all participants in the EW group and all participants in the control group who complete the first (baseline) questionnaire. The investigators generate four models to examine the impact of condition, time, and a condition by time interaction on each outcome (i.e., PHQ-9, GAD-7, WHODAS, PCQ). Time is treated as a categorical factor, to account for the possibility of non-linear effects. For any models with evidence of this interaction, estimated means are plotted to inspect trajectories, with 95% confidence intervals to gauge effects. In a per-protocol analysis, the investigators generate the same mixed effect models adding number of completed sessions as a predictor, to determine whether completing fewer sessions impacts outcomes (setting all control participants to 0).

The approach to analysing qualitative interviews is guided by reflexive thematic analysis (Braun \& Clark, 2019), which is a recursive and constructive process of identifying and analysing themes. Given the focus on emotional and cognitive effects, the investigators adopt a predominantly theoretical or deductive approach, to provide a nuanced account of how participants experienced these effects of EW. To accommodate the diversity of experiences, the investigators incorporated inductive analysis. Analysis is primarily semantic and essentialist, since the investigators wish to understand participants' explicit experiences or accounts of EW, rather than deriving latent themes.