Published on 27.06.19 in Vol 4, No 2 (2019): Apr-Jun
Preprints (earlier versions) of this paper are available at http://preprints.jmir.org/preprint/13857, first published Feb 27, 2019.
Web-Based Benefit-Finding Writing for Adults with Type 1 or Type 2 Diabetes: Preliminary Randomized Controlled Trial
Background: The high prevalence of diabetes distress and subclinical depression in adults with type 1 and type 2 diabetes mellitus (T1DM and T2DM, respectively) indicates the need for low-intensity self-help interventions that can be used in a stepped care approach to address some of their psychological needs. However, people with diabetes can be reluctant to engage in mental health care. Benefit-finding writing (BFW) is a brief intervention that involves writing about any positive thoughts and feelings concerning a stressful experience such as an illness, avoiding potential mental health stigma. It has been associated with increases in positive affect and positive growth and has demonstrated promising results in trials in other clinical populations. However, BFW has not been examined in people with diabetes.
Objective: This study aimed to evaluate the efficacy of a Web-based BFW intervention for reducing diabetes distress and increasing benefit finding in diabetic adults with T1DM or T2DM compared to a control writing condition.
Methods: Adults with T1DM or T2DM and diabetes distress were recruited online through the open access Writing for Health program. After completing baseline questionnaires, they were randomly allocated to receive online BFW or an active control condition of online writing about the use of time (CW). Both groups completed 15-minute online writing sessions, once per day, for 3 consecutive days. Online measures were administered at baseline, 1 month, and 3 months postintervention. Participants were also asked to rate their current mood immediately prior to and following each writing session.
Results: Seventy-two adults with T1DM or T2DM were recruited and randomly allocated to receive BFW (n=24) or CW (n=48). Participants adhered to the BFW regimen. Greater increases in positive affect immediately postwriting were found in the BFW group than in the CW group. However, there were no significant group-by-time interactions (indicating intervention effects) for benefit finding or diabetes distress at either the 1-month or 3-month follow-up. Both the BFW and CW groups demonstrated small, significant decreases in diabetes distress over time.
Conclusions: BFW was well tolerated by adults with diabetes in this study but did not demonstrate efficacy in improving diabetes distress or benefit finding compared to an active control writing condition. However, due to recruitment difficulties, the study was underpowered and the sample was skewed to individuals with minimal diabetes distress and none to minimal depression and anxiety at baseline. Future research should continue to investigate the efficacy of variants of therapeutic writing for adults with T1DM or T2DM, using larger samples of participants with elevated diabetes distress.
Trial Registration: Australiand New Zealand Clinical Trials Registry ACTRN12615000241538; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=368146
JMIR Diabetes 2019;4(2):e13857
Diabetes mellitus is group of disorders posing a global public health challenge. In 2018, the International Diabetes Federation estimated that the global prevalence of diabetes in adults in 2017 was 8.4% (451 million adults) . Psychological comorbidity is common and burdensome in people with type 1 or type 2 diabetes mellitus (T1DM or T2DM, respectively) [ - ], with many experiencing negative thoughts and emotions toward diabetes and its treatment [ , ]. Diabetes distress is a construct correlated with, but distinct from, depression in people with diabetes and includes distress associated with the treatment regimen, eating, hypoglycemia, complications, interpersonal relationships, and health care professionals [ , ]. Approximately 70% of people with T2DM display high levels of diabetes-related distress without meeting the criteria for major depressive disorder (MDD) [ , ]. Diabetes distress is associated with poor glycemic control and acts as a unique contributor to poor diabetes self-care [ , ]. Several studies have found that diabetes distress mediates the association between depressive symptoms and hemoglobin A1c (HbA1c) [ , ]. Further, diabetes distress is also a risk factor for depression [ ].
Similarly, in people with diabetes, anxiety and depression are associated with poorer diabetes self-management [- ] and increased disease severity and complications [ - ]. Of note, subthreshold depressive symptoms (which fall short of the full diagnostic criteria for MDD or dysthymia) are more common in people with diabetes than MDD, but are less likely to be diagnosed than MDD [ ]. Approximately half of the people with T2DM will experience at least one episode of subthreshold depression over 5 years [ ]. Subthreshold depressive symptoms in people with diabetes are associated with poorer quality of life [ , ] and poorer diabetes self-care [ ]. These findings have increased attention to addressing diabetes distress and mild depressive symptoms in patients with diabetes and may provide an avenue for preventing MDD [ , , ] and improving diabetes self-management [ , , ].
The international guidelines for diabetes management now recognize the importance of psychological care . Screening for both diabetes distress and depressive symptoms in people with diabetes, followed by appropriate interventions, has been recommended [ , ]. A stepped-care approach has been suggested, with mild distress or subthreshold depression managed within primary care, utilizing evidence-based self-help interventions [ , ]. Web-based interventions, predominantly clinician-assisted interventions, are effective for reducing comorbid depression in people with diabetes [ ]. However, many interventions for diabetes distress or depression rely upon face-to-face group sessions [ ], numerous telephone calls [ ], or clinician support [ ]. Overall, there is growing recognition of the need for low-intensity interventions to address diabetes distress or subthreshold depression, which are cost-effective and easily disseminated to large numbers of patients.
However, many people with diabetes are reluctant to seek mental health care . This may be due to the stigma of mental ill health and the belief that distress associated with diabetes is normal and should not be pathologized [ ]. Hence, interventions that do not explicitly refer to “depression” or “anxiety” may appeal more to some people with diabetes [ ].
Therapeutic writing is a brief intervention that aims to improve physical or mental health . The most common form of therapeutic writing is expressive writing (EW), in which thoughts and feelings regarding a stressful event are disclosed in writing, typically for 15-20 minutes for 3-4 days within a short period of time [ ]. EW has been examined in over 250 studies investigating its effects on physical or mental health in a wide range of populations, including healthy participants, people with psychological problems, and people with chronic health conditions [ , - ]. However, there are limitations to EW. Results of EW studies are quite variable, and the effect sizes are often small [ , ]. Further, EW often involves an immediate increase in distress [ ], even when followed by longer-term benefits [ ], which limits its suitability for wide dissemination without therapist support.
Indeed, trials of EW in people with diabetes have yielded mixed results. One study, for example, found fewer depressive symptoms at follow-up , while another found reduced stress but no effect on HbA1c levels, diabetes self-care, or diabetes distress [ ]. A later trial found that EW was associated with a worsening in depressive symptoms, with no change in diabetes distress [ ]. Of note, in the latter study, the EW task involved writing about any stressful experience over the past month rather than a diabetes-specific task.
Such findings have led researchers to investigate other variations of therapeutic writing, to maximize benefits and increase positive affect (and reduce distress) during the intervention. By modifying writing instructions, researchers have sought to increase the likelihood that participants engage in desired cognitive processes, with the aim of increasing the benefits gained from the writing task [, ]. The integration of positive psychology into therapeutic writing is one such modification [ ].
Benefit-finding writing (BFW) involves writing about any positive thoughts and feelings about a stressful experience such as an illness. Until recently, research has largely overlooked the utility of positively focused writing following stressful events or illness. However, there is emerging evidence that the experience of a medical illness often has sequelae that patients view as positive or beneficial . Benefit finding is defined as “identifying positive life changes resulting from adversity and negative life stressors, including illness” [ ]. It is correlated with posttraumatic growth [ ], which has recently been found to be associated with greater positive affect and less negative affect in the daily lives of people with some chronic health conditions [ ]. Benefit finding has been associated with increased psychosocial well-being and decreased depression in a range of clinical populations [ , ]. Benefit finding in diabetes has been associated with lower symptoms of depression, increased adherence to diabetes self-care, and greater perceived coping effectiveness [ ]. Therefore, interventions to increase benefit finding in people with diabetes could be useful [ ].
Trials of BFW in both nonclinical populations [, - ] and clinical populations [ - ] have demonstrated promising results. Compared to EW, BFW has been found to result in less distress and increased positive affect immediately postwriting [ , ]. Results from pilot trials in clinical populations suggest that BFW has benefits with regard to the symptoms of depression [ ], anxiety [ ], fatigue [ ], and the number of cancer-related medical appointments [ ]. In addition, positive affect journaling has recently been found to reduce symptoms of depression and anxiety in general medical patients with elevated anxiety [ ]. Thus, the limited research on BFW to date suggests that it may have the same longer-term health benefits as EW, but with the added advantage of immediate increases in positive affect. However, BFW has not been examined in people with diabetes.
To our knowledge, this is the first study of BFW in people in diabetes. We aimed to examine BFW in adults with T1DM or T2DM who were experiencing diabetes distress, but were not experiencing a mood or anxiety disorder. This is because people with diabetes who meet the criteria for MDD or an anxiety disorder are likely to require more intensive treatment, with several evidence-based interventions available . Our study included participants reporting mild or minimal symptoms of depression or anxiety, in addition to any degree of diabetes distress.
The primary aim of this randomized controlled trial (RCT) was to evaluate the efficacy of a Web-based BFW intervention for adults with T1DM or T2DM (compared to a CW condition) for reducing diabetes distress and increasing benefit finding in diabetes. The secondary outcomes examined were self-rated depression and anxiety symptoms, diabetes self-care, health, and health care utilization. It was hypothesized that participants randomized to the BFW condition would demonstrate significant decreases in diabetes distress and symptoms of depression and anxiety, and report reduced visits to health care professionals compared to those in the control group. It was also hypothesized that participants randomized to the BFW condition would experience significant increases in benefit finding in relation to diabetes and improvements in diabetes self-care and perceived health, relative to those in the control group, and that these effects would be evident at 1-month and 3-month follow-up assessments.
This study also aimed to conduct manipulation checks of the BFW intervention instructions by investigating immediate emotional responses to the writing tasks and their linguistic content. It was hypothesized that, compared to those in the CW group, participants in the BFW group would (1) have significantly greater increases in positive affect immediately after writing but not significantly different changes in negative affect; (2) use a significantly higher proportion of positive emotion words, negative emotion words, and cognitive processing words, in line with previous research ; and (3) rate their responses as significantly more personal and more meaningful, but not more distressing.
Finally, as people with diabetes can be reluctant to engage in psychological care , this study aimed to examine the feasibility of internet-based BFW by adults with T1DM or T2DM. Feasibility was assessed by adherence and acceptability of the intervention.
The design consisted of a CONSORT-EHEALTH –compliant, 2 (conditions) x 3 (time) RCT ( ). The protocol for the trial has been published [ ]. Participants were randomized to either internet-based BFW or internet-based CW (use-of-time writing). Both conditions involved an intervention of 3 days of online writing using the Writing for Health website (see below). Outcome measures were administered at three time points for both groups: baseline, 1 month, and 3 months postintervention. Self-rated current mood was also assessed immediately prior to and following each writing session. An online feedback questionnaire was administered postintervention.
Participants and Recruitment
Ethics approval was obtained from the Human Research Ethics Committee at St. Vincent’s Hospital, Sydney, Australia (HREC/13/SVH/379). All participants provided informed consent before engaging in any research-related activity. This study was prospectively registered with the Australia and New Zealand Clinical Trials Registry (ACTRN12615000241538).
The study was advertised from February 2015 to November 2016 on the websites, social media, and publications of Australian diabetes-related organizations and advertisements in waiting rooms of diabetes services and general practitioners throughout Sydney, Australia. Due to slow recruitment, the study was also advertised in Diabetes Australia’s printed Circle Magazine, which was mailed to 140,000 members. Finally, letters were mailed to 500 adult members of the National Diabetes Services Scheme in Australia, informing them of the study. In all advertising and recruitment materials, adults with T1DM or T2DM were directed to the Writing for Health website to find out more about the study.
Eligibility and exclusion criteria for adults with T1DM or T2DM living in Australia are shown in. The exclusion threshold was set at 8 or above on both the Generalized Anxiety Disorder-7 items (GAD-7) scale or the Patient Health Questionnaire-9 items (PHQ-9) scale, as sensitivity and specificity values have been found to be acceptable for a cutoff of 8 on the GAD-7 scale for identifying generalized anxiety disorder [ ] and the PHQ-9 scale for identifying MDD [ ]. Participants were not provided any compensation for taking part in the study.
Potential participants applied for the study via the open access Writing for Health website (). All stages of this study were conducted online through Writing for Health, including information about the study, consent to participate, screening questionnaires with automated feedback, participant registration, randomization to one of two conditions, the writing interventions, feedback questionnaire, and follow-up questionnaires at 1 month and 3 months postintervention.
Potential participants read the study information in Writing for Health and provided informed consent online (by checking a box) to participate. The study was described as investigating whether the writing exercises in the Writing for Health program improve the mental and physical well-being of people with diabetes. Both types of writing exercises were described to potential participants. However, the research hypotheses were not revealed.
Potential participants completed automated screening questionnaires, which also provided baseline data. Excluded applicants received an onscreen message informing them that the program is not suitable for them, with links to appropriate resources. All potential participants were provided with online feedback on the severity of their depression and anxiety symptoms.
Participants who met the eligibility criteria proceeded to online registration with the program, completed further online questionnaires (for further baseline data, see below), and were randomized to one of two online writing conditions (BFW or CW). Randomization (1:1) was automatically generated by Writing for Health after participants registered with the program. Although the randomization process was concealed from the researchers, they were not blinded to the condition of each participant. At the conclusion of the 3-day writing intervention, participants were invited to complete an online Feedback Questionnaire.
Automated reminder emails were sent by Writing for Health to participants on each day of their 3-day writing intervention and when it was time to complete their follow-up questionnaires at 1 month and 3 months postintervention. Participants were also provided automated feedback on the severity of their depression and anxiety symptoms at their online follow-up assessments.
Direct contact between participants and researchers did not occur in the standard course of the trial. However, in accordance with the study safety protocol , a study psychologist contacted a participant by email or telephone if he/she indicated distress after a writing session or scored in the severe range for depression or anxiety at the 1-month or 3-month follow-up, or indicated possible suicidal thoughts, to assess any need for psychological support and refer the participant to appropriate services, if required.
The online Writing for Health website was developed for this study by mental health researchers (including psychologists and a psychiatrist) at St Vincent’s Hospital, Sydney, Australia, and the University of New South Wales, and was hosted on a secure server at the university. Feedback on the usability of the website was previously provided by five adults with T1DM or T2DM during the development process.
Both the BFW and CW conditions involved participants in three 15-minute online writing sessions (once per day for 3 consecutive days), according to the instructions provided. Participants from both conditions continued to receive usual care from their health services.
Intervention Condition: Internet-Based Benefit-Finding Writing for Diabetes
Participants in the BFW condition were asked to write about any positive thoughts and feelings that they had had about their experiences with diabetes. The instructions () were adapted from those used by Stanton and colleagues (2002) in BFW for women with breast cancer [ ]. The same instructions were provided for all three writing sessions, consistent with previous studies of BFW [ , ].
Control Condition: Internet-Based Use-of-Time Writing (Control Writing)
Participants in the CW condition were asked to write in detail about how their time was spent that day (first writing session) and the plans for how their time will be spent the following day (second writing session) and week (third writing session). Participants were instructed to be as objective as possible and to focus on the facts and details of how their time was spent (or will be spent), and not to focus on their emotions. These CW instructions were adapted from the control conditions used in previous trials of therapeutic writing [, , , ].
Primary Outcome Measures
The 17-item Diabetes Distress Scale (DDS17)  is a self-report measure of psychosocial stress associated with diabetes, with four reliable subscales: Emotional Burden (feeling overwhelmed by diabetes), Physician-Related Distress (worries about access, trust, and care), Regime-Related Distress (concerns about diet, physical activity, and medications), and Interpersonal Distress (not receiving understanding and appropriate support from others). Cut-off points on the DDS17 have been established for little or no distress, moderate distress, and high distress [ ].
The 17-item Benefit Finding Scale  was originally developed to investigate benefit finding in women with early stage breast cancer. In the current study, the stem question is modified from, “Having had breast cancer has...” to “Having had diabetes has...” Participants were asked to respond to each of the 17 perceived benefits, such as “has lead me to be more accepting of things” and “has brought my family closer together” on a 5-point scale: 1 - not at all, 2 - a little, 3 - moderately, 4 - quite a bit, and 5 - extremely. This scale has previously been adapted for use in diabetes (with one item removed) and found to have one large factor and good internal consistency (Cronbach alpha=0.89) in a population of adolescents with T1DM [ ].
Secondary Outcome Measures
The PHQ-9  was used in the initial online screening to assess current symptoms of depression. It is a brief, widely used, reliable, valid 9-item self-report measure of both severity of depression over the past 2 weeks and is used to make a Diagnostic and Statistical Manual of Mental Disorders-IV criteria–based diagnosis of depression. It has established cut-off scores of 5, 10, 15, and 20, representing mild, moderate, moderately severe, and severe depression, respectively. The total score ranges between 0 and 27, with scores ≥10 having a sensitivity of 88% and a specificity of 88% for major depression [ ].
The GAD-7  was used in the initial online screening to assess the current symptoms of anxiety. It is a brief, widely used, reliable, and valid 7-item self-report measure of the severity of anxiety. Scores range from 0 to 21; scores of 5, 10, and 15 represent mild, moderate, and severe anxiety symptoms, respectively. A total score of 8 has been identified as an important threshold for identifying the presence of an anxiety disorder [ ].
The Summary of Diabetes Self-Care Activities (SDSCA) measure (revised) is an 11-item self-report measure of self-care of diabetes (including diet, exercise, blood sugar testing, foot care, and smoking) widely used both clinically and in research . Items in the revised version were selected based on their psychometric properties, sensitivity to change, and ease of scoring and interpretation [ ]. In a critical appraisal of 26 different measures of diabetes outcomes, the SDSCA measure (revised) was one of only three measures that meets all the criteria of suitability, validity, reliability, and sensitivity to change [ ].
Self-rated health was assessed by the question, “In general, how would you rate your health at present?” The five response options were: very good, good, fair, poor, and very poor. Responses to this question have previously been found to be significantly associated with the blood glucose indicator HbA1c (poorer self-rated health associated with higher HbA1c levels) and number of self-reported diabetes-related symptoms in patients with T2DM .
Health care utilization was assessed by the question “In the past month, how many times have you visited a doctor or other health care professional?” at three time-points: baseline, 1-month follow-up, and 3-month follow-up.
The measures below were assessed to conduct manipulation checks.
International Positive and Negative Affect Schedule—Short Form
The International Positive and Negative Affect Schedule Short Form (I-PANAS-SF) is a reliable and valid 10-item measure of positive and negative affect, comprised of 10 words that represent positive and negative affect . The values of the correlations between this short-form and the positive and negative affect scales of the full 20-item form of the PANAS are 0.92 and 0.95, respectively [ ]. Instructions were modified to assess state rather than trait affect, using the instructions of the 20-item PANAS - Immediate Version [ ]. Participants were instructed to indicate the degree of specific affect they feel “right now, at the present moment” on a scale of 1 to 5 (1=very slightly/not at all, 5=extremely).
Essay Evaluation Measure
The Essay Evaluation Measure is a 3-item measure adapted from previous studies [, ], which asks participants to rate, immediately after each writing session, how meaningful, personal, and distressing their writing exercise was, on a 7-point scale (0=not at all, 6=extremely). Similar questions have been used as manipulation checks in previous therapeutic writing studies [ , ]. Immediately after each writing session, participants in the BFW intervention condition were also asked if they were able to identify any positive thoughts or feelings about living with diabetes in their writing session.
The Feedback Questionnaire is a 12-item self-report measure developed to assess participants’ experiences and perceptions of the Writing for Health program. Item content was informed by self-report measures from other evaluations of internet-based interventions [, ] and included its usability, ease of use, credibility, and most and least helpful aspects.
Sociodemographic information (age, gender, education, and occupation) and diabetes-related information (type, duration of illness, management, and complications) were also collected.
Statistical analyses were conducted using SPSS 22 (IBM Corp, Armonk, NY). Group differences in baseline characteristics were analyzed with two-sample t tests and Chi-square tests. The 1-month and 3-month follow-up data were analyzed on an intention-to-treat (ITT) basis using the SPSS mixed procedure. Time was treated as a repeated measures factor (3 levels: T1, T2, T3), with an unstructured residual variance-covariance matrix, and with fixed effects for group, time, and time-by-group interactions. This procedure allowed inclusion of individuals with missing data at follow-up assessments .
Planned contrasts were used to examine differences between the groups in changes in outcome variables between baseline and each of the two follow-up occasions and changes over time within groups. All effects were tested at P<.05. Within-group and between-group Cohen d effect sizes for the primary and secondary outcome measures were also calculated.
Prior to outcome analyses, the distribution of variables was examined for skewness, with variables transformed (log or square root) as needed to reduce skewness to between -1 and 1. All four DDS17 subscale scores and the Blood Sugar Testing subscale of the SDSCA required log transformation; the DDS17, PHQ-9 and GAD-7 total scores required square root transformation. Outliers (scores more than 3 SDs above the mean) were addressed using the winsorizing technique, in which outliers are replaced with the mean value plus 3 SDs . Winsorizing is recommended to reduce the impact of outliers on the Type 1 error rate [ ]. Only one outlier value was identified and winsorized.
In addition, manipulation check analyses were conducted to validate the writing intervention instructions in three ways:
- Scores on the Essay Evaluation Measure were compared between the two groups using analysis of variance (ANOVA).
- To examine immediate emotional responses to the writing interventions, scores on the I-PANAS-SF [ ] administered immediately before and after each writing session were analyzed using a 2 (group) x 3 (session) x 2 (positive affect and negative affect) repeated measures multivariate ANOVA.
- The content of the written scripts in both groups was assessed using the Linguistic Inquiry Word Count 2015 software program (2015) [ ] to examine for differences in positive emotion words and cognitive insight words. This validated method provides a content analysis of the language used in the scripts, quantifying the number of words used from specific categories (eg, emotions and cognitive processes) and has previously been used in numerous studies of therapeutic writing [ , , - ]. Linguistic Inquiry Word Count scores for word categories were compared between the two groups using ANOVAs.
Sample Size and Power Analysis
For this preliminary study, we powered the study to detect a moderately large effect size (Cohen d) of 0.7, consistent with previous BFW studies [, ]. Based on a statistical power of 0.80 and a probability level (Cronbach alpha) of .05, a sample size of 34 per group (ie, 68 for the two groups) was needed for two-tailed tests to detect an effect size (Cohen d) of 0.7. Given the expected attrition rate of approximately 34% [ ], our target total sample size was 104.
However, the unexpected difficulty encountered in recruiting participants to the trial caused extensive delays in the study timeline. Many different methods were employed to recruit the required sample size, and the recruitment period was extended to assist the effort. Yet, by the end of the extended recruitment period, only 88 of the 104 participants had been recruited (BFW: n=37; CW: n=51). Of these, 72 participants (BFW: n=24; CW: n=48) had completed their baseline measures and were therefore eligible for inclusion in ITT analyses. Revised power analyses were conducted to determine the effect sizes that could be detected with the obtained sample size. Moreover, the unintended unequal allocation of participants eligible for analyses in the two groups (resulting in a 2:1 ratio) would further reduce the statistical power, as a 2:1 ratio requires 12% more participants than a trial using 1:1 to detect the same effect size with equivalent power . Therefore, for ITT analyses in this study (BFW: n=24; CW: n=48), moderately large effect sizes of Cohen d ≥0.73 could be detected by two-tailed tests, given a statistical power of 0.80 and a Cronbach alpha of .05.
Participant Recruitment and Attrition
A participant flow diagram is displayed in. Over a 21-month period, only 169 individuals provided consent to commence the screening procedure during that period. Of these, 162 commenced the online screening procedure and 102 were eligible to participate, 88 of whom proceeded to online registration and were randomized to either the BFW group (n=37) or the CW group (n=51). Of these participants, 25 allocated to the BFW group and 48 allocated to the CW group completed baseline assessment. One participant in the BFW group formally withdrew from the study after completing the baseline questionnaires but before commencing the intervention. Thus, 24 participants in the BFW group and 48 in the CW group were eligible for inclusion in the ITT analyses (N=72).
At the 1-month follow-up, 21 BFW participants and 33 CW participants completed the questionnaires, while at the 3-month follow-up, 17 BFW participants and 23 CW participants completed the questionnaires. Within the BFW group, 21 participants completed all three writing sessions (88% of those who completed baseline data and 95% those who completed the first writing session). In comparison, 37 of the CW participants completed all three writing sessions (77% of those who completed baseline data and 86% of those who completed the first writing session). There was no significant difference between the BFW group (95%) and the CW group (86%) in the proportion of participants who completed all three writing sessions, of those who commenced the first session (χ21=1.341, P=.25).
displays the demographic and clinical baseline characteristics, and shows the outcome measures at baseline. A total of 46 participants (64% of the sample) had T1DM and the remaining 26 participants (36%) had T2DM. The majority were women (n=57, 79%) and highly educated, with 63% (n=45) reporting that they had a university degree. Over one-third of the sample was retired (n=27, 38%). The mean response score of 3.96 (SD 0.82) on the current health item (ranging from 1=very poor to 5=very good) corresponded to “good” self-reported current health.
|Characteristic||Total sample (N=72)||BFWa (n=24)||CWb (n=48)||P valuec|
|Age (years), mean (SD)||53.79 (15.93)||53.54 (17.26)||53.92 (15.41)||.80|
|Female, n (%)||57 (79)||20 (83)||37 (77)||.64|
|Married/de facto, n (%)||44 (62)||15 (65)||29 (60)||.51|
|Employed (full-time or part-time), n (%)||30 (42)||9 (39)||21 (44)||.40|
|Retired, n (%)||27 (38)||10 (44)||17 (35)||.29|
|University educated, n (%)||45 (63)||14 (61)||31 (65)||.72|
|Type 1 diabetes, n (%)||46 (64)||15 (62)||31 (65)||.87|
|Years since diabetes diagnosis, mean (SD)||17.69 (14.26)||16.86 (13.68)||18.08 (14.65)||.78|
|Diabetes complications, n (%)||16 (22)||4 (17)||12 (22)||.69|
|Diabetes managementd, n (%)|
|Insulin||51 (71)||17 (71)||34 (71)||.98|
|Medication||24 (33)||6 (25)||18 (38)||.60|
|Blood sugar testing||67 (93)||24 (100)||43 (90)||.68|
|Healthy eating plan||59 (82)||18 (75)||41 (85)||.73|
|Exercise||54 (75)||16 (67)||38 (79)||.61|
|Current antidepressant medication, n (%)||13 (18)||6 (25)||7 (15)||.24|
|Ever been depressed, n (%)||33 (46)||12 (50)||21 (44)||.50|
|Ever had a mental illness diagnosed, n (%)||22 (31)||7 (29)||15 (31)||.88|
aBFW: benefit-finding writing.
bCW: control writing.
cP values for variables based on means are from a two-sample t test. P values for variables based on percentages are from a Chi-square test.
dCategories were not mutually exclusive and participants could indicate more than one management strategy to best describe their situation.
As reported in, participants’ baseline levels of depression, anxiety, and diabetes distress were very low. Baseline scores on the PHQ-9 (mean 2.25, SD 1.62) indicated that most participants were in the “minimal to none” range (0 to 4) for depression symptoms [ ]; similarly, baseline scores on the GAD-7 (mean 1.71, SD 1.73) indicated that most participants were in the “minimal to none” range (0 to 4) for anxiety symptoms [ ]. Further, baseline scores on the DDS (mean 1.74, SD 0.66) and all four subscales of the DDS17 (means ranging from 1.65 to 1.95) were in the range of “DDS < 2.0: little or no diabetes distress” [ ]. However, almost one half of the sample (n=33, 46%) reported ever experiencing depression for 2 weeks or more in the past ( ).
There were no significant differences between the intervention (BFW) and control (CW) groups in any baseline characteristics (demographics, clinical characteristics, or outcome measures), with the exception of a single item of the SDCSA measure (revised) assessing fruit and vegetable consumption (t70=2.484, P=.02). The CW group had a significantly higher mean number of days (5.58 days, SD 1.41 days) than the BFW group (4.30 days, SD 2.34 days) in which they ate five or more serves of fruits and vegetables over the past 7 days.
|Measure||Total sample (N=72)||BFWa (n=24)||CWb (n=48)||P valuec|
|Primary outcome measures|
|Benefit Finding Scale, mean (SD)||43.81 (16.67)||39.83 (17.16)||45.79 (16.23)||.45|
|Diabetes Distress Scale, mean (SD)|
|Total score||1.73 (0.64)||1.74 (0.66)||1.73 (0.63)||.60|
|Emotional Burden||1.65 (0.68)||1.66 (0.66)||1.65 (0.70)||.49|
|Physician-Related Distress||1.68 (0.71)||1.70 (0.70)||1.67 (0.71)||.39|
|Regimen-Related Distress||1.73 (0.66)||1.80 (0.71)||1.69 (0.64)||.67|
|Interpersonal Distress||1.95 (0.87)||1.83 (0.77)||2.01 (0.92)||.56|
|Secondary outcome measures|
|PHQ-9d, mean (SD)||2.11 (1.59)||2.25 (1.62)||2.04 (1.58)||.31|
|GAD-7e, mean (SD)||1.60 (1.71)||1.71 (1.73)||1.54 (1.71)||.24|
|Revised Summary of Diabetes Self-Care Activities|
|General Diet, mean (SD)||5.29 (1.68)||4.93 (2.14)||5.46 (1.41)||.86|
|Specific Diet—Fruit and Veg, mean (SD)||5.17 (2.10)||4.30 (2.34)||5.58 (1.87)||.02f|
|Specific Diet—High-Fat Foods, mean (SD)||3.39 (2.16)||3.61 (2.45)||3.29 (2.03)||.56|
|Exercise, mean (SD)||3.79 (2.26)||3.39 (2.28)||3.98 (2.25)||.45|
|Blood Glucose Testing, mean (SD)||5.54 (2.28)||5.50 (2.27)||5.56 (2.31)||.76|
|Foot Care, mean (SD)||2.53 (2.52)||2.17 (2.67)||2.70 (2.45)||.19|
|Smoking Status, n (%)||7 (10)||1 (4)||6 (13)||.23|
|Self-reported health, mean (SD)||3.96 (0.82)||4.00 (0.80)||3.94 (0.84)||.78|
|Number of visits to a health professional in the past 30 days, mean (SD)||1.82 (2.17)||2.04 (1.97)||1.71 (2.28)||.56|
aBFW: benefit-finding writing.
bCW: control writing.
cP values for variables based on means are from a two-sample t test. P values for variables based on percentages are from a Chi-square test.
dPHQ-9: Patient Health Questionnaire-9 items.
eGAD-7: Generalized Anxiety Disorder-7 items.
fStatistically significant at P<.05.
Essay Evaluation Measure
Consistent with predictions, participants in the BFW group rated their writing exercises as significantly more personal (F1,56=6.00, P=.02, Ƞp2=0.09) and significantly more meaningful (F1,56=6.87, P=.01, Ƞp2=0.11) than those in the CW group. The two groups did not differ significantly in terms of the ratings of how distressing the writing exercises were (F1,56=2.76, P=.10, Ƞp2=0.05).
The majority of participants in the BFW group reported being able to identify positive thoughts or feelings about diabetes in their writing sessions: 20 (91%) in Writing Session 1, 21 (100%) in Writing Session 2, and 20 (95%) in Writing Session 3.
Linguistic Content Analyses
Consistent with predictions, participants in the BFW group used a significantly greater proportion of positive emotion words (F1,56=128.37, P<.001, Ƞp2=0.70), negative emotion words (F1,56=55.23, P<.001, Ƞp2=0.50), causal words (F1,56=49.71, P<.001, Ƞp2=0.47), and cognitive insight words (F1,56=180.32, P<.001, Ƞp2=0.76) than those in the CW group. The two groups did not differ in the total number of words used per session (F1,56=0.13, P=.72, Ƞp2=0.00).
Positive and Negative Affect
Change scores (post-pre) for the Positive Affect and Negative Affect subscales of I-PANAS-SF were computed for each of the three writing sessions (). As predicted, the BFW group had significantly greater increases in positive affect than the CW group (F1,56=7.76, P=.01 , Ƞp2=0.12). The increases in negative affect following writing did not differ significantly between the BFW and CW groups (F1,56=2.37, P=.13 , Ƞp2=.04).
Effect of Writing Interventions on Primary Outcome Measures
Between-Group Differences in Primary Outcome Measures at 1-Month and 3-Month Follow-Ups
presents the observed means and estimated marginal means for the primary outcome measures at baseline and 1-month and 3-month follow-ups. reports the between-group results of the SPSS mixed procedure analyses and the Cohen d effect sizes. displays the estimated means of the primary outcome measures in the two groups at baseline, 1-month follow-up, and 3-month follow-up.
There were no significant group-by-time interactions in the mixed model analyses for the BFS or DDS17 scores (F2,45.75=2.18, P=.12 and F2,45.75=0.34, P=.97, respectively). As displayed in, there were no significant group-by-time interactions for any of the planned contrasts examining differences between the BFW and CW groups in changes from baseline to 1-month follow-up or 3-month follow-up for the BFS, total DDS17 score, or any of the four subscales of the DDS17 (all P>0.05). In addition, as displayed in , all between-group Cohen d effect sizes for the primary outcomes were minimal to small.
|Primary outcome for each group||Observed means (SD)||Estimated meansa (SD)|
|Baselineb||1-month follow-upc||3-month follow-upd||Baseline||1-month follow-up||3-month follow-up|
|Benefit Finding Scale|
|BFWe||39.83 (17.16)||39.91 (18.68)||40.06 (18.22)||39.83 (16.52)||38.10 (18.78)||37.61 (19.07)|
|CWf||45.79 (16.23)||46.97 (19.12)||49.09 (19.77)||45.79 (16.56)||45.04 (19.61)||49.29 (20.58)|
|Diabetes Distress Scale Total|
|BFW||1.74 (0.66)||1.62 (0.56)||1.54 (0.65)||1.30 (0.24)||1.25 (0.20)||1.23 (0.24)|
|CW||1.73 (0.63)||1.67 (0.57)||1.57 (0.48)||1.30 (0.21)||1.26 (0.21)||1.24 (0.28)|
|Diabetes Distress Scale Subscales|
|BFW||1.66 (0.66)||1.75 (0.81)||1.58 (0.62)||0.44 (0.39)||0.48 (0.44)||0.42 (0.39)|
|CW||1.65 (0.70)||1.83 (0.77)||1.79 (0.61)||0.43 (0.34)||0.51 (0.48)||0.52 (0.48)|
|BFW||1.70 (0.70)||1.42 (0.91)||1.44 (0.81)||0.27 (0.20)||0.12 (0.20)||0.15 (0.20)|
|CW||1.67 (0.71)||1.34 (0.49)||1.26 (0.41)||0.27 (0.21)||0.14 (0.21)||0.11 (0.28)|
|BFW||1.80 (0.71)||1.74 (0.69)||1.58 (0.57)||0.52 (0.34)||0.48 (0.34)||0.42 (0.39)|
|CW||1.69 (0.64)||1.67 (0.66)||1.57 (0.63)||0.47 (0.28)||0.43 (0.35)||0.40 (0.35)|
|BFW||1.83 (0.77)||1.46 (0.59)||1.55 (1.02)||0.39 (0.29)||0.24 (0.29)||0.21 (0.34)|
|CW||2.01 (0.92)||1.87 (1.07)||1.59 (0.89)||0.43 (0.28)||0.32 (0.35)||0.27 (0.35)|
aEstimated means based on intention-to-treat sample of benefit-finding writing (n=24) and control writing (n=48), with variables transformed (log or square root) to address skewness, SDs, and CIs shown in parentheses.
bBenefit-finding writing: n=24, Control writing: n=48.
cBenefit-finding writing: n=21, Control writing: n=33.
dBenefit-finding writing: n=17, Control writing: n=23.
eBFW: benefit-finding writing.
fCW: control writing.
|Outcome||Baseline to 1-month follow-upa||Baseline to 3-month follow-upb|
|Estimated mean difference (95% CI)||P value||Cohen d (95% CI)||Estimated mean difference (95% CI)||P value||Cohen d (95% CI)|
|Benefit Finding Scale||0.98 (–5.19 to 7.15)||.32||–0.06 (–0.63 to 0.50)||5.72 (–0.52 to 11.97)||.07||–0.34 (–0.80 to 0.19)|
|Diabetes Distress Scale, total score||0.01 (–0.07 to 0.09)||.82||–0.05 (–0.54 to 0.44)||0.01 (–0.06 to 0.09)||.83||–0.05 (–0.53 to 0.45)|
|Diabetes Distress Scale subscales|
|Emotional Burden||0.05 (–0.16 to 0.25)||.65||–0.11 (–0.59 to 0.39)||0.11 (–0.07 to 0.29)||.23||–0.31 (–0.79 to 0.20)|
|Physician-Related Distress||0.02 (–0.10 to 0.13)||.74||–0.10 (–0.59 to 0.38)||–0.05 (–0.15 to 0.06)||.38||0.19 (–0.31 to 0.67)|
|Regimen-Related Distress||–0.00 (–0.12 to –0.11)||.95||–0.02 (–0.49 to 0.47)||0.02 (–0.11 to 0.16)||.74||–0.10 (–0.58 to –0.40)|
|Interpersonal Distress||0.04 (–0.11 to 0.19)||.61||–0.14 (–0.63 to 0.36)||0.01 (–0.14 to 0.16)||.88||–0.07 (–0.56 to 0.42)|
aDifference = (baseline – 1-month follow-up [for benefit-finding writing]) – (baseline – 1-month follow-up [for control writing]).
bDifference = (baseline – 3-month follow-up [for benefit-finding writing]) – (baseline – 3-month follow-up [for control writing]).
Within-Group Differences in Primary Outcome Measures at 1-Month and 3-Month Follow-Ups
displays the results of mixed procedure analyses for within-group changes on the primary outcomes from baseline to 1-month and 3-month follow-ups. There were no significant within-group changes from baseline in terms of benefit finding in relation to diabetes (BFS scores). However, significantly lower diabetes distress (DDS17 total score) was observed in both groups at the 3-month follow-up, but not the 1-month follow-up, compared to the baseline. The within-group changes in the four subscales of DDS17 were also examined. Both the BFW and CW groups displayed significant within-group reductions in two subscales of the DDS17 (Physician-Related Distress and Interpersonal Distress) at both 1-month and 3-month follow-ups. As shown in , medium effect sizes were observed for these within-group DDS17 subscale changes. However, neither group displayed any significant within-group changes in the other two subscales of the DDS17 (Emotional Burden and Regimen-Related Distress).
|Outcome and group||Baseline to 1-month follow-upa||Baseline to 3-month follow-upb|
|Estimated mean difference (95% CI)||P value||Cohen d effect sized (95% CI)||Estimated mean difference (95% CI)||P value||Cohen d effect size d (95% CI)|
|Benefit Finding scale|
|BFWc||1.74 (–3.13 to 6.60)||.48||0.10 (–0.47 to 0.66)||2.22 (–2.59 to 7.03)||.48||0.12 (–0.44 to 0.69)|
|CWd||0.76 (–3.04 to 4.55)||.69||0.04 (–0.36 to 0.44)||–3.50 (–7.48 to 0.48)||.08||–0.19 (–0.59 to 0.21)|
|Diabetes Distress Scale total|
|BFW||0.05 (–0.20 to 0.11)||.14||0.23 (–0.58 to 1.03)||0.07 (0.01 to 0.13)||.03||0.29 (–0.51 to 1.10)|
|CW||0.04 (–0.01 to 0.08)||.13||0.19 (–0.38 to 0.76)||0.06 (0.01 to 0.11)||.02||0.24 (–0.33 to 0.81)|
|Diabetes Distress Scale subscales|
|BFW||–0.04 (–0.20 to 0.13)||.65||–0.10 (–0.90 to 0.70)||0.02 (–0.11 to 0.16)||.75||0.05 (–0.75 to 0.85)|
|CW||–0.08 (–0.21 to 0.04)||.18||–0.19 (–0.76 to 0.38)||–0.09 (–0.20 to 0.03)||.13||–0.22 (–0.78 to 0.35)|
|BFW||0.15 (0.06 to 0.24)||<.001||0.75 (–0.08 to 1.58)||0.11 (0.03 to 0.20)||.007||0.60 (–0.22 to 1.42)|
|CW||0.13 (0.06 to 0.20)||<.001||0.62 (–0.04 to 1.20)||0.16 (0.09 to 0.22)||<.001||0.65 (–0.53 to 1.23)|
|BFW||0.03 (–0.06 to 0.13)||.49||0.12 (–0.68 to 0.92)||0.09 (–001 to 0.20)||.08||–0.27 (–1.08 to 0.53)|
|CW||0.03 (–0.04 to 0.11)||.36||0.13 (–0.44 to 0.69)||0.07 (–0.02 to 0.16)||.11||–0.22 (–0.79 to 0.35)|
|BFW||0.15 (0.03 to 0.27)||.02||0.52 (–0.30 to 1.33)||0.17 (0.06 to 0.29)||.004||0.57 (–0.25 to 1.39)|
|CW||0.11 (0.02 to 0.22)||.02||0.35 (0.22 to 0.95)||0.16 (0.07 to 0.26)||<.001||0.51 (0.07 to 1.08)|
aChange = baseline – 1-month follow-up.
bChange = baseline – 3-month follow-up.
cBFW: benefit-finding writing.
dCW: control writing.
Effect of Writing Interventions on Secondary Outcome Measures
Between-Group Differences in Secondary Outcome Measures at 1-Month and 3-Month Follow-Ups
includes the observed means and estimated marginal means for the secondary outcome measures at baseline and follow-ups. The results for the between-group differences in secondary outcomes are displayed in . There was only one significant group-by-time interaction for any of the secondary outcome measures examined. For self-reported health, there was a significant difference between the BFW and CW groups in changes from baseline to 1-month follow-up (P=.04), reflecting an increase in self-reported health in the CW group and a decrease in the BFW group (neither of these within-group changes were significant). However, there was no significant difference between the two groups in changes in self-reported health from baseline to the 3-month follow-up (P=.81). There were no significant group-by-time interactions for scores of the PHQ-9, GAD-7, revised SDSCA, or health care utilization.
Within-Group Differences in Secondary Outcome Measures at 1-Month and 3-Month Follow-Ups
The results for the within-group differences in secondary outcomes are displayed in. Both the BFW and CW groups displayed significant increases in depression (PHQ-9 scores) and anxiety (GAD-7) scores over time. In both the BFW and CW groups, there were significant within-group increases in both the PHQ-9 and GAD-7 scores from baseline to 1-month follow-up and a significant increase in GAD-7 scores from baseline to the 3-month follow-up. The CW group also displayed a significant increase in PHQ-9 scores from baseline to the 3-month follow-up. The effect sizes for these within-group changes ranged from small to moderately large. The observed mean PHQ-9 and GAD-7 scores in both groups at both follow-ups remained in the nonclinical range (2.79 to 4.47), which is in the “none-minimal” range (<5) for depression and anxiety severity.
There were few other significant within-group changes in the secondary outcomes. In the BFW group, there was a significant decrease in Exercise scores on the revised SDSCA at the 1-month follow-up, but not at the 3-month follow-up. In the CW group, there was a significant increase in the Foot Care subscale of the revised SDSCA at the 1-month follow-up, but not at the 3-month follow-up. There were no significant within-group changes over time in either group for the General Diet, Specific Diet, or Blood Glucose Testing subscales of the revised SDSCA or self-reported health or health care utilization.
Mean distress ratings reported after each session were low for both groups in all three writing sessions, ranging from 0.14 to 1.05. Two participants in the BFW group (8%) and two in the CW group (4%) were contacted by email following a high distress rating (≥5/6); all four participants reported that their distress was short-lived. Four participants in the BFW group (17%) and two in the CW group (4%) were contacted by telephone due to elevated scores on the PHQ-9 or GAD-7 at follow-up(s), in line with the study safety protocol . One participant (in the BFW group) was referred to a mental health service after reporting a recurrence of previous depression at the 3-month follow-up. All six participants contacted by telephone had reported a history of depression or a diagnosis of a mental illness at baseline. There were no privacy breaches or technical difficulties during the trial.
The BFW group (n=22) and CW group (n=38) did not differ in their mean total score for the Feedback Questionnaire (BFW: mean 16.18, SD 4.89; CW: mean 15.36, SD 5.34; t58=0.59, P=.56). In the BFW group, 71% (n=15) of the participants reported that they were “mostly” to “very” satisfied with the Writing for Health program compared to 55% (n=21) in the CW group (χ21=1.47, P=.23). Approximately half of the participants in both the BFW group (n=11, 51%) and CW group (n=18, 47%) reported that they would be “mostly” to “very” confident in recommending Writing for Health to a friend with diabetes (χ21=0.14, P=.71). Further, approximately one-third of both groups (BFW: n=7, 33%; CW: n=13, 34%) reported that the writing exercises were “mostly” to “very” helpful in reducing stress (χ21=0.01, P=.95).
This RCT examined the efficacy of BFW, compared to a CW condition, among adults with T1DM or T2DM who were not currently experiencing MDD or an anxiety disorder. It was hypothesized that participants randomized to the BFW group, compared to those in the CW group, would have significant increases in benefit finding, significant reductions in diabetes distress (primary outcomes), and significant improvements in the secondary outcomes at both the 1-month and 3-month follow-ups. However, these hypotheses were not supported by our results. In addition, there were no significant intervention effects on the primary outcomes.
All the hypotheses for the validation analyses were supported by our results. This suggests that the results for the outcome measures cannot be explained by participants not following the writing instructions or failing to engage with the writing tasks as expected. Specifically, consistent with other BFW studies [, , ], participants in the BFW group rated their writing sessions as significantly more personal and meaningful than those in the CW group. Second, as in previous findings [ , , ], the BFW group had a significantly greater increase in positive affect following writing, relative to the CW group. Third, consistent with previous findings, linguistic analyses revealed that BFW participants had a greater use of positive emotion words [ , - ], negative emotion words [ , ], and cognitive processing words [ , , , ] than CW participants. Further, the majority of BFW participants reported that they were able to identify at least one positive aspect of living with diabetes in their writing session that day. Overall, these findings suggest that the BFW participants followed their writing instructions and, as expected, engaged in both emotional expression and cognitive processing to a greater degree than control group participants. However, there were no significant between-group differences in the primary and secondary outcome variables. A second review of the literature produced several potential explanations.
To our knowledge, this was the first study to examine benefit finding as an outcome of therapeutic writing, so it is not possible to directly compare this finding with previous trials. However, posttraumatic growth, a concept closely related to benefit finding , has been found to increase following BFW to a greater degree than following EW [ ]. Further, more intensive interventions have previously been found to increase benefit finding in medical populations [ , ]. For example, in adults with T2DM, benefit finding was found to increase following 14 sessions of telephone health coaching [ ]. Our BFW intervention (of three 15-minute sessions) was substantially briefer, with less clinician contact. Although the intervention brevity and content were chosen to avoid potential resistance to psychological care by people with diabetes, it is conceivable that a more intensive, or simply more efficacious, intervention is required to increase benefit finding in people with diabetes.
Contrary to our hypotheses, the BFW group did not show significant reductions in diabetes distress, compared to the control group, at either the 1-month or 3-month follow-up. Rather, both groups had significant within-group reductions in diabetes distress, specifically in Physician-Related Distress and Interpersonal Distress, which are related to diabetes-related support from health professionals and family and friends, respectively. Thus, one possibility is that the process of participating in a research study about experiences of living with diabetes, albeit with limited clinician support, was sufficient to increase the level of perceived social support in relation to diabetes, regardless of the writing intervention instructions. This explanation is consistent with the findings of the REDEEM trial, in which three brief online interventions, all accompanied by support telephone calls, resulted in significantly reduced diabetes distress in nondepressed adults with T2DM . As suggested by the authors [ ], it appeared that nondepressed adults with diabetes distress were highly responsive to professional attention and normalization of diabetes distress. The finding of a small reduction in diabetes distress in both groups in this study is consistent with the findings of a meta-analysis of interventions for reducing diabetes distress [ ], which found that even generalist and psychoeducation interventions resulted in small reductions in diabetes distress in adults with T1DM or T2DM.
Further, the lack of a significant reduction in diabetes distress, compared to a control group, is consistent with the findings of an RCT of standard EW for adults with T2DM  and a pilot RCT of EW about stressors occurring in the past month in nondepressed adults with T2DM [ ]. It is also in line with the meta-analysis showing that only interventions of six or more sessions resulted in significant reductions in diabetes distress compared to a control group [ ]. Thus, overall, it appears that while attention to diabetes distress and a generalist intervention may be sufficient to result in small reductions in diabetes distress in nondepressed adults, a longer or more efficacious intervention is required to provide significant reductions in diabetes distress compared to an active control condition.
This study did not find the predicted improvements in the secondary outcomes of depression, anxiety, diabetes self-care, self-reported health, and health care utilization. Of note, participants in both groups demonstrated significant increases in mean depression and anxiety scores over a period of 3 months, albeit with mean scores remaining in the “none-minimal” range for symptoms. There were no significant between-group differences in these changes. This finding was inconsistent with previous findings of a reduction in symptoms of depression or anxiety following diabetes-specific EW , writing about use of time [ ], or online positive affect journaling in medical patients with elevated anxiety [ ]. Given that the baseline depression and anxiety scores in our study were so low (2.11 and 1.60, respectively), the increase in mean scores may reflect a regression to the mean [ ]. The baseline means were unexpectedly substantially below the inclusion criteria cut-off scores of 8 and lower than the normative means of adults with T1DM or T2DM in Australia (5.6 to 7.7 on the PHQ-9 and 4.0 to 5.3 on the GAD-7) [ ]. Further, half of the sample reported a history of depression, and it is known that the adults with diabetes and past depression are at an increased risk of a recurrence of depression [ ]. A minority of participants with a history of depression experienced worsening of depression or anxiety symptoms over the 3-month study period, which may have accounted for the increase in mean scores in both groups. Regardless of the reason for these changes, this finding highlights the need for monitoring of depression and anxiety in trial participants with diabetes and past depression.
Given that one of the key potential benefits of BFW, relative to EW, is the absence of short-term distress that typically accompanies EW , it was important for this study to assess immediate emotional responses to BFW in people with diabetes. This study found that participants in the BFW group reported no more distress than those writing about use of time and had greater increases in positive affect immediately after writing. This supports the hypothesis that BFW may be more suitable for online dissemination than forms of therapeutic writing that typically lead to short-term distress [ ]. However, participants in this study had few symptoms of depression or anxiety, and it is not known how depressed adults with diabetes would respond to BFW.
Feasibility of the Intervention
Feasibility of Web-based BFW for adults with T1DM or T2DM was assessed by adherence to and acceptability of the intervention. Adherence to BFW was high, with 95% of those who commenced the first writing session also completing all three writing sessions. Further, manipulation checks indicated that those in the BFW group were able to write about the positive aspects of diabetes and engaged with the intervention as intended. However, acceptability of the intervention was only moderate. Although over two-thirds of the participants in the BFW group reported satisfaction with the Writing for Health program, only half of both groups would recommend it to a friend.
Further, given the small sample recruited over a 21-month period, there was no high level of demand for Web-based BFW in adults with diabetes. Other studies of therapeutic writing in participants with diabetes have also reported difficulty in recruiting their target sample sizes [, ]. This is in line with the observation that participants’ reasons for undertaking unfacilitated therapeutic writing are often unclear, particularly among those with a chronic illness [ ].
Thus, this study identified some limitations to the feasibility of Web-based BFW for adults with diabetes, with apparent low interest in the intervention and only moderate acceptability, despite high adherence among those who commenced the writing sessions. Recent qualitative findings (published subsequent to this study) regarding the perceptions of written reflection about T2DM  have shed some light on the appropriateness of therapeutic writing for adults with diabetes. Specifically, it was found that while some adults with T2DM reported that writing about diabetes increased their commitment to diabetes self-management, others perceived that a written reflection was inapplicable to their diabetes. Some found that writing about diabetes was difficult, the timing in terms of their disease trajectory or life priorities was inappropriate, and they required a meeting with a diabetes nurse in order to benefit from the task. The authors concluded that written reflection about diabetes may only be suitable for some adults with T2DM and that it may be more appropriate as part of a “blended” approach rather than a standalone intervention [ ]. The findings from this study regarding the feasibility of BFW for diabetes support this recommendation.
Limitations and Future Directions
There were several limitations to this study. A clear limitation was the small sample size, which reduced the statistical power such that only moderately large effect sizes for differences between the two groups, if present, could have been detected. Second, given the preliminary nature of this trial, multiple comparisons of several independent variables were conducted without control of the alpha levels to reduce the risk of Type 1 errors (ie, “false positive”). However, given that there were few significant results, this is unlikely to be a significant issue.
Perhaps, most importantly, the generalizability of the results of this study was limited by the profile of the sample who registered to participate. Although the conservative eligibility criteria were set in line with the exploratory nature of the study, baseline levels of diabetes distress, depression, and anxiety were low, leaving very little room for improvement. It has recently been suggested that therapeutic writing may offer the most benefit for those with moderate levels of symptoms, as those with very few symptoms cannot improve their symptoms with this treatment and those with severe symptoms may require stronger treatment . Therefore, future trials of therapeutic writing in adults with diabetes should apply a minimum threshold for diabetes distress, to examine its efficacy among those with elevated symptoms, as well as the maximum cut-off scores for depression and anxiety to exclude those with a likely mental illness.
Physical symptoms and biological indicators such as HbA1c were not assessed. However, given that physical symptoms tend to be more responsive to therapeutic writing interventions than psychological symptoms [, ], future studies of therapeutic writing in people with diabetes should include a self-report measure of physical symptoms.
Finally, the current study found that BFW was only moderately acceptable to adults with T1DM or T2DM. Given the reluctance of many people with diabetes to engage in mental health care, future research should examine predictors of engagement with low-intensity interventions, especially those such as BFW, given that it appears to appeal to only some adults with diabetes. Further, since a participatory design is known to increase the acceptability and use of interventions , the development of future therapeutic writing interventions should include the participation of adults with diabetes at every design stage. Although this study set out to use BFW because previous research suggested that it may be suitable for online dissemination due to a lack of distress involved [ , , ], one future direction could be the co-design of a “combined” therapeutic writing, which would allow users to express negative thoughts and feelings about diabetes as well as perceived benefits, with strategies to encourage cognitive change and coping [ , ].
The results of this preliminary RCT found that Web-based BFW for nondepressed adults with T1DM or T2DM was no more efficacious than a CW condition in improving diabetes distress or benefit finding over a period of 3 months. Possibly due to very low baseline levels of depression and anxiety in the sample, BFW in this study was not efficacious in improving the symptoms of depression or anxiety, diabetes self-care, self-reported health, or health care utilization compared to CW.
BFW was adhered to and was associated with increases in positive affect and no more distress than the control condition. This suggests that BFW for diabetes may be more suitable for online dissemination than traditional EW, which typically results in short-term distress. Hence, future research should continue to investigate the efficacy of BFW for adults with T1DM or T2DM, using a larger sample of participants with elevated diabetes distress. Further, engaging a co-design process may improve the perceived helpfulness of therapeutic writing in adults with diabetes. Despite these issues, further research in this population is warranted, as therapeutic writing offers potential as a simple and cost-effective low-intensity intervention, especially for people with diabetes who may not wish to consider mental health interventions.
We would like to acknowledge the support of Faces in the Street: Urban Mental Health Research Institute, St Vincent’s Hospital, Sydney, Australia, and the contribution of Ms Lisa Robins and Ms Therese Fletcher in monitoring the trial. This research was conducted by JC as part of her PhD candidature with the School of Psychiatry, Faculty of Medicine, University of New South Wales. JC’s PhD candidature was supported by an Australian Postgraduate Award scholarship.
JC, KW, and JP conceived of the study and initiated the study design and protocol. JC managed the trial implementation, under the supervision of KW and JP, and drafted this paper. All authors contributed to and approved the final manuscript.
Conflicts of Interest
Multimedia Appendix 1
CONSORT-EHEALTH checklist (V 1.6.1).PDF File (Adobe PDF File), 3MB
Multimedia Appendix 2
Instructions for Web-based benefit-finding writing for diabetes.PDF File (Adobe PDF File), 38KB
Multimedia Appendix 3
Results of secondary outcome measures: observed and estimated means, with SDs at baseline and 1-month and 3-month follow-ups.PDF File (Adobe PDF File), 51KB
Multimedia Appendix 4
Estimated differences in mean change of secondary outcomes between baseline and 1-month and 3-month follow-ups for the benefit-finding writing group and the control writing group.PDF File (Adobe PDF File), 58KB
Multimedia Appendix 5
Within-group estimated changes in secondary outcomes between baseline and 1-month and 3-month follow-ups.PDF File (Adobe PDF File), 71KB
- Cho NH, Shaw JE, Karuranga S, Huang Y, da Rocha Fernandes JD, Ohlrogge AW, et al. IDF Diabetes Atlas: Global estimates of diabetes prevalence for 2017 and projections for 2045. Diabetes Res Clin Pract 2018 Apr;138:271-281. [CrossRef] [Medline]
- Anderson RJ, Freedland KE, Clouse RE, Lustman PJ. The prevalence of comorbid depression in adults with diabetes: a meta-analysis. Diabetes Care 2001 Jun;24(6):1069-1078. [Medline]
- Roy T, Lloyd CE. Epidemiology of depression and diabetes: a systematic review. J Affect Disord 2012 Oct;142 Suppl:S8-21. [CrossRef] [Medline]
- Snoek FJ, Bremmer MA, Hermanns N. Constructs of depression and distress in diabetes: time for an appraisal. Lancet Diabetes Endocrinol 2015 Jun;3(6):450-460. [CrossRef] [Medline]
- Dennick K, Sturt J, Speight J. What is diabetes distress and how can we measure it? A narrative review and conceptual model. J Diabetes Complications 2017 May;31(5):898-911. [CrossRef] [Medline]
- Polonsky WH, Fisher L, Earles J, Dudl RJ, Lees J, Mullan J, et al. Assessing psychosocial distress in diabetes: development of the diabetes distress scale. Diabetes Care 2005 Mar;28(3):626-631. [Medline]
- Fisher L, Glasgow RE, Strycker LA. The relationship between diabetes distress and clinical depression with glycemic control among patients with type 2 diabetes. Diabetes Care 2010 May;33(5):1034-1036 [FREE Full text] [CrossRef] [Medline]
- Fisher L, Mullan JT, Arean P, Glasgow RE, Hessler D, Masharani U. Diabetes distress but not clinical depression or depressive symptoms is associated with glycemic control in both cross-sectional and longitudinal analyses. Diabetes Care 2010 Jan;33(1):23-28 [FREE Full text] [CrossRef] [Medline]
- Gonzalez JS, Shreck E, Psaros C, Safren SA. Distress and type 2 diabetes-treatment adherence: A mediating role for perceived control. Health Psychol 2015 May;34(5):505-513 [FREE Full text] [CrossRef] [Medline]
- Schinckus L, Dangoisse F, Van den Broucke S, Mikolajczak M. When knowing is not enough: Emotional distress and depression reduce the positive effects of health literacy on diabetes self-management. Patient Educ Couns 2018 Dec;101(2):324-330. [CrossRef] [Medline]
- Belvederi Murri M, Mamberto S, Briatore L, Mazzucchelli C, Amore M, Cordera R. The interplay between diabetes, depression and affective temperaments: A structural equation model. J Affect Disord 2017 Dec;219:64-71. [CrossRef] [Medline]
- Schmitt A, Reimer A, Kulzer B, Haak T, Gahr A, Hermanns N. Negative association between depression and diabetes control only when accompanied by diabetes-specific distress. J Behav Med 2015 Jun;38(3):556-564. [CrossRef] [Medline]
- Ehrmann D, Kulzer B, Haak T, Hermanns N. Longitudinal relationship of diabetes-related distress and depressive symptoms: analysing incidence and persistence. Diabet Med 2015 Oct;32(10):1264-1271. [CrossRef] [Medline]
- Anastasiades M, Gupton O, Fritz Y, Calzada P, Stillman M. Diabetes, depression and nonadherencexploring hopelessness as a mediating factor: A preliminary study. Mental Health in Family Medicine 2016 [FREE Full text]
- Gonzalez JS, Safren SA, Delahanty LM, Cagliero E, Wexler DJ, Meigs JB, et al. Symptoms of depression prospectively predict poorer self-care in patients with Type 2 diabetes. Diabet Med 2008 Sep;25(9):1102-1107 [FREE Full text] [CrossRef] [Medline]
- Collins MM, Corcoran P, Perry IJ. Anxiety and depression symptoms in patients with diabetes. Diabet Med 2009 Feb;26(2):153-161. [CrossRef] [Medline]
- Lloyd CE, Dyer PH, Barnett AH. Prevalence of symptoms of depression and anxiety in a diabetes clinic population. Diabet Med 2000 Mar;17(3):198-202. [Medline]
- Egede LE, Nietert PJ, Zheng D. Depression and all-cause and coronary heart disease mortality among adults with and without diabetes. Diabetes Care 2005 Jun;28(6):1339-1345. [Medline]
- Schmitz N, Gariépy G, Smith KJ, Clyde M, Malla A, Boyer R, et al. Recurrent subthreshold depression in type 2 diabetes: an important risk factor for poor health outcomes. Diabetes Care 2014 Apr;37(4):970-978. [CrossRef] [Medline]
- Li C, Ford ES, Zhao G, Ahluwalia IB, Pearson WS, Mokdad AH. Prevalence and correlates of undiagnosed depression among U.S. adults with diabetes: the Behavioral Risk Factor Surveillance System, 2006. Diabetes Res Clin Pract 2009 Feb;83(2):268-279. [CrossRef] [Medline]
- Wilhelm K, Reddy J, Crawford J, Robins L, Campbell L, Proudfoot J. The importance of screening for mild depression in adults with diabetes. Translational Biomedicine 2017;8(1). [CrossRef]
- Schram MT, Baan CA, Pouwer F. Depression and quality of life in patients with diabetes: a systematic review from the European depression in diabetes (EDID) research consortium. Curr Diabetes Rev 2009 May;5(2):112-119 [FREE Full text] [Medline]
- Bot M, Pouwer F, Ormel J, Slaets JPJ, de Jonge P. Predictors of incident major depression in diabetic outpatients with subthreshold depression. Diabet Med 2010 Nov;27(11):1295-1301. [CrossRef] [Medline]
- International Diabetes Federation Guideline Development Group. Global guideline for type 2 diabetes. Diabetes Res Clin Pract 2014 Apr;104(1):1-52. [CrossRef] [Medline]
- Hermanns N, Caputo S, Dzida G, Khunti K, Meneghini LF, Snoek F. Screening, evaluation and management of depression in people with diabetes in primary care. Prim Care Diabetes 2013 Apr;7(1):1-10 [FREE Full text] [CrossRef] [Medline]
- Petrak F, Baumeister H, Skinner TC, Brown A, Holt RIG. Depression and diabetes: treatment and health-care delivery. Lancet Diabetes Endocrinol 2015 Jun;3(6):472-485. [CrossRef] [Medline]
- Diabetes UK. 2010. Emotional and psychological care and support in diabetes: Report from the emotional and psychological support working group of NHS Diabetes and Diabetes UK URL: https://www.diabetes.org.uk/Documents/Reports/Emotional_and_Psychological_Support_and_Care_in_Diabetes_2010.pdf [accessed 2019-06-09] [WebCite Cache]
- Franco P, Gallardo AM, Urtubey X. Web-Based Interventions for Depression in Individuals with Diabetes: Review and Discussion. JMIR Diabetes 2018 Sep 14;3(3):e13 [FREE Full text] [CrossRef] [Medline]
- Pibernik-Okanović M, Hermanns N, Ajduković D, Kos J, Prašek M, Šekerija M, et al. Does treatment of subsyndromal depression improve depression-related and diabetes-related outcomes? A randomised controlled comparison of psychoeducation, physical exercise and enhanced treatment as usual. Trials 2015 Jul 15;16:305 [FREE Full text] [CrossRef] [Medline]
- Fisher L, Hessler D, Glasgow RE, Arean PA, Masharani U, Naranjo D, et al. REDEEM: a pragmatic trial to reduce diabetes distress. Diabetes Care 2013 Sep;36(9):2551-2558 [FREE Full text] [CrossRef] [Medline]
- Shin JK, Poltavskiy E, Kim TN, Hasan A, Bang H. Help-seeking behaviors for serious psychological distress among individuals with diabetes mellitus: The California Health Interview Survey, 2011-2012. Prim Care Diabetes 2017 Dec;11(1):63-70 [FREE Full text] [CrossRef] [Medline]
- Clarke J, Proudfoot J, Vatiliotis V, Verge C, Holmes-Walker DJ, Campbell L, et al. Attitudes towards mental health, mental health research and digital interventions by young adults with type 1 diabetes: A qualitative analysis. Health Expect 2018 Dec;21(3):668-677 [FREE Full text] [CrossRef] [Medline]
- Nyssen OP, Taylor SJC, Wong G, Steed E, Bourke L, Lord J, et al. Does therapeutic writing help people with long-term conditions? Systematic review, realist synthesis and economic considerations. Health Technol Assess 2016 Apr;20(27):vii-xxxvii, 1 [FREE Full text] [CrossRef] [Medline]
- Pennebaker JW, Beall SK. Confronting a traumatic event: toward an understanding of inhibition and disease. J Abnorm Psychol 1986 Aug;95(3):274-281. [Medline]
- Frattaroli J. Experimental disclosure and its moderators: a meta-analysis. Psychol Bull 2006 Nov;132(6):823-865. [CrossRef] [Medline]
- Wilhelm K, Crawford J. Expressive Writing and Stress-Related Disorders. In Harkens K, Hayden EP. editors. The Oxford Handbook of Stress and Mental Health. Oxford: Oxford University Press; 2018.
- Smyth JM, Stone AA, Hurewitz A, Kaell A. Effects of writing about stressful experiences on symptom reduction in patients with asthma or rheumatoid arthritis: a randomized trial. JAMA 1999 Apr 14;281(14):1304-1309. [Medline]
- Baikie KA. Emotional and physical health benefits of expressive writing. Advances in Psychiatric Treatment 2005 Sep 01;11(5):338-346. [CrossRef]
- Bodor NZ. The University of Texas at Austin: Texas Scholar Works. Austin: University of Texas; 2002. The health effects of emotional disclosure for individuals with Type 1 diabetes URL: https://repositories.lib.utexas.edu/handle/2152/11914 [accessed 2019-06-09] [WebCite Cache]
- Ulbrecht J, Trief P, Wallston K, Heron K, Smyth J. Short-term effects of expressive writing as an adjunctive treatment in T2DM on clinical statuspatient well-being. Diabetes 2013 Jul;62(Supplement 1):A2 [FREE Full text] [CrossRef]
- Dennick K, Bridle C, Sturt J. Written emotional disclosure for adults with Type 2 diabetes: a primary care feasibility study. Prim Health Care Res Dev 2015 Apr;16(2):179-187. [CrossRef] [Medline]
- Sloan DM, Marx BP, Epstein EM, Lexington JM. Does altering the writing instructions influence outcome associated with written disclosure? Behav Ther 2007 Jun;38(2):155-168 [FREE Full text] [CrossRef] [Medline]
- Guastella AJ, Dadds MR. Cognitive-behavioural emotion writing tasks: a controlled trial of multiple processes. J Behav Ther Exp Psychiatry 2008 Dec;39(4):558-566. [CrossRef] [Medline]
- Smyth JM, Johnson JA, Auer BJ, Lehman E, Talamo G, Sciamanna CN. Online Positive Affect Journaling in the Improvement of Mental Distress and Well-Being in General Medical Patients With Elevated Anxiety Symptoms: A Preliminary Randomized Controlled Trial. JMIR Ment Health 2018 Dec 10;5(4):e11290 [FREE Full text] [CrossRef] [Medline]
- Antoni MH, Lehman JM, Kilbourn KM, Boyers AE, Culver JL, Alferi SM, et al. Cognitive-behavioral stress management intervention decreases the prevalence of depression and enhances benefit finding among women under treatment for early-stage breast cancer. Health Psychol 2001 Jan;20(1):20-32. [Medline]
- Helgeson VS, Reynolds KA, Tomich PL. A meta-analytic review of benefit finding and growth. J Consult Clin Psychol 2006 Oct;74(5):797-816. [CrossRef] [Medline]
- Tedeschi RG, Calhoun LG. The Posttraumatic Growth Inventory: measuring the positive legacy of trauma. J Trauma Stress 1996 Jul;9(3):455-471. [Medline]
- Jones DR, Johnson JA, Graham-Engeland JE, Park CL, Smyth JM. Is Perceived Growth Associated with Momentary Indicators of Health and Well-Being in People with Asthma or Rheumatoid Arthritis? Appl Psychol Health Well Being 2018 Jul;10(2):254-271. [CrossRef] [Medline]
- Algoe SB, Stanton AL. Is benefit finding good for individuals with chronic disease? In: Park CL, Lechner SC, Stanton AL, Antoni MH, editors. Medical Illness and Positive Life Change: Can Crisis Lead to Personal Transformation?. Washington DC: American Psychological Association; 2009.
- Tran V, Wiebe DJ, Fortenberry KT, Butler JM, Berg CA. Benefit finding, affective reactions to diabetes stress, and diabetes management among early adolescents. Health Psychol 2011 Mar;30(2):212-219 [FREE Full text] [CrossRef] [Medline]
- Nazarian D, Smyth JM. An Experimental Test of Instructional Manipulations in Expressive Writing Interventions: Examining Processes of Change. Journal of Social and Clinical Psychology 2013 Jan;32(1):71-96. [CrossRef]
- Lichtenthal WG, Cruess DG. Effects of directed written disclosure on grief and distress symptoms among bereaved individuals. Death Stud 2010 Jul;34(6):475-499 [FREE Full text] [CrossRef] [Medline]
- Crowley JP. Expressive Writing to Cope with Hate Speech: Assessing Psychobiological Stress Recovery and Forgiveness Promotion for Lesbian, Gay, Bisexual, or Queer Victims of Hate Speech. Hum Commun Res 2013 Dec 06;40(2):238-261. [CrossRef]
- Crowley JP, Denes A, Makos S, Whitt J. The Longitudinal Effects of Expressive Writing on Victims of Relational Transgressions. Health Commun 2018 Feb 01:1-13. [CrossRef] [Medline]
- Lovell B, Moss M, Wetherell MA. Assessing the feasibility and efficacy of written benefit-finding for caregivers of children with autism: a pilot study. Journal of Family Studies 2015 Jun 18;22(1):32-42. [CrossRef]
- Harvey J, Sanders E, Ko L, Manusov V, Yi J. The Impact of Written Emotional Disclosure on Cancer Caregivers' Perceptions of Burden, Stress, and Depression: A Randomized Controlled Trial. Health Commun 2018 Jul;33(7):824-832. [CrossRef] [Medline]
- Stanton AL, Danoff-Burg S, Sworowski LA, Collins CA, Branstetter AD, Rodriguez-Hanley A, et al. Randomized, controlled trial of written emotional expression and benefit finding in breast cancer patients. J Clin Oncol 2002 Oct 15;20(20):4160-4168. [CrossRef] [Medline]
- Danoff-Burg S, Agee JD, Romanoff NR, Kremer JM, Strosberg JM. Benefit finding and expressive writing in adults with lupus or rheumatoid arthritis. Psychology & Health 2006 Oct;21(5):651-665. [CrossRef]
- Guastella AJ, Dadds MR. Cognitive-Behavioral Models of Emotional Writing: A Validation Study. Cogn Ther Res 2006 Oct 10;30(3):397-414. [CrossRef]
- Eysenbach G, CONSORT- E. CONSORT-EHEALTH: improving and standardizing evaluation reports of Web-based and mobile health interventions. J Med Internet Res 2011;13(4):e126 [FREE Full text] [CrossRef] [Medline]
- Crawford J, Wilhelm K, Robins L, Proudfoot J. Writing for Health: Rationale and Protocol for a Randomized Controlled Trial of Internet-Based Benefit-Finding Writing for Adults With Type 1 or Type 2 Diabetes. JMIR Res Protoc 2017 Mar 14;6(3):e42 [FREE Full text] [CrossRef] [Medline]
- Plummer F, Manea L, Trepel D, McMillan D. Screening for anxiety disorders with the GAD-7 and GAD-2: a systematic review and diagnostic metaanalysis. Gen Hosp Psychiatry 2016;39:24-31. [CrossRef] [Medline]
- Manea L, Gilbody S, McMillan D. Optimal cut-off score for diagnosing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. CMAJ 2012 Feb 21;184(3):E191-E196 [FREE Full text] [CrossRef] [Medline]
- Baikie KA, Geerligs L, Wilhelm K. Expressive writing and positive writing for participants with mood disorders: an online randomized controlled trial. J Affect Disord 2012 Feb;136(3):310-319. [CrossRef] [Medline]
- Fisher L, Hessler DM, Polonsky WH, Mullan J. When is diabetes distress clinically meaningful?: establishing cut points for the Diabetes Distress Scale. Diabetes Care 2012 Feb;35(2):259-264 [FREE Full text] [CrossRef] [Medline]
- Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001 Sep;16(9):606-613 [FREE Full text] [Medline]
- Spitzer RL, Kroenke K, Williams JBW, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006 May 22;166(10):1092-1097. [CrossRef] [Medline]
- Kroenke K, Spitzer RL, Williams JBW, Monahan PO, Löwe B. Anxiety disorders in primary care: prevalence, impairment, comorbidity, and detection. Ann Intern Med 2007 Mar 6;146(5):317-325. [Medline]
- Toobert DJ, Hampson SE, Glasgow RE. The summary of diabetes self-care activities measure: results from 7 studies and a revised scale. Diabetes Care 2000 Jul;23(7):943-950 [FREE Full text] [Medline]
- Eigenmann CA, Colagiuri R, Skinner TC, Trevena L. Are current psychometric tools suitable for measuring outcomes of diabetes education? Diabet Med 2009 Apr;26(4):425-436. [CrossRef] [Medline]
- Nielsen ABS, Gannik D, Siersma V, Olivarius NDF. The relationship between HbA1c level, symptoms and self-rated health in type 2 diabetic patients. Scand J Prim Health Care 2011 Sep;29(3):157-164 [FREE Full text] [CrossRef] [Medline]
- Thompson ER. Development and Validation of an Internationally Reliable Short-Form of the Positive and Negative Affect Schedule (PANAS). Journal of Cross-Cultural Psychology 2016 Jul 26;38(2):227-242. [CrossRef]
- Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol 1988 Jun;54(6):1063-1070. [Medline]
- Lepore SJ. Expressive writing moderates the relation between intrusive thoughts and depressive symptoms. J Pers Soc Psychol 1997 Nov;73(5):1030-1037. [Medline]
- Thorndike FP, Saylor DK, Bailey ET, Gonder-Frederick L, Morin CM, Ritterband LM. Development and Perceived Utility and Impact of an Internet Intervention for Insomnia. E J Appl Psychol 2008;4(2):32-42 [FREE Full text] [Medline]
- Nazarian D, Smyth JM. Context Moderates the Effects of an Expressive Writing Intervention: A Randomized Two-Study Replication and Extension. Journal of Social and Clinical Psychology 2010 Oct;29(8):903-929. [CrossRef]
- Graham J, Cumsille P, Elek-Fisk E. Methods for handling missing data. In: Schinka JA, Velicer WF, editors. Handbook of psychology: Research methods in psychology. Hoboken, NJ: John Wiley & Sons Inc; 2013:109-141.
- Liao H, Li Y, Brooks G. Outlier Impact and Accommodation Methods: Multiple Comparisons of Type I Error Rates. J Mod App Stat Meth 2016 May 01;15(1):452-471. [CrossRef]
- Pennebaker JW, Boyd RL, Jordan K, Blackburn K. The development and psychometric properties of LIWC. Austin, TX: University of Texas; 2015.
- Pennebaker JW, Mayne TJ, Francis ME. Linguistic predictors of adaptive bereavement. J Pers Soc Psychol 1997 Apr;72(4):863-871. [Medline]
- Pennebaker JW, Seagal JD. Forming a story: the health benefits of narrative. J Clin Psychol 1999 Oct;55(10):1243-1254. [CrossRef] [Medline]
- Robinson H, Jarrett P, Vedhara K, Broadbent E. The effects of expressive writing before or after punch biopsy on wound healing. Brain Behav Immun 2017 Mar;61:217-227. [CrossRef] [Medline]
- Meinert CL. Clinical Trials Design, Conduct, and Analysis. Oxford, UK: Oxford University Press; 1986.
- Lewandowski GW. Promoting positive emotions following relationship dissolution through writing. The Journal of Positive Psychology 2009 Jan;4(1):21-31. [CrossRef]
- Segal DL, Tucker HC, Coolidge FL. A Comparison of Positive Versus Negative Emotional Expression in a Written Disclosure Study Among Distressed Students. Journal of Aggression, Maltreatment & Trauma 2009 May 21;18(4):367-381. [CrossRef]
- Low CA, Stanton AL, Danoff-Burg S. Expressive disclosure and benefit finding among breast cancer patients: mechanisms for positive health effects. Health Psychol 2006 Mar;25(2):181-189. [CrossRef] [Medline]
- King LA. The Health Benefits of Writing about Life Goals. Pers Soc Psychol Bull 2016 Jul 02;27(7):798-807. [CrossRef]
- King LA, Miner KN. Writing about the Perceived Benefits of Traumatic Events: Implications for Physical Health. Pers Soc Psychol Bull 2016 Jul 02;26(2):220-230. [CrossRef]
- McCullough ME, Root LM, Cohen AD. Writing about the benefits of an interpersonal transgression facilitates forgiveness. J Consult Clin Psychol 2006 Oct;74(5):887-897. [CrossRef] [Medline]
- Sears SR, Stanton AL, Danoff-Burg S. The yellow brick road and the emerald city: benefit finding, positive reappraisal coping and posttraumatic growth in women with early-stage breast cancer. Health Psychol 2003 Sep;22(5):487-497. [CrossRef] [Medline]
- Groarke A, Curtis R, Kerin M. Cognitive-behavioural stress management enhances adjustment in women with breast cancer. Br J Health Psychol 2013 Sep;18(3):623-641. [CrossRef] [Medline]
- Wolever RQ, Dreusicke M, Fikkan J, Hawkins TV, Yeung S, Wakefield J, et al. Integrative health coaching for patients with type 2 diabetes: a randomized clinical trial. Diabetes Educ 2010;36(4):629-639. [CrossRef] [Medline]
- Sturt J, Dennick K, Hessler D, Hunter BM, Oliver J, Fisher L. Effective interventions for reducing diabetes distress: systematic review and meta-analysis. International Diabetes Nursing 2015 Jul 06;12(2):40-55. [CrossRef]
- Barnett AG, van der Pols JC, Dobson AJ. Regression to the mean: what it is and how to deal with it. Int J Epidemiol 2005 Feb;34(1):215-220. [CrossRef] [Medline]
- Ventura AD, Browne JL, Holmes-Truscott E, Hendrieckx C, Pouwer F, Speight J. Diabetes MILES-2 2016 Survey Report. Melbourne, Vic: Diabetes Victoria; 2016.
- Lustman PJ, Griffith LS, Freedland KE, Clouse RE. The course of major depression in diabetes. Gen Hosp Psychiatry 1997 Mar;19(2):138-143. [Medline]
- Lie SS, Karlsen B, Niemiec CP, Graue M, Oftedal B. Written reflection in an eHealth intervention for adults with type 2 diabetes mellitus: a qualitative study. Patient Prefer Adherence 2018;12:311-320 [FREE Full text] [CrossRef] [Medline]
- DeSmet A, Thompson D, Baranowski T, Palmeira A, Verloigne M, De Bourdeaudhuij I. Is Participatory Design Associated with the Effectiveness of Serious Digital Games for Healthy Lifestyle Promotion? A Meta-Analysis. J Med Internet Res 2016 Apr 29;18(4):e94 [FREE Full text] [CrossRef] [Medline]
- Frederiksen Y, O'Toole MS, Mehlsen MY, Hauge B, Elbaek HO, Zachariae R, et al. The effect of expressive writing intervention for infertile couples: a randomized controlled trial. Hum Reprod 2017 Dec;32(2):391-402. [CrossRef] [Medline]
- Gellaitry G, Peters K, Bloomfield D, Horne R. Narrowing the gap: the effects of an expressive writing intervention on perceptions of actual and ideal emotional support in women who have completed treatment for early stage breast cancer. Psychooncology 2010 Jan;19(1):77-84. [CrossRef] [Medline]
|BFW: benefit-finding writing|
|CW: control condition of online writing|
|EW: expressive writing|
|GAD-7: Generalized Anxiety Disorder-7|
|HbA1c: hemoglobin A1c|
|I-PANAS-SF: International Positive and Negative Affect Schedule - Short Form|
|ITT: intention to treat|
|MDD: major depressive disorder|
|PHQ-9: Patient Health Questionnaire-9 items|
|RCT: randomized controlled trial|
|SDSCA: Summary of Diabetes Self-Care Activities|
|T1DM: type 1 diabetes mellitus|
|T2DM: type 2 diabetes mellitus|
Edited by C Richardson; submitted 27.02.19; peer-reviewed by B Auer; accepted 20.03.19; published 27.06.19
©Joanna Crawford, Kay Wilhelm, Judy Proudfoot. Originally published in JMIR Diabetes (http://diabetes.jmir.org), 27.06.2019.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Diabetes, is properly cited. The complete bibliographic information, a link to the original publication on http://diabetes.jmir.org/, as well as this copyright and license information must be included.