Efficacy and cost-effectiveness of a web-based intervention with mobile phone support to treat depressive symptoms in adults with diabetes mellitus type 1 and type 2: design of a randomised controlled trial

Study design

This study is designed as a randomised controlled trial with two parallel groups. The IG will receive access to the diabetes-specific, web-based online program (GET.ON M.D.). The CG will receive general online psychoeducation about depression. Moreover, the CG will have the opportunity to work with an unguided version of the online program 12 months after baseline assessment. The participants in both conditions will have unrestricted access to treatment as usual during the study.

Measurement points are scheduled equally for both groups, at baseline and 8 weeks after randomisation. In addition, there will be follow-up measurements 6 and 12 months after randomisation (Figure 1).

This study was approved by the Psychology Ethics Committee of the Philipps-University of Marburg, Germany.

Recruitment

Firstly, we will advertise the study in German journals for people with diabetes. Secondly, we cooperate with a large-scale German health insurance company: insurants with diabetes will be informed via post about the possibility of participating in this study. Furthermore, we will contact large-scale diabetes patients organisations with the request to inform their members about our study on their websites. In addition, the study will be advertised on social networks such as Facebook and Twitter.

The recruitment message for the study focuses on the opportunity for people with diabetes mellitus (type 1 or type 2) to help reduce depression. Recruitment is still ongoing (start March 2013) with a total recruitment time of 8 months.

Study procedures

The open access website (https://geton-training.de) contains detailed information about the study. Interested visitors may leave their email addresses on this website. Afterwards, they will be registered in a secured web-based platform and invited for an online screening. The people who do not pass the screening will receive an explanation for their exclusion along with the German patient guidelines for depression [45] via email. The people who report suicidal ideation will receive an email with the advice to seek professional help and detailed information on available services. The participants who pass the screening will be asked for informed consent via email and will have to fill out a baseline questionnaire. These participants will also be invited to complete the Structured Clinical Interview (SCID-I for DSM IV, by telephone) [46]. The SCID is used to assess whether the included participants have major or minor depression at baseline and whether they have had episodes of depression in the past. The SCID also includes a suicide protocol that enables us to identify people who are currently suffering from suicidal ideation. The telephone SCID is comparable to a face-to-face interview [47] and administered by a trained interviewer. After the telephone interview, the participants will be randomised.

To increase the probability that the participants will complete the follow-up procedure (8 weeks, 6 months and 12 months after baseline), they are offered a monetary incentive for completing each online survey (50 Euros for the completion of all follow-up surveys). Payment will be arranged after each complete questionnaire has been received.

Study population

This study focuses on adults with diabetes mellitus type 1 or type 2 and depressive symptoms.

Sample size

Informed by previous studies [25, 32], the sample size is based on the expected difference in the primary outcome (CES-D) between the intervention group and the control group at post-test. This sample size provides 80% power to detect a mean difference of d = 0.35 in the effect sizes for the primary outcome (alpha = 0.05, two-tailed test; calculated using Power Analysis and Sample Size Software (PASS 12)). We will need 130 participants per group; hence a total of 260 participants will be included.

Randomisation

An automated, web-based randomisation program (randomisation.eu) will assign the participants to either the intervention or the control group in a ratio of 1:1.

Description of the two groups

Intervention content

Psychoeducation (session 1)

The first session consists of psychoeducation [55]. The participants are informed that diabetes and depression are linked and can impact each other. In two videos, they are provided psychoeducational information, and they receive an overview of the whole program. After defining their goals concerning depression and diabetes, the participants choose a first positive activity for the next day.

Behavioural activation (sessions 2 and 5)

The concept of behavioural activation is introduced with a video. The participants are familiarised with the connection between pleasant activities and mood. With the help of examples, the participants create an individual list of activities. Based on the principles of action planning and coping planning, the participants schedule pleasant activities for a few days [51]. Starting with session 2, this scheduling component is integrated into each session.

In session 5, the key focus is on physical activity. Based on the transtheoretical model of health behaviour change [56], the participants can reflect on their own stage of change and receive adapted information and tasks.

Problem solving (sessions 3 and 4)

Throughout sessions 3 and 4, the main focus is on problem solving [52]. After listing experienced problems, the participants are prompted to classify their problems as solvable or unsolvable. Following a video, the participants become familiar with a six-step plan to handle solvable problems. Next, they are encouraged to apply this plan to one of their problems. In session 4, the participants retrospectively review their plans and choose to either continue to work on the specific problem or start to work on a new one. In addition, they learn techniques to reduce rumination on unsolvable problems.

Plan for the future (session 6)

In the last session, the participants assess whether they have seen improvements in mood. They can learn about additional ways to get help, if needed. In addition, the participants reflect on which techniques suit them, and are encouraged to make a plan for the future. At the end, they write a letter to themselves about what they want to achieve during the next four weeks.

Optional sessions and booster session

The participants are offered two optional sessions (healthy sleep and coping with obesity) that are available at the end of sessions 2 through 5. It is known that insomnia is a predictor and a symptom of depression [57]. The sleep session offers information on the principles of sleep hygiene, sleep restriction and stimulus control [58]. There is also a relationship between obesity and depression [59]. In particular, people with diabetes type 2 often suffer from obesity [60]. The aim of this session is to cope with the psychological consequences of obesity, for example frustration, feelings of self-blame and guilt.

Four weeks after finishing the intervention, the participants have the opportunity to work on an optional booster session. The aim of this session is to evaluate progress and to strengthen their newly acquired skills.

Improving efficacy and adherence

Taking into account that non-compliance presents as a problem in web-based programs [38], in many depressed people [36, 37] and people with diabetes [35], we developed two ways to increase the number of sessions completed by the intervention group.

Improving efficacy and adherence I: text message coach

Preliminary results have shown that mobile phone-based interventions and mobile phone supportive messages have the potential to improve psychological wellbeing and decrease stress or depression [61]. A pilot study demonstrated that daily text messaging, as a supplement to face-to-face CBT, is feasible and acceptable to patients. Positive experiences with the text messages were reported by 65% of the participants [62]. Studies have also indicated that text messages about diabetes self-management can also lead to benefits in some cases [63, 64]. The participants in this study will have the opportunity to apply for a standardised text message coach: they will receive daily text messages to support their efforts to integrate the techniques they learn within the program into everyday life. The text messages address four different areas: behavioural activation, problem solving, diabetes and motivation enhancement.

Improving efficacy and adherence II: phone call

Clarke et al. (2005) have demonstrated that reminders (by telephone or postcards) are a pivotal factor in effectiveness and attrition [65]. In the GET.ON M.D. study, a participant who does not complete a session within one week will receive a reminder email from his or her personal trainer. If a participant does not log in to the program during the next seven days after receiving this reminder, he or she will receive a phone call from his or her trainer.

The aim of this call is both to motivate the participant to proceed with the training and to identify the exact reasons for not logging in (e.g. forgetfulness, technical problems, lack of interest).

Control group - psychoeducation

The CG will have access to online psychoeducation based on the German S3-Guideline/National Disease Management Guideline “Unipolar Depression” [45]. Psychoeducation is helpful in reducing depressive symptoms [55]. It informs the participants about the nature of depression including information about symptoms and sources of help. After the last measurement point (12 months after baseline), the CG will have access to an unguided version of the GET.ON M.D. program.

Outcome assessment

Primary outcome measures

The primary outcome is the level of depression, which is measured by the CES-D, a well-validated self-report depression questionnaire that will be administered online [48, 49]. The CES-D consists of 20 items that are answered on a 4-point Likert scale; the total score ranges between 0 and 60. The internal consistency is reported to be α = 0.89, and the split-half reliability is r = 0.91 [48].

Secondary outcome measures

The secondary outcome measures include the depression subscale of the Hospital Anxiety and Depression Scale (HADS [66, 67]); the Penn State Worry Questionnaire (PSWQ [68], 3-item short version [69]); the Behavioural Activation Depression Scale (BADS [70], 9-item short version [71]); the Social Problem Solving Inventory-Revised (SPSI-R [72]); the long-term blood sugar level (self-reported haemoglobin A1c values (HbA1c)); the Problem Areas in Diabetes questionnaire (PAID [73], German 5-item short version [74]); the Acceptance and Action Diabetes Questionnaire (AADQ [75], German version [76]); the Diabetes Self-Management Questionnaire (DSMQ [77]); the negative effects of online health trainings questionnaire ([78], self-adapted version for online health trainings); the Kessler Psychological Distress Scale (K10 [79]); and the Attitudes Towards Seeking Psychological Help Scale (ATSPHS, [80], 10-item short version [81]).

Additional measurements

We will collect socio-demographic data (age, gender, education and occupation); lifestyle data (body mass index, smoking and alcohol use); data on diabetes (type, duration of illness and comorbidity); Patient Questionnaire on Therapy Expectation and Evaluation (PATHEV [82]); data on satisfaction with the program (a self-designed questionnaire based on the German version of the Client Satisfaction Questionnaire [83, 84]); data about the method of recruitment (for example through a diabetes journal or Facebook) and additional information about the participant’s familiarity with the Internet.

Measurements for the economic evaluation

Quality of life

Quality-adjusted life years (QALYs) will be calculated based on the EuroQol (EQ-5D) [87] and the Short Form Health Survey 12 (SF-12) [88]. The EQ-5D is a validated tool for measuring general health-related quality of life. It consists of five items (mobility, self-care, usual activities, pain/discomfort and anxiety/depression), each of which is rated as causing “no problems”, “some problems” or “extreme problems” [89]. Moreover, to generate the QALYs, a visual analogue scale (VAS) ranging from 0 (worst imaginable health state) to 100 (best imaginable health state) is also an integral part of the EQ-5D.

The SF-12 is based on the 36-item Short Form Health Survey (SF-36), a health-related quality-of-life questionnaire [90]. The SF-12 focuses on eight dimensions of health (physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional and mental health).

Statistical analysis

Economic evaluation

The economic analysis will be performed from a societal perspective. We will conduct a cost-effectiveness analysis (CEA) and a cost-utility analysis (CUA). Using the incremental cost-effectiveness ratio (ICER), we will show the differences in the costs and benefits between the two groups. The robustness of the ICER will be determined using non-parametric bootstrapping. Furthermore, we will also use bootstrapping to quantify the uncertainty around the ICER, which will be shown on the cost-effectiveness plane [97] and as a cost-effectiveness acceptability curve [98, 99]. To perform the CUA, we calculate QALYs from the EQ-5Q and also from the SF-12.

Limitations

One limitation of this study design is that many diabetes patients are elderly, especially those with diabetes type 2 and may have limited access to the Internet. In 2002, only 8% of the Germans 60 or older had access to the Internet: this number increased to 39% ten years later [102]. The results from this study cannot be generalised to the overall population of patients with diabetes (especially type 2) and comorbid depression.

Furthermore, it is known that lower social status and lower education are associated with an increased risk for diabetes [103] and that women (28%) have significantly higher rates of depression compared to men (18%) [2]. As studies show, web-based interventions are mostly used by higher educated persons and females [32, 104, 105]. Therefore, we expect selection bias, restricting the external validity, particularly with regard to lower educated and males with diabetes.