Background
Major depressive disorder (MDD) is related to high disease burden for both people affected and society [
1]. In a recent literature review investigating global variation in the prevalence and incidence of MDD, a global point prevalence of 4.7%, a lifetime prevalence between 10 and 15% and a global incidence of 3.0% have been reported [
2,
3]. It is estimated that existing psychological and pharmacological treatments have the potential to avert only 36% of the burden of MDD, and only when assuming perfectly efficient provision of existing treatments in terms of coverage, patient compliance, an d clinician competence [
4,
5]. Thus, we are either in need of more powerful interventions for treating depression or we should aim at diminishing the likelihood of developing depression in the first place, highlighting prevention of depression as a promising approach. Recent research suggests that psychological preventive interventions such as cognitive behavioral therapy (CBT) or interpersonal psychotherapy have the potential to prevent a clinically significant number of new depression cases [
6]. A meta-analysis of 32 randomized controlled trials (RCTs) reported a reduced incidence rate for MDD of 21% (incidence rate ratio = 0.79, 95% confidence interval: 0.69–0.91) when comparing psychotherapy-based preventive interventions with usual care or wait list conditions.
While the effectiveness of preventive interventions seems sufficiently documented, it remains challenging to identify target populations that benefit most from preventive measures [
7]. According to Cuijpers and colleagues [
8] two factors need be taken into account: the “impact” and the “effort” of preventive measures. An adequate “impact” means that prevention must lead to a substantial reduction of total disease burden. Therefore, a substantial proportion of new cases must be prevented if assembled risk indicators are fully blocked. A reasonable level of “effort” is primarily defined as a low number needed to be treated (NNT) to prevent one new case of MDD. Additionally, persons at risk should be easily identifiable and interventions should not only be cost-effective but also low priced to allow for their implementation at a population level.
From this viewpoint, chronically medically ill patients appear to be a meaningful target population for the prevention of MDD, given the substantially increased prevalence for MDD in this population compared to the general population [
9,
10]. In addition, comorbid MDD in medically ill patients is associated with numerous negative implications such as problems in the physician-patient relationship, increased risky health-related behaviors, higher medical symptom burden, medical complications, lowered quality of life and increased mortality [
11,
12]. Within the group of medically ill persons, back pain is one of the most common conditions [
10,
13] and is associated with a two to three-fold increased risk for MDD [
14]. In addition, depression is one of the core predictors of persistent pain symptoms, increased pain related disability, and poor treatment outcomes, and is associated with increased morbidity and health care costs as well as diminished quality of life [
11,
12,
15‐
17].
The benefits of prevention can be multiplied by focusing on patients who already show some depressive symptoms due to several reasons. First, subthreshold depressive symptoms are an additional risk indicator for MDD [
18,
19]. Multiple risk groups have increased specificity for prevention measures which leads to a reduction of NNT (“effort”) [
20] and leading to greater cost-effectiveness of preventive interventions. Second, by lowering the NNT, the number of persons who are not in need of a preventive intervention, but receive it, will be reduced. Third, subthreshold depression itself is a considerable disease burden for people affected and for society [
21,
22]. A successful preventive intervention will not only reduce the risk of developing MDD but also improve depression symptom severity at all levels of depression, as shown by a recent meta-analysis (pooled effect size g = 0.35, 95%-CI: 0.23–0.47; [
23]). Fourth, uptake rates of a depression prevention intervention may be higher in a target population of patients with depressive symptoms, as treatment utilization was found to be associated with severity of baseline depression [
24].
The internet is an appropriate prevention medium for scaling up preventive interventions as units of delivery are reasonably priced and can be easily administered [
7,
25]. It has several additional advantages as discussed elsewhere [
26‐
28].
In prior studies, Internet- and mobile-based interventions (IMI) have shown to be effective in the treatment of MDD [
29,
30] as well as in the treatment of subthreshold depression, indicating their potential to be utilized for preventive interventions [
25,
31‐
34]. Human support (guidance) has repeatedly been shown to have a positive effect on effectiveness of and adherence to IMIs [
35,
36].
The proposed study aims to investigate the effectiveness and cost-effectiveness of an IMI (eSano BackCare-DP) to prevent the onset of depression for chronic back pain patients (PROD-BP) with subthreshold depressive symptoms. The study will be embedded into routine orthopedic care in order to examine the intervention’s effectiveness in an unselected sample of all eligible chronic back pain patients (i.e. the implementation will not be limited to a self-selected population of people who are already attracted to depression prevention interventions and the Internet). It is expected that
1)
eSano BackCare-DP is effective in preventing the onset of MDD compared to treatment as usual (TAU) over a 12-month follow-up period,
2)
eSano BackCare-DP is cost-effective compared to TAU.
3)
Compared to TAU, eSano BackCare-DP is superior in terms of (a) depression response, (b) work capacity, (c) quality of life, (d) pain related disability and (e) pain intensity.
Furthermore, the distributions of principal confounders in each group will be explored.
Assessments
Assessments will be conducted at pre-treatment (T0) and at 9 weeks, 6- and 12-months follow-up (T1, T2, T3; see Table
1). All self-report assessments, potentially effect-modifying covariates and demographic variables will be provided using a web-based interface integrated into the intervention platform. Section A (Affective Syndromes) of the SCID [
43], the Hamilton Rating Scale for Depression (HAM-D-17 [
56]) and the Quick Inventory of Depressive Symptomatology (QIDS-C16; [
57]) will be performed via telephone interviews at T0 and T3. Moreover, we translated the SCID-V-RV into German to be able to comment on DSM-IV and -V [
58] diagnoses.
Table 1
Outcome assessments and assessment time point
Inclusion/exclusion criteria |
Chronic Back pain | MR + TI | x | x | | | |
Depressive symptomatology | PHQ-9 | x | x | x | x | x |
Inclusion criteria a), c), d) | TI/SRQ | | x | | | |
Acute/past 6 months depressive episode, dysthymia or bipolar disorder | SCID | | x | | | x |
Current/past 6 months/on waiting list for psychotherapy | TI | | x | | | |
Suicidality | SCID/HAM-D/QIDS | | x | | | x |
PHQ-9 | x | x | x | x | x |
Primary outcome |
Onset of Depression | SCID | | x | | | x |
Secondary outcomes |
Severity of depressive symptoms | PHQ-9 | x | x | x | x | x |
HAM-D/QIDS | x | | | x |
Quality of life | AQOL-6D/EQ-5D-5 L | | x | x | x | x |
Pain intensity | Rating scale | | x | x | x | x |
Pain related disability | ODI | | x | x | x | x |
Pain self-efficacy | PSEQ | | x | x | x | x |
Ability to work | SPE | | x | x | x | x |
Economic evaluation |
Costs | TiC-P | | x | | x | x |
Quality of life | AQOL-6D/EQ-5D-5 L | | x | x | x | x |
Covariates |
Demographic variables | SRQ/MR | x | x | | | |
Depression type and chronicity | SCID | | x | | | x |
Patient adherence | Attrition rate | | | x | x | |
Patient satisfaction | CSQ-8a
| | | x | | |
Side-effects of intervention | INEPa
| | | x | x | |
Back Pain type and chronicity | MR | x | | | | |
Internet affinity | IAS | | x | | | |
Videos promoting the importance of collecting data will be implemented into online assessments to enhance compliance with completing measures. In psychological intervention trials, blinding of study participants and eCoaches is not possible. However, all members of the research team conducting telephone administered outcomes will remain blinded. Therefore, both participants and interviewers will be reminded of the reason and importance of blinding at the beginning of each interview. Moreover, performance bias will be minimized, as the web-based intervention is separated from other health care services.
Procedure on suicidal ideation
The telephone interviews (SCID, HAM-D, QIDS) and questionnaires (PHQ-9) include a suicide screening to identify participants who are currently suffering from suicidal ideation. We will follow a suicide protocol adapted from prior trials [
31,
44] if participants score on any suicidality item. Participants who report low suicidal ideation (HAM-D, QIDS or PHQ-9 item score = 1) will receive an email with detailed information on available health services and the advice to seek professional help if symptoms increase. If participants express moderate to high suicidal ideation during the assessment or express any suicidal thoughts or intentions to their eCoach, a trained psychotherapist from the study team [EM, LS, HB, SaS] will contact the participant and initiate further actions.
Discussion
This study will be the first to investigate the effectiveness and cost-effectiveness of a psychological Internet- and mobile-based intervention for the prevention of depression in a chronic pain population. Due to its recruitment strategy from routine medical health care, the entire potential target group can be reached within a naturalistic setting. Results will have implications for researchers, health care providers and public health policy makers.
Conducting IMI-trials commonly involves possible limitations, which we try to overcome using the following measures. First, web-based interventions can have moderate to high drop-out rates [
90‐
93], and drop-out rates can be expected to be even higher in preventive interventions due to the lower symptom burden of participants. We will approach this problem in different ways: a) by focusing on patients with current depressive symptoms b) by providing guidance via eCoaches, which has been shown to have an adherence-facilitating effect [
35,
36], and c) by explicitly facilitating at risk participants’ motivation to use the intervention after discharge from orthopedic rehabilitation care. In a prior IMI with diabetes patients, participants completed an average of 78.3% of all sessions [
44]. In a sample of subthreshold depressed patients, participants completed an average of 82.2% [
31]. These results correspond to findings from a recent meta-analysis on adherence to internet-based CBT (ICBT) [
94]. Van Ballegooijen and colleagues concluded that adherence to guided ICBT could be equal to adherence to face-to-face CBT. Participants do not necessarily have to complete all sessions to benefit from IMIs. They may also stop the treatment because they have recovered [
95] or experienced improvement in symptoms, thereby reducing the likelihood of developing depression. These cases would represent a prevention success rather than a treatment drop-out [
90].
A second limitation of most Internet- and mobile-based trials to date (including those mentioned above) are their highly selective online recruitment strategies. This may explain the promising results concerning drop-out rates, as participants already connected to the Internet comprised the intervention groups. As a down-side, however, those recruitment strategies lead to a lack of external validity [
27,
96,
97]. In our study, we address this problem through the integration of the intervention into routine care. Thereby, the entire potential target group will be offered the opportunity to take part in the preventive intervention. The two different recruitment strategies will allow for analyses on different dissemination and implementation strategies of IMIs into routine healthcare. Thus, we can estimate what kind of patients, and to which extent, make use of the offer to take part in a preventive IMI within the whole group of chronic back pain patients.
Third, IMIs can have negative side effects [
98‐
100]. For this reason, we followed the key recommendations of Rozental and colleagues [
98]. We increased the flexibility of the treatment schedule by giving participants the possibility of delay at the beginning of each session, and increased flexibility of therapist contact for patients. Additionally, we prolonged treatment duration by adding two booster sessions after the main treatment modules. Furthermore, negative side effects of treatment will be assessed on a regular basis and reasons for drop-out from intervention will be assessed.
The specific strengths of this study are the following: a) Prevention studies are regularly methodologically limited because they lack a diagnosis at baseline and/or follow-up [
6]. By carrying out the SCID prior to study start and at 12-months follow-up a high content validity can be ensured. b) With a target sample of 406 participants, the study will be optimally powered, overcoming the small scale trial limitations of most prior prevention studies [
6,
20]. Following the ITT principle contributes to reducing overestimation of clinical effectiveness. c) The intervention is specifically tailored to the special needs of the target group of chronic back pain patients. This has been discussed as having an uptake and adherence facilitating effect [
27]. We aim to further facilitate adherence through the integration of the intervention into patients’ routine healthcare, which enables clinicians to inform participants about the characteristics and effectiveness of IMIs. This may have a positive impact on their acceptance [
101]. d) Using the internet as the medium for prevention might allow for scaling up of preventive interventions on a public mental health level. e) The direct implementation of the intervention into the health-care system increases external validity in contrast to prior RCTs [
31,
102].
High prevalence rates underscore that the integration of depression prevention into curative care systems for the medically ill is one of the major emerging global health challenges. If this study - the first of its kind – shows to be effective, the intervention could be implemented into general (chronic) back pain and mental health treatment protocols as well as adapted to other chronically ill patient groups, thus helping to reduce the disease burden of depression for both affected persons and society. Thus, the results of this study will be of major public health relevance.
Acknowledgements
We thank Prof. Dr. Dr. Jürgen Bengel, Dr. Kristin Kieselbach and Prof. Dr. Anja Göritz for their support by providing their expertise in the treatment of chronic pain patients. We thank colleagues who were part of the development of prior interventions [
44,
46] that partly build the basis of eSano BackCare-DP. Moreover we like to thank our many study assistants for their support in the development of the intervention. We thank Ellen Meierotto for supervising the SCID-interviewers. We thank Yannik Terhorst and Susanne Stollewerk for proofreading and Mary Wyman for language editing of the manuscript. Many thanks go to the Data Safety and Monitoring Board, consisting of Prof. Dr. Martin Hautzinger, Prof. Dr. Martin Härter and Dr. Levente Kriston as well as the Clinical Trials Unit Freiburg for their support in carrying out the study project.
Special thanks go to the cooperating orthopedic rehabilitation facilities Schoen Klinik, Bad Staffelstein; Rehaklinik Sonnhalde, Donaueschingen; RehaKlinikum Bad Säckingen; Städt. Rehakliniken Bad Waldsee; Schwarzwaldklinik Orthopädie - Abteilung Medizinische Rehabilitation, Bad Krozingen; Rheintalklinik Bad Krozingen; REGIO-Reha Tagesklinik, Freiburg and Universitäts- und Rehabilitationskliniken, Ulm as well as the further orthopedic rehabilitation units of the second recruitment strategy.