EFFORT-D study process evaluation: challenges in conducting a trial into the effects of running therapy in patients with major depressive disorder

Major depressive disorder (MDD) has huge effects on wellbeing and daily personal and professional functioning. It is the second leading cause of disease burden based on years lived with disability [1], and the loss of daily-adjusted life years is considerable [2]. Depressive symptoms are common in various populations [3]. Recurrence of symptoms of MDD occurs in an unfavourable but fluctuating course in 44% of the patients and a severe chronic course in 12–32% of the patients [4, 5]. Pharmacotherapy and psychotherapy, i.e. cognitive behaviour therapy and interpersonal therapy, are the mainstays of treatment [6, 7]. Effective use of antidepressant medication is often compromised by side effects, while both the availability of trained therapists and lack of motivation on the part of patients may limit the utility of antidepressant psychotherapy [8]. Although for instance cognitive therapy treatment has a positive effect, about 50% of the patients relapse after 2 years with an estimated 80% relapse after 5 years [9]. Meta-analysis of standard MDD treatment showed a clinically and statistically heterogeneous outcome of treatment as usual: about 30% of the patients remitted from depression, but depressive symptoms also worsened in 12% [10].

Exercise is currently seen as a promising alternative to pharmacotherapy and psychotherapy, because it has low costs and few side effects, and is also considered safe for most patients. Previous research showed that a low level of physical activity is highly prevalent in individuals with depression [11] and that exercise has a positive effect on physical health in such a population [12‐14]. Physical activity also has additional advantages in the large group of patients that suffer from a combination of mental and physical problems, such as depression and diabetes, in which prevention of metabolic syndrome by enhancing physical activity is vital [15, 16].

Using exercise therapy as (part of the) regular treatment for MDD seems, therefore, a logical consequence. However, studies focussing on exercise therapy as (part of the) treatment for depression usually involved patients with mild to moderate depression. These patients were mostly recruited through public announcements, and attracted participants who are different from the population with MDD that is hospitalised in recognised mental health care institutions. Of the 28 studies selected by Mead et al. [17] in their Cochrane review on exercise and depression, only two were randomized controlled trials (RCTs) that included adult inpatients [18, 19]. In a follow-up Cochrane review of Cooney et al. [20], only seven of 39 studies recruited participants from clinical populations, i.e. hospital inpatients or outpatients referred by a general practitioner (GP). The studies included in the latter review, that reported positive effects on quality of life and symptom severity in severely depressed inpatients, were all short-term studies.

One of the few studies into severe MDD inpatients, which was not included in the aforementioned Cochrane reviews, had an add-on design with, besides exercise, treatment as usual consisting of pharmacotherapy and/or electroconvulsive therapy [20]. Although this study lasted only 2 weeks, it showed a significant effect on depressive symptoms as well as physical and psychological domains of quality of life; however, 50% of the eligible patients were not motivated to participate. The eight trials with a longer follow-up in the Cochrane review of Cooney et al. [21] using an intervention time between 10 and 16 weeks and follow-up measurements with a maximum of 10 months, showed only a moderate effect on depression with a pooled standardised mean difference (SMD) of − 0.33. This is in line with the review of Krogh et al. [22] studying RCTs in adults with a clinical depression. This review showed a significant SMD of − 0.40 in 13 trials, with an inverse relationship between duration of the intervention and the magnitude of the association between exercise and depression. Also, a pooled analysis of five trials with a long-term follow-up (9–26 months), suggested no long-term benefit (SMD − 0,01). Consequently, when implementing exercise in for instance a guideline for treatment of MDD, a diminishing effect of exercise in the long run should be considered.

Limited evidence is found that short-term interventions are as effective as long-term interventions in MDD [23]. The latest review in this field advised moderate exercise to be suitable as part of a multidimensional approach [24]. A limitation of existing studies is that bias must be considered regarding inadequate blinding of participants and observers. When studies with adequate allocation concealment, intention-to-treat-analysis and blinded outcome assessment were analysed, the pooled SMD (− 0.18) for the primary outcome of depression scores was small and no longer significant [17, 21, 25].

An important aspect of MDD is the unpredictable duration of a depressive period. In an epidemiological cohort study, half of those with MDD and a minor depressive episode recovered within 6 months with standard treatment, wherein only a short period without treatment seemed, however, defensible. Treatment was urgently advised for those with greatest need, a lower functioning at onset and experiencing comorbidity [5]. However, it is unknown how long an intervention should take. The abovementioned controversy between moderate positive results from meta-analyses and treatment of severe (inpatient) MDD in the harsh clinical daily practice still exists, leaving clinicians and patients uncertain about the optimal use of exercise as a part of their treatment. As a result, more high-quality research into the relationship between MDD and exercise as a (add-on) treatment was needed.

The EFFect Of Running Therapy on Depression (EFFORT-D) study, which has been described in a previous publication, was both aimed at severe (inpatient) MDD and a long-term exercise intervention in an add-on RCT design [26]. During the EFFORT-D study, which was carried out between 2012 and 2015, both an effect evaluation (results to be published) and a process evaluation were conducted. The process evaluation enabled us to determine influencing factors to the outcome results in different phases and identify topics for future research, such as implementation strategies in clinical practice and advices for depressive patients. To our knowledge, only few process evaluation studies in this field have been published [27‐29]. The importance of a process evaluation has been demonstrated by Linnan and Steckler [30]. Although their examples focused on preventive population interventions, their framework can also be applied to our study focusing on curative interventions for patients. Therefore, the aim of this process evaluation was to gain insight into the context, reach and dose of the exercise intervention of the EFFORT-D study.

This process evaluation showed that economic conditions had an important negative influence on the implementation of the EFFORT-D study. A severely reduced reach, combined with a large loss of dose delivered was shown. Once included in the intervention group, it turned out to be very difficult to motivate patients with MDD to participate in the aerobic exercise therapy and the long-term measurements. The duration of the intervention seemed to be too long according to the evaluative questionnaire. During this study, the research team faced substantial challenges in several areas. An unexpected high proportion (almost three quarters) of the eligible participants was excluded before randomization. As a result, only about one-fifth of the intended number of patients could be included in this study. Also, contextual difficulties, due to an organisational merger and an economic-driven reorganisation, reduced the reach and the number of available patients in the research population. Forty percent of the eligible participants reported that they were unable to incorporate the exercise programme in their daily lives. They argued that they were too tired and the required investment in time was too high. This can be interpreted as a lack of motivation and it seems to form a structural problem for implementation of exercise in MDD. This motivational issue was also found to be a perceived barrier to exercise in another qualitative study [27]. Patients in that study were screened for preferences of exercise as a treatment for depression. They reported mood problems, tiredness and lack of motivation as the biggest barriers for exercise in both men and women. In another systematic review on motivating factors and barriers towards exercise, mood problems and stress were also identified as major barriers for engaging in exercise during severe mental illness [31]. Comparable trials in MDD inpatients or severe depressed outpatients who were referred by GPs or psychiatrists also show 25–50% of the eligible participants that refuse to participate. This may be an indication that severity of depression might be the underlying factor for refusal [20, 32]. However, severity of depression seemed not to be an important factor of non-adherence to exercise in a pooled analysis from two clinical trials [33]. This result could be explained by the fact that a population with mild to moderate depression was included in these trials. Finally, in 20% of the outpatients referred by a GP in our study, depressive symptoms fell below the required threshold for inclusion during the intake procedure. This was possibly due to the mechanism of spontaneous recovery [34] or was a result of treatment that was already started by the GP. The fact that the dropout rate in both arms of our study was comparable could indicate that the intervention itself was not the main cause of the high dropout. It might also mean that strain of the follow-up testing was too high for these severely depressed participants. This was, however, not confirmed by the evaluative questions answered by the exercise leaders and research assistant. The follow-up arguments for quitting the study showed that participating in the early stages of the intervention did not provide enough incentive to remain on board. These kinds of motivational problems are also described in patients with affective disorders such as MDD, bipolar disorder [35] and with schizophrenia [36].

The concept of ‘autonomous motivation’ to exercise, meaning an intrinsic motivation and an identification with the benefits of a healthy lifestyle, seems to play a vital role in maintaining involvement in the exercise programme. Somewhat hopeful here is the fact that 14% of our eligible participants were excluded because they already trained twice a week for 30 min. Possibly, the use of a depression guideline of the GPs explains this effect, because it recommends prescribing exercise for mild to moderate depression. It seems that the EFFORT-D intervention and study design in its current form does not seem to be feasible in these MDD patients, certainly not for hospitalised inpatients. So, regarding implementation of exercise programmes in the treatment of MDD, motivational themes need a lot of attention throughout the programme. This is very important, because people with severe psychiatric disorders, such as MDD, are at higher risk for chronic physical conditions [37‐39] and tend to be very sedentary [40]. It might seem more adequate to treat MDD patients using the activation principles of patients with other severe mental illnesses such as schizophrenia or bipolar disorder. As described in recent publications, aiming the exercise intervention on improvement of quality of life and cardiorespiratory fitness, besides the improvement of depression, is indicated [11, 38, 41]. A broader multidisciplinary lifestyle intervention, incorporating light-to-moderate physical activity in daily life, in addition to lifestyle interventions aimed at smoking and eating habits, seems effective. A disadvantage of such a multidimensional approach is that it will be difficult to determine the effect of the physical exercise separately, but an advantage is a probably broader health gain. An alternative approach might be indoor training, as used in other studies of depressed patients [32, 42] and in studies of patients with schizophrenia [43]. This could also prevent training absences during wintry conditions with icy running tracks. However, a combination of indoor and outdoor exercise is preferred due to indications that a green environment contributes to wellbeing [44, 45]. Another option is to launch a pre-study training, in which participants are given the opportunity to personalise their programme with help from a physical therapist with the aim of increasing their fitness level and preventing injuries. This is in line with recommendations based on a qualitative study of patients’ viewpoints in general practice involving 33 participants [28]. Once participants have a sufficient fitness level, they can participate in the group intervention. An advantage of this personalised approach is that the form of exercise can be adapted to the patients’ preferences and possibilities, for instance cycling in the case of joint problems. In addition, also E-learning environments, supporting technical tools such as electronic devices for self-registration of activity and smartphone applications (including WhatsApp and social media) can be considered. These can potentially encourage participants in daily life and promote more adherence to exercise [46].

Prevention of early dropouts needs more attention in future research. It is a paradox, however, that the estimated length of an optimal exercise study is 3–4 months, whereas the preferred length of an exercise programme is longer. A single study, with a long follow-up of 10 months, showed that it is worthwhile to invest in long-term motivation to persist in exercising because of maintenance of the therapeutic effect and prevention of a relapse [47]. The latest results of the Dutch NEMISIS-study show that the median time for a depressive episode was 6 months in MDD and 3 months for a minor depressive episode with a full remission of 75% after 1 year for both [5]. This means that exercise as an adjunctive treatment probably can contribute to remission in the first 6 months, after this period it is probably meaningful to prevent a relapse.

This process evaluation had several limitations. First, evaluative questioning of participants was not appropriate, because the expected bias in favour of exercise was estimated to be high: practically only the very small group of completers of the intervention at T12 were able to give feedback. This diminished potential informative and qualitative suggestions for future research from a patient perspective. In particular, evaluation of the larger group of dropouts could have given more detailed information about blockades for exercise or suggestions to solve these, but that was not feasible. Second, the questions directed at the trainers and research assistant for this process evaluation were sent by e-mail. Instead, a (semi)-structured interview would have led to more in-depth information. At last the numbers of included patients were small so the conclusion that this type of exercise in these particular patients does not fit MDD patients cannot be generalised easily. Nevertheless, this process evaluation of an exercise intervention in patients with MDD is one of the few delivered and may be of value for both researchers as well as clinicians in the field. Researchers can profit from the described pitfalls, whereas the clinicians may recognise their own patients and prevent obstacles when planning an (add-on) exercise intervention.

Our process evaluation showed that both the intervention as well as the study design of EFFORT-D were not feasible for patients with MDD in terms of motivation and compliance. In future interventions, applying electronic self-registration devices is advised, which not only can motivate the participant through direct feedback, but they also enable researchers to objectively measure the intensity and time spent on physical activity. Besides, social media use could enhance more adequate communication (for instance forming an easy accessible WhatsApp or Facebook group). Further research is needed to define the optimal duration of an exercise intervention in MDD. Finally, a tailored and stepped-care intervention method is advised in case of aerobic exercise training such as running therapy or Nordic walking, starting with a pre-study phase in which a physical therapist can prepare the potential participant. Flexibility in aerobic activity fitting the individual participant, and combining in- and outdoor facilities, may also help to retain the participants and enhance motivation for exercise, making it part of their lifestyle.