Enthusiasm for homework and improvement of psychological distress in subthreshold depression during behavior therapy: secondary analysis of data from a randomized controlled trial

Subthreshold depression, sometimes also called subsyndromal depression or minor depression, refers to a depressive state that does not meet the full diagnostic criteria for major depressive disorder (MDD) [1]. It would be classified as Depressive Disorder Not Otherwise Specified in DSM-IV [2] or Other Specified Depressive Disorder in DSM-5 [3]. Patients with subthreshold depression often present with both depression and anxiety symptoms, albeit both under any diagnostic thresholds [4]. It has been increasingly recognized that subthreshold depression is highly prevalent in the general population [5] and in the primary care [6], is clinically relevant because it significantly affects the quality of life and functioning of the sufferers [7] and carries a high risk of developing major depressive disorder [8], and is societally important because it is associated with high economic costs [9].

Cognitive behavioral therapy (CBT), the psychotherapy for depression and anxiety with the largest number of randomized controlled trials (RCTs) supporting its efficacy [10, 11], has been shown to be effective for subthreshold depression [12] as well. In CBT, patients self-monitor their own behaviors, emotions and thoughts and also practice newly learned cognitive or behavioral coping skills outside the therapy sessions as homework. Homework is usually assigned in every session of CBT, asking patients to practice newly learned cognitive or behavioral skills and to generalise such skills to their daily situations in which their problems occur [13]. In general, homework has been shown to facilitate improvement in depression, anxiety or other client problems through CBT [14] but patients often find it difficult to complete the homework [15‐19].

In order to understand and possibly facilitate the process of CBT, it is therefore important to examine which aspects of the homework are important in this therapeutic process. Client characteristics, therapist characteristics, characteristics of the task, and interrelationships among these components have been considered to influence the homework compliance [20]. Among the first components, the clients’ motivational level may be one of the important elements for enhancing homework compliance. Some psychological interventions, such as motivational interviews, are aimed at enhancing enthusiasm [21]. However, no study has yet examined whether the enthusiasm to complete homework assignments thus enhanced can lead to improved process and outcome of CBT.

Moreover, the extant studies of homework in CBT suffers one crucial methodological weakness. In the literature, homework compliance has usually been assessed either post-hoc after the treatment is over, thus risking the recall bias [22, 23], or only once out of the 10 or more sessions of the program, thus possibly not reflecting the overall compliance [24, 25]. Even when homework compliance was measured several times, only the average of those several values was used to predict the outcome of the treatment [26, 27]. To the best of our knowledge, no study to date has examined the session-by-session relationships between homework and an improvement through CBT.

We have previously conducted a randomized controlled trial of CBT administered via telephone among employees with subthreshold depression in a large company in Japan. This telephone CBT (tCBT) program was shown to have a large effect, in comparison with a waiting list control group, with an effect size of around 0.7 for the primary outcome of general psychological distress including depression and anxiety [28]. In this trial, we applied motivational interviewing techniques and measured the enthusiasm of the participants for completing homework at the end of each session.

The current study therefore focused on enthusiasm rather than compliance and aimed to examine the session-by-session relationships between the participants’ enthusiasm to complete the homework and the therapeutic outcomes using structural equation modeling (SEM). We hypothesized that high enthusiasm to complete the homework assignments would be associated with improvement of psychological distress in the subsequent session.

This study is a secondary analysis of data from a previous RCT. The original RCT was approved by the Ethics Review Committee of Nagoya City University Graduate School of Medicine, and the present re-analysis has been approved by the Ethics Committee of Kyoto University Graduate School of Medicine. The original report of this RCT adhered to CONSORT guidelines and the research design and methods of the RCT have been described in detail in a previous publication [28]. Summarized briefly, the study involved an RCT to compare the effectiveness of tCBT in addition to an Employee Assistance Program (EAP) and EAP alone for the treatment of subthreshold depression among workers at a large manufacturing company in Japan. At the initial screening, potential participants were asked to provide their written informed consent to fill in the screening questionnaire after full explanation of the purpose and procedure of the study. Then those who fulfilled the eligibility criteria were asked to provide their final informed consent to participate in the intervention study. Two thousand three hundred twenty-two employees were assessed for eligibility, of whom 145 met the eligibility criteria and invited to participate in the RCT. Twenty-seven declined and 118 finally took part in the original RCT. Fifty-eight were allocated to the tCBT arm and started the telephone sessions immediately. Sixty were allocated to the EAP alone arm (waiting list) and had to wait for four months, complete the end-of-treatment questionnaire and, if they still desired, began receiving the tCBT sessions (See Appendix: Figure 1)

The present study is the first to examine the session-by-session relationships between enthusiasm to engage in homework and changes in psychological distress through the course of behavior therapy for subthreshold depression using structural equation modeling. The best fitting model suggested that distress severity at the beginning of a session did not influence the enthusiasm for completing the homework from that session (estimated r = 0.0), which then, however, was negatively associated with the distress severity of the following session (r = −0.21 to −0.24).

The largest strength of this research is the examination of the session-by-session relationships between the enthusiasm to complete homework and psychological distress severity. A number of preceding studies have shown that homework in CBT was associated with the improvement of depression and anxiety; in these studies, however, homework compliance was measured either as average throughout the CBT sessions or at one out of the many sessions, while depression was typically measured at the end of treatment. In this study, by contrast, we examined the session-by-session relationships between homework and distress and found that enthusiasm to complete homework did indeed lead to a reduction in distress at the time of the subsequent session.

Another strength of our research is that we measured enthusiasm to engage in homework at the end of each session with regard to the particular homework assignments of that session. During the sessions, the therapists used Socratic questions to enhance the clients’ motivations while also modifying the assignments and troubleshooting possible barriers if necessary. In other words, in contrast with prior studies, which studied homework compliance in terms of the completion of the assigned homework ex post facto, we measured a variable that therapists can work on collaboratively with their clients and found that such enthusiasm did lead to distress symptoms reduction during the following week. It is also important to note that this enthusiasm was independent of the distress measured at the beginning of the session. In other words, therapists can strive to enhance the clients’ enthusiasm for homework regardless of the initial psychological distress severity.

It is important to note that not only distress, but also enthusiasm thus measured showed strong session-to-session correlations. It is natural to expect that distress severity at one session would predict distress severity at the subsequent session. The very strong correlation between enthusiasm measurements throughout the sessions may be a reflection of the patient’s personality (e.g., tenacity to engage in assignments), his/her determination and preference for BT, and/or a stable collaborative relationship between the therapist and client. One could argue that there may be little room for the clinician to work on, if the former were the only factor behind the observed strong correlations. The observational nature of the current study precludes any further elucidation in this regard. However, our study does suggest that enhancing this factor through a collaborative therapist-patient relationship could contribute to a reduction in distress in the coming week.

There are several limitations in this study that should be acknowledged. First, as this research was a secondary analysis of a completed trial, the sample size was limited by the available data and was not as large as one would have preferred. This may have led to some possible type II errors (dismissing true associations). On the other hand, we were able to observe several statistically significant correlations that would be clinically meaningful. Second, as our sample consisted of employees with subthreshold depression who were not seeking clinical help and who were working, our results may not be readily generalizable to moderately or severely depressed clinical cases. Moreover, the majority of our samples were men, and this would also affect to generalizability. This limitation was primarily due to the high proportion of male employees in the company where the original RCT took place. Third, as the participants had made most of their improvement through the BT sessions alone, our session-by-session analyses of the relationships between homework enthusiasm and improvement in psychological distress were limited to the BT sessions and we could not investigate them through the CT sessions conducted in the latter half of the program. Thus our results would apply to BT homeworks but might not to CBT homeworks in general. Forth, although we estimated enthusiasm using two questions with satisfactory internal consistency reliability, as described in the Methods section, there is no standardised method to measure enthusiasm, so more refinement in this direction is desirable. Fifth, client characteristics other than enthusiasm and level of distress, such as perfectionism and fear of failure [20], and therapist characteristics were not examined in the present study. Further research is needed to examine interplay of such characteristics for homework completion and improvement of psychological distress or depression.

In conclusion, the clinical and research implications of the present study may be as follows. Clinicians now have empirical evidence when they strive to work collaboratively with their clients to enhance their enthusiasm for homework, since it was shown that a higher enthusiasm leads to a reduction in distress during the subsequent week; moreover, they can and should do this regardless of the clients’ distress severity at that session because these two factors are not correlated. Research into homework in CBT should no longer be content in making global measurements once at the end of the treatment, but such research should examine its vicissitudes and mutual influences throughout the sessions using statistical methods that allow longitudinal modeling. Only then can such research illuminate what each clinician and patient can do in each session.