Insomnia describes the inability to sleep, including difficulty falling asleep, difficulty staying asleep, difficulty sleeping, and early waking, despite an attempt to sleep at the right time and in the right environment. This condition reduces an individual’s quality of life by causing various daytime dysfunctions, including sleepiness during the day [1]. Insomnia can be caused by a variety of factors, including changes in social life such as decreased interpersonal interaction and socializing, as well as mental and physical illnesses, decreased physical function, and physiological changes in the brain. Insomnia impairs daytime activity and triggers additional illnesses such as depression, lifestyle-related diseases, and cancer. Insomnia is a common complaint in adults, with 13.5% to 20.6% of the general population experiencing daytime sleepiness >3 times per week [2]. It has been reported that 14.9% of people in Japan have profound daytime sleepiness [3]. Another study found that 1 in 5 Japanese adults had sleep problems [4]. In Japan, the economic losses caused by sleep deprivation and insomnia have been estimated to be 15 trillion yen (US $8.67 billion) [5,6].

From the perspective of optimizing medical costs associated with lifestyle-related diseases, such as diabetes, hypertension, and hyperlipidemia, lifestyle guidance is increasingly being introduced as a complement to the use of medication. However, the evidence base to support the introduction of lifestyle guidance is lacking. The provision of face-to-face advice paired with an insomnia improvement program delivered as easy-to-use internet-based life guidance will likely reduce medical expenses.

Insomnia practice guidelines in the United States and Australia recommend cognitive behavioral therapy (CBT) as the first choice for the treatment of insomnia. In Japan, there are very few medical institutions that can provide CBT, and there are no barriers to prescribing psychotropic drugs outside of psychiatry and psychosomatic medicine. However, pharmacotherapy is associated with side effects such as dependence, tolerance, anterograde amnesia, muscle relaxant effects, carryover effects, and rebound insomnia. Furthermore, it increases the risk of double prescription and overdose and requires long-term treatment. However, the success of pharmacotherapy is limited. According to the Japanese Society of Sleep Studies guidelines, it is preferable to prioritize CBT over pharmacotherapy for insomnia.

CBT for insomnia is a treatment that focuses on anxiety and biased thinking about sleep and improves insomnia by reviewing an individual’s lifestyle, existing anxiety and tension, and any thinking related to the maintenance of insomnia. Sleep hygiene instructions can provide important knowledge about sleep, including how to adjust living conditions so that good-quality sleep can be obtained, and practice measures for maintaining good-quality sleep throughout life. This study used internet-based CBT (ICBT) without email support to provide sleep hygiene guidance as a trial treatment. ICBT was paired with a self-help intervention as a form of noninvasive self-medication.

Positive psychology focuses on positive psychological traits such as happiness, optimism, and satisfaction with life and was developed by Seligman et al [7] of the University of Pennsylvania based on the reflection that traditional psychology was excessively focused on normalizing negative traits such as mental illness. Positive psychology is based on the idea of nurturing [7]. A typical method of positive psychology comprises the exercise of writing 3 good things every day before going to bed every night [7]. It is a simple diary-like exercise that involves listing 3 good things that happened that day and providing a written explanation of why those things happened. In this study, this exercise was called the Three Good Things (TGT) exercise.

ICBT has gained increasing attention as a possible treatment for improving insomnia, and many studies on ICBT have been conducted. ICBT provides web-based CBT and sleep hygiene guidance, simulating an ordinary CBT session (counseling) that is typically conducted by the patient and therapist face-to-face. With the approval of the Chiba University School of Medicine Hospital clinical trials ethics review board (approval number G27040), the combination of guided ICBT for insomnia and routine care (ie, usual care [UC]) has been shown to significantly reduce the Pittsburgh Sleep Quality Index (PSQI) score [8,9]. In a randomized controlled trial (RCT) of 23 patients with insomnia whose symptoms persisted even after taking sleep medications such as benzodiazepines, the adjusted mean change in total PSQI score from baseline to 6 weeks was −6.11 in the ICBT+UC group (n=11), which was significantly better than the 0.40 change seen in those who underwent UC alone (n=12; P<.001). In addition, a significant improvement was seen for the adjusted mean changes in PSQI score; sleep onset latency (SOL); sleep efficiency (SE); number of awakenings; and mean change in depression at 3, 6, and 12 weeks from baseline in the ICBT+UC group. No adverse events were reported.

Seligman et al [7] previously evaluated the effects of the TGT exercise in a RCT. According to that study, the group that performed this exercise daily for 1 week not only increased self-reported levels of happiness but also reported decreased depression compared with the control group, and the effect persisted after 6 months. To the best of our knowledge, no studies have examined whether the TGT exercise can reduce insomnia. However, recent reviews have shown that there is a correlation between positive emotions and sleep [10]. TGT may improve sleep by improving positive emotions.

This study aims to determine the effectiveness of an unguided, web-based self-help intervention in reducing insomnia. Using an RCT design, a nonclinical population of adults with insomnia was allocated to one of three groups: (1) the ICBT group, (2) the TGT group, and (3) the waitlist group. It is hypothesized that unguided ICBT and TGT would result in higher quality sleep in the treatment groups than in the waitlist group.

We report this RCT trial in accordance with the CONSORT-EHEALTH (Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth) version 1.6 checklist (Multimedia Appendix 1).

This study comprised an exploratory, parallel-group (3 groups), randomized, open-label, controlled study, incorporating a nonintervention group (waiting list control [WLC]). The registration of the participants was started in February 2019.

An email was sent to the registered monitors owned by the internet research company that commissioned the research, and a preliminary survey was conducted on the internet in relation to the inclusion and exclusion criteria given in the following sections, after which informed consent was obtained.

Potential participants were required to meet all the following inclusion criteria at the time of the preliminary survey: (1) aged 20 to 70 years (regardless of sex); (2) patients with mild or severe sleep disorders, who scored ≥6 on the Athens Insomnia Scale (AIS) and ≥6 on the PSQI; (3) sleep disorder occurring at least three nights per week and lasting for at least 3 months; (4) access to internet use with PCs, smartphones, and tablets; and (5) understood the explanations in Japanese and freely provided web-based consent.

Exclusion criteria of the study included (1) those with moderate or severe anxiety symptoms, scoring ≥10 on the Generalized Anxiety Disorder-7 (GAD-7) questionnaire; (2) those with moderate or severe depression symptoms, scoring ≥10 points on the Patient Health Questionnaire-9 (PHQ-9); (3) the confirmed presence of alcohol or drug dependence, diagnosed by a medical institution within the past year, excluding tobacco smokers; (4) suicidal thoughts, with a score of ≥2 in the Q9 part of the PHQ-9; (5) the confirmed diagnosis of cerebral organic diseases such as sleep apnea syndrome, restless leg syndrome, epilepsy, dementia, or cerebrovascular disease; (6) the confirmed diagnosis of a severe progressive physical illness, such as cancer or heart failure; (7) the use of pharmacotherapy for mental disorders, including insomnia symptoms; and (8) night shift work between 8 PM and 2 AM.

Individuals who met the above conditions and provided consent were enrolled as study participants. This study excluded patients commonly treated with pharmacotherapy for insomnia.

Case registration was performed via the trial website. Among the registered members (candidate participants) of monitors owned by the internet research company, those who met the conditions were automatically assigned participant identification codes when they accessed the servers. Members voluntarily underwent a preliminary survey after explanation and agreement, following the distribution of the recruitment materials. In the first preliminary survey, AIS, PSQI, GAD-7, and PHQ-9 were administered to select and screen participants for inclusion. Only those who met the inclusion criteria and did not meet any of the exclusion criteria were able to proceed to the next preliminary survey. The second preliminary survey was conducted 2 weeks after the first preliminary survey and contained the same content as the first survey. Only those who met the inclusion criteria and did not meet any of the exclusion criteria proceeded to take part in the study. Responses from the second preliminary survey were used in the analysis as baseline values before the intervention. After the participants were randomly assigned, the method of accessing the intervention program assigned to each participant was provided by email.

Allocation modifiers included baseline PSQI scores (≥12 and ≤11) and gender (male and female).

This study was approved by the clinical trials ethics review board of the Chiba University Hospital (registration number G30022) and was registered as a clinical trial (UMIN000034927). A document explaining consent was presented to the participants on a webpage, accompanied by a verbal explanation as part of a video animation by the principal investigator. After viewing these materials, individuals who freely agreed to participate were recruited into the study.

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

Figure 1 shows the research flow based on the CONSORT (Consolidated Standards of Reporting Trials) guidelines. Of the 21,394 individuals contacted through the internet research company, consent to participate in the study was obtained from 312 (1.46%) individuals who met the eligibility criteria and were randomly assigned to the ICBT, TGT, or WLC group. Of the 106 people assigned to ICBT, 23 (21.7%) did not start the ICBT program at least once, and 4 (3.8%) who performed the ICBT program at least once were classified as trial deviations, leaving 79 (74.5%) individuals with a FAS.

For details of the deviation related to PSQI, there was an error in the formula set by the internet survey company that conducted the web questionnaire, and the correct score was ≤5; however, it was judged to be ≥6 by incorrect calculation. Therefore, of the 312 individuals, 4 (1.3%) individuals who had initially been selected as participants and who should not have been selected underwent the ICBT program; these individuals were reported as deviations to the Chiba University Hospital clinical trials ethics review committee and accepted as such.

Separately, there was an error in the formula used by the internet survey company that conducted the web questionnaire. In the error, the correct PSQI total score was ≥6; however, it was judged incorrectly to be ≤5. The correct PHQ-9 total score should have been calculated from the 1st to 9th items (the 10th item should have been excluded); however, the internet survey company that conducted the web questionnaire mistakenly calculated from the 1st to the 10th item. In addition, the correct GAD-7 total score should have been calculated from the first to seventh items (the eighth item was not added); however, the internet survey company that conducted the web questionnaire mistakenly calculated from the first to eighth items. For these reasons, of the 21,394 individuals, 83 (0.39%) were not selected in the first screening, and 25 (0.12%) were not selected in the second screening.

Of the 106 participants assigned to the ICBT group, 74 (69.8%) provided data at 4 weeks, 68 (64.2%) provided data at 8 weeks, and 55 (51.9%) had an attendance of ≥75% during the ICBT program; thus, the PPS was 55. Of the 103 participants assigned to the TGT group, 15 (14.6%) were excluded who did not start the TGT program at least once; thus, the FAS was 88. In the TGT group, of the 103 participants, 84 (79.2%) provided data at 4 weeks, and 81 (78.6%) provided data at 8 weeks. Approximately 76.7% (79/103) of participants had an attendance of ≥75% during the TGT program; thus, the PPS was 79.

Of the 312 participants, 103 (33%) were assigned to the WLC group; thus, the FAS score for this group was 103. Data were obtained successfully for 88.3% (91/103) of these participants at 4 weeks and for 84.5% (87/103) of participants at 8 weeks; thus, the PPS was 103. The registration of participants started in February 2019, and the follow-up ended in May 2019.

There were no adverse events. When adverse events occurred for participants, we asked them to report the adverse events by email. However, there were no reports.

Table 1 shows the characteristics of the participants at baseline. Female participants comprised 41.5% (112/270; ICBT group 32/79, 41%; TGT group 39/88, 44%; and WLC group 41/103, 39.8%) of the study population, with a mean age of 50.4 (SD 10.8; ICBT group: 49.8, SD 11.1; TGT group: 50.5, SD 11.0; and WLC group: 51.0, SD 10.4) years. There were no significant differences in sex, age, marital status, educational history, or employment status among the 3 groups.

Table 2 presents the baseline data for the primary and secondary outcome measures. There were no significant differences among the 3 groups for the PSQI, AIS, GAD-7, or PHQ-9 scores.

The results of this study suggest that a self-help internet intervention without email support is effective as a noninvasive self-medication for adults with insomnia. The ICBT group showed a significant change of −6.11 in the PSQI score. The participant characteristics of this study differed significantly from those of similar previous studies. A previous study by van Straten et al [22] showed the efficacy of guided ICBT using an RCT design, in which 1500 patients with insomnia were invited by email, and 118 patients were assigned to the ICBT group (n=59) and a WLC group (n=59). However, the reported baseline mean PSQI score in the ICBT group in that study was relatively severe (12.4, SD 2.1). The mean change in PSQI scores in the ICBT group was −3.5, and in the WLC group it was −0.1. Our previous study of ICBT guided by email support reported a severe baseline PSQI mean score of 13.5 (SD 2.7); furthermore, participants had persistent symptoms despite taking sleep medications such as benzodiazepines [9].

In this study, the unguided ICBT group showed a change of −2.19, the TGT group showed a change of −1.79, and the WLC group showed a change of −0.64; the mean change in PSQI scores was smaller than those reported by van Straten et al [22] and Sato et al [9]. This could be because of the baseline average score of PSQI being 9.8, which indicates relatively mild insomnia. In addition, given that the ICBT was unguided, participant motivation to continue in this study may not have been well-supported.

The TGT group showed a significant decrease in PSQI scores compared with the WLC group only at the end of the program, that is, at the 4-week time point. To the best of our knowledge, no previous studies have examined the effects of TGT on insomnia through an RCT. However, as no significant improvement in positive emotions (indicated by the actual items on the CES-D) was observed in the TGT group, this casts doubt on the existence of a mechanism for the improvement of insomnia through the improvement of positive emotions. Contrary to previous studies, the TGT exercise did not lead to the improvement of positive emotions, and thus, it is unclear whether TGT really contributes to the improvement of happiness.

In this study, comparing the ICBT group and the TGT group was not the original purpose; however, the number of participants who received >75% of the intervention in the ICBT group was 70% (55/79) of participants compared with 90% (79/88) of participants who received >75% of the intervention in the TGT group. This could have been because of difficulties with regular participation in the ICBT program.

At the 4- and 8-week follow-ups, the ICBT group showed a significant improvement in PSQI scores over the WLC group, whereas the TGT group did not show a significant improvement over the WLC group. Therefore, ICBT could have longer-term effectiveness. If further research shows that ICBT results in sustained effects on insomnia, and TGT is easier to participate in more regularly, it may be necessary to consider an unguided intervention that combines both ICBT and TGT to achieve optimal outcomes.

In this study, there were errors in the calculation set by the internet research company that conducted the web questionnaire for PSQI, PHQ-9, and GAD-7 scores, which may have affected the eligible participants. In future studies, the distribution of inclusion and exclusion criteria should be carefully monitored to avoid similar mistakes.

Some participants in the ICBT and TGT groups did not participate more than once. However, in the WLC group, all participants were included in the final analysis. To ensure consistency with the ICBT and TGT groups in future research, it will be necessary to select those who have participated at least once in the WLC group for inclusion in the final analysis, or alternatively a placebo sham program could be implemented.

In conclusion, this RCT provided evidence that 4 weeks of unguided ICBT and TGT exercise for adults with insomnia, both administered as self-help internet interventions without email support, may improve insomnia symptoms compared with the WLC group. The findings also suggest that the effects of ICBT may last longer than TGT, whereas TGT may be easier to participate in more regularly. However, this study experienced errors during the selection process. Further research is warranted to examine the effectiveness of a combined intervention of ICBT and TGT for insomnia.