We had an online research agency (Cross Marketing Inc, Tokyo) oversee our Web-based survey. After being provided a thorough understanding of our research and agreeing to voluntarily participate in the study, 600 participants were recruited from the Kanto district in Japan through the online research provider. They consisted of 200 individuals with insomnia who were using a medication for insomnia, 200 individuals with insomnia who were not taking any insomnia medication, and 200 individuals without insomnia by their self-reports. Each group of 200 participants consisted of 100 men, including 20 men each in their twenties, thirties, forties, fifties, and sixties, and 100 women, including 20 women each in their twenties, thirties, forties, fifties, and sixties, to match the groups by age and gender.

The candidate respondent received brief text-based information about the survey, including its objectives and conditions for participation (informed consent). The survey consisted of two parts. The first part asked demographic questions including gender, age, living area, employment status, and presence and severity of insomnia using the Athens Insomnia Scale (AIS), a self-reported psychometric questionnaire consisting of eight items developed by Soldatos et al [19].

The survey’s second part consisted of the following questions about treatment for insomnia.

Frequency of sleep medication use and the level of satisfaction with sleep medication use:

Knowledge of CBT:

Preference for CBT before drug therapy, preference for CBT versus drug therapy, and preference for internet-based CBT versus face-to-face CBT:

We used descriptive analyses (numbers, frequencies, percentages, means and standard deviations). We compared differences in each survey item among the three respondent groups, using the chi-square test, ANOVA, and a residual analysis [20]. An alpha level of .05 was used. All data were analyzed using SPSS for Windows version 21 (SPSS, Chicago, IL, USA).

The study was approved by the Regional Ethical Review Board, Faculty of Medicine, Chiba University (2017-5-19; No. 2711).

As designed, a total of 600 respondents (300 men and 300 women; mean age 45, SD 14 years, range 20-69 years) completed the online survey (Table 1). As shown by the respondents’ use of the AIS, the respondents with insomnia who were taking an insomnia medication and respondents with insomnia who were not taking an insomnia medication had significantly more severe insomnia compared to the respondents without insomnia. When we defined an AIS total score of 6 or higher as insomnia, 90.0% (180/200) of the respondents with insomnia were using an insomnia medication and 94.5% (189/200) of the respondents with insomnia were not taking an insomnia medication compared to 52.0% (104/200) of the respondents without insomnia (see Table 1). There were also significant differences among the three respondent groups in employment status and AIS score (see Table 1).

The respondents with insomnia who were taking an insomnia medication (n=200) were asked about the frequency of their insomnia medication use; 68.0% (136/200) reported that they used an insomnia medication every night, and 95.5% (191/200) reported using such a medication at least once per week (Table 2). The respondents who indicated that they used an insomnia medication were also asked about their level of satisfaction with the sleep medication use: 54.0% (108/200) were satisfied, 27.5% (55/200) were neutral, and 18.5% (37/200) were dissatisfied.

All respondents were asked about their knowledge of CBT. There were significant differences among the three respondent groups in their knowledge of CBT and the effects of CBT (Table 3): 55.0% (330/600) of the respondents had no knowledge of CBT. Among the three groups, the percentage of those who had no knowledge of CBT were as follows: 39.0% (78/200) of the respondents with insomnia who were taking an insomnia medication, 54.0% (108/200) of the respondents with insomnia who were not taking an insomnia medication, and 72.0% (144/200) of the respondents without insomnia. In the group of respondents with insomnia who were using medication, the response “I have heard of CBT, and I know that it is an insomnia treatment” was significantly more frequent comparing the three respondent groups’ answers. Conversely, the response “I have never heard of CBT, and I did not know that it is an insomnia treatment” was significantly less comparing the three respondent groups’ answers. Among the group of respondents with insomnia who were not taking insomnia medication, the response “I have heard of CBT, and I know that it is an insomnia treatment” was significantly less frequent comparing the three respondent groups’ answers, and “I have heard of CBT, and I did not know that it is an insomnia treatment” was significantly more frequent comparing the three respondent groups’ answers. In the group without insomnia, the response “I have heard of CBT, and I know that it is an insomnia treatment” was significantly more frequent, and the response “I have never heard of CBT, and I did not know that it is an insomnia treatment” was significantly less frequent comparing the three respondent groups’ answers (Table 4).

Even among the respondents who had heard of CBT (n=270), 68.5% did not know that CBT is effective for insomnia. Among the respondents with insomnia who were using an insomnia medication, the response “Do you know that CBT is effective for insomnia? Yes, I know” was significantly more frequent comparing the three respondent groups’ answers, and “No, I do not know” was significantly less frequent comparing the three respondent groups’ answers. Among the survey respondents with insomnia who were not using an insomnia medication, the response “Do you know that CBT is effective for insomnia? Yes, I know” was significantly less frequent comparing the three respondent groups’ answers, and “No, I do not know” was significantly more frequent comparing the three respondent groups’ answers (Table 3).

All respondents were asked whether they had undergone CBT before drug therapy, and 47.7% (286/600) responded they had undergone CBT. There were no significant differences in the rate among the three groups: with insomnia using medication (51.5%,103/200), with insomnia not using medication (47.5%, 95/200), and without insomnia (44.0%, 88/200) (Table 5).

Notably, 57.2% (343/600) of the total respondents preferred CBT over drug therapy for insomnia. Among those with insomnia who were using an insomnia medication, the statement “I choose CBT” was significantly less frequent comparing the three respondent groups’ answers, and “I choose drug therapy” was significantly more frequent comparing the three respondent groups’ answers. In both groups of respondents with insomnia who were not taking an insomnia medication and the respondents without insomnia, the statement “I choose CBT” was significantly more frequent comparing the three respondent groups’ answers, and “I choose drug therapy” was significantly less frequent comparing the three respondent groups’ answers (Table 5). Although respondents with insomnia who were using insomnia medications had a relatively lower preference for CBT (40.5%, 81/200), both those with insomnia who were not taking an insomnia medication and the respondents without insomnia had a relatively higher preference for CBT (64.0%, 128/200 and 67.0%, 134/200, respectively).

Of the total number of respondents, 47.0% (282/600) preferred internet-based CBT for insomnia to face-to-face CBT. In the group with insomnia taking an insomnia medication, the statement “I choose face-to-face CBT” was chosen significantly more frequently comparing the three respondent groups’ answers, and “I choose computerized CBT on the internet” was chosen less frequently when comparing the three respondent groups’ answers. Among the respondents with insomnia not using insomnia medication, “I choose face-to-face CBT” was chosen significantly less frequently when comparing the three respondent groups’ answers, and “I choose computerized CBT on the internet” was more chosen more frequently when comparing the three respondent groups’ answers (Table 5).

Although the group with insomnia using an insomnia medication had a relatively lower preference for internet-based CBT (40.5%, 81/200), the group with insomnia not using medication for it had a relatively higher preference for internet-based CBT (55.5%, 111/200).

Our Web-based survey of 600 individuals in Japan revealed that approximately half (57.2%, 343/600) of the respondents preferred CBT for insomnia to drug therapy, and half (47.0%, 282/600) preferred internet CBT for insomnia over face-to-face CBT.

Culver et al [21] reported that 57.7% of female veterans (N=1538) in the United States rated nonmedication treatment of insomnia as very acceptable, whereas only 33.5% rated medication treatment as very acceptable. Sedov et al [22] reported that 50.9% of a series of pregnant women in Canada (N=187) described CBT as their first choice for the treatment of insomnia, 11.8% selected pharmacotherapy, and 37.3% selected acupuncture if they experienced insomnia. Together, these results suggest a preference for CBT for insomnia over pharmacotherapy [22].

Regarding internet-based CBT, Cheung et al [23] conducted semistructured interviews in Australia, and they reported that 56.86% of their patients with insomnia (N=51) had a preference for face-to-face CBT, and 43.13% had a preference for internet-based CBT. Their results are similar to ours.

The treatment plan for an individual with insomnia should be tailored according to his or her values and preferences. Further research is needed to increase the availability of more effective internet CBT in addition to face-to-face CBT and pharmacotherapy.

Although our online survey obtained valuable information, our study has some limitations including the sampling methods. First, instead of random sampling, we used a stratified sample based on gender and age ranging from people in their twenties to their sixties from an internet inquiry to conduct our online survey. Individuals younger than 20 or older than 70 were excluded. However, all age groups are affected by insomnia, and the incidence tends to increase with age. In a study conducted in the United States, Ancoli-Israel et al [24] reported that 9% of 1000 subjects aged 18 years and older and 20% of 1000 subjects aged 65 years and older had chronic insomnia. Future studies should include people older than 70 years old. Older people have less access to online surveys than the general population. In that case, the use of a face-to-face survey and/or a telephone survey may be necessary to obtain data from people in their seventies. Second, we were unable to elucidate the insomnia severity or age. Third, the number of people using CBT was unclear. The questionnaires should include a number of insomnia patients using CBT with and without medication. Finally, the reason for preference was unclear. The questionnaires should include a reason for choice of medication, face-to-face CBT, and internet-based CBT.

The responses to our online survey indicate that approximately half of the respondents preferred CBT over drug therapy for insomnia, and half preferred internet-based CBT for insomnia over face-to-face CBT.