Differences in the effect of Internet-based cognitive behavioral therapy for improving nonclinical depressive symptoms among workers by time preference: Randomized controlled trial

Kotaro Imamura, Toshi A. Furukawa, Yutaka Matsuyama, Akihito Shimazu, Kazuto Kuribayashi, Kiyoto Kasai, Norito Kawakami

Research output: Contribution to journalArticle


Background: Previous randomized controlled trials (RCTs) have shown a significant intervention effect of internet-based computerized cognitive behavioral therapy (iCBT) on improving nonclinical depressive symptoms among healthy workers and community residents in a primary prevention setting. Time preference is one's relative valuation for having a reward (eg, money) at present than at a later date. Time preference may affect the effectiveness of cognitive behavioral therapy. Objective: This RCT aimed to test the difference of intervention effect of an iCBT program on improving nonclinical depressive symptoms between two subgroups classified post-hoc on the basis of time preference among workers in Japan. Methods: All workers in one corporate group (approximate n=20,000) were recruited. Participants who fulfilled the inclusion criteria were randomly allocated to either intervention or control groups. Participants in the intervention group completed 6 weekly lessons and homework assignments within the iCBT program. The Beck Depression Inventory-II (BDI-II) and Kessler's Psychological Distress Scale (K6) measures were obtained at baseline and 3-, 6-, and 12-month follow-ups. Two subgroups were defined by the median of time preference score at baseline. Results: Only few (835/20,000, 4.2%) workers completed the baseline survey. Of the 835 participants, 706 who fulfilled the inclusion criteria were randomly allocated to the intervention or control group. Participants who selected irrational time preference options were excluded (21 and 18 participants in the intervention and control groups, respectively). A three-way interaction (group [intervention/control] × time [baseline/follow-up] × time preference [higher/lower]) effect of iCBT was significant for BDI-II (t 1147.42 =2.33, P=.02) and K6 (t 1254.04 =2.51, P=.01) at the 3-month follow-up, with a greater effect of the iCBT in the group with higher time preference. No significant three-way interaction was found at the 6- and 12-month follow-ups. Conclusions: The effects of the iCBT were greater for the group with higher time preference at the shorter follow-up, but it was leveled off later. Workers with higher time preference may change their cognition or behavior more quickly, but these changes may not persist.

Original languageEnglish
Article numbere10231
JournalJournal of medical Internet research
Issue number8
Publication statusPublished - 2018 Aug 1
Externally publishedYes



  • Internet-based computerized cognitive behavioral therapy
  • Nonclinical depressive symptoms
  • Time preference
  • Workers

ASJC Scopus subject areas

  • Health Informatics

Cite this