0093 Efficacy of a Chatbot-Based Sleep Intervention on Sleep Quality Improvement among Young Adults. (25th May 2022)

• Record Type: Journal Article
• Authors: Oh, Yoo Jung; Zhang, Jingwen; Ji, Xiaopeng; Liao, Wang; Feng, Bo
• Is Part Of: Sleep. Volume 45(2022)Supplement 1
• Page Start: A42
• Page End: A42
• Publication Date: 2022-05-25
• Subjects: Sleep -- Physiological aspects -- Periodicals; Sleep disorders -- Periodicals; Sommeil -- Aspect physiologique -- Périodiques; Sommeil, Troubles du -- Périodiques; Sleep disorders; Sleep -- Physiological aspects; Sleep -- physiological aspects; Sleep Wake Disorders; Psychophysiology; Electronic journals; Periodicals; 616.8498
• DOI: 10.1093/sleep/zsac079.091
• Languages: English
• ISSNs: 0161-8105

Abstract

Abstract: Introduction: Insufficient sleep duration and poor sleep quality are prevalent in adolescents and young adults. Artificial-intelligent agents such as chatbots have been integrated into digital health interventions to support symptom management and promote health behaviors. This pilot study tested the efficacy of a sleep advice chatbot in young adults with self-reported sleep disturbances. Methods: Young adults aged 18 to 28, who scored 55 and above (t-score) on the PROMIS Sleep Disturbance (Short Form) were enrolled into the study. Using a 2 x 2 research design, participants were randomly assigned into four groups: chatbot 1 (sleep advice), chatbot 2 (sleep advice & emotional support), human 1(sleep advice), human 2 (sleep advice & emotional support). Sleep advice chatbots were developed using Google's Dialogflow. After the conversation, participants rated the effectiveness of the sleep advice on a 5-point Likert scale. Participants completed the PROMIS Sleep Disturbance seven days after the study. Results: Of 293 participants who were enrolled, mean age was 19.75 (SD = 1.57) and 85.1% were females. Average sleep disturbance scores measured before and after the study was 25.47 (SD = 4.91) and 20.97 (SD = 5.55), showing a significant decrease in sleep disturbance (p < .001). Mean advice effectiveness was 3.96 (SD = 1.10), with human advice provider receiving better scores (p < 0.05). Controlling for age and sex, two-way ANOVA analysis showed a significant interaction between advice giver type (chatbot vs. human) and advice type (sleep advice vs. sleep & emotional support) on alleviating sleep disturbances [F(1, 262) = 4.07, p = .045, partial η2 = .015]. Specifically, participants who received sleep advice from a chatbot had greater decreases in sleep disturbances (Mdiff = -4.69, SD = 3.65) than from a human (Mdiff = -3.78, SD = 5.92). On the other hand, with emotional support, advice given by humans was more effective (Mdiff = -5.56, SD = 5.86) than chatbots (Meandiff = -3.44, SD = 6.31). Conclusion: The results of this study suggest the potential efficacy of a chatbot-based sleep advice intervention among young adults. Future research can consider using a combination of chatbot and human expert-based sleep interventions. Support (If Any):