Reliability and validity of the revised Chinese version of the Children's Sleep Habits Questionnaire (CSHQ-CH-R)

The Chinese version of the Children's Sleep Habits Questionnaire (CSHQ-CH) has been widely used in clinical practice and research since it was developed in 2007. This study aimed to examine the psychometric properties of the revised version (CSHQ-CH-R) and establish cut-off values for screening for sleep disturbances in Chinese preschool-aged and school-aged children.

A community sample of 3850 preschool and school-aged children aged 3-12 years were obtained from 6 kindergartens and 2 primary schools in Shanghai, Beijing and Chengdu, China, from December 2019 to February 2020. A clinical sample of 119 children aged 3-12 years who were diagnosed with sleep disorders were recruited from a tertiary pediatric hospital in Shanghai from September 2019 to November 2021. Caregivers completed the original 33 items of the CSHQ-CH plus an additional 15 items through an online platform. Item analysis was conducted to develop the CSHQ-CH-R, which included discriminative validity, redundancy and ambiguity. Internal consistency, interobserver reliability, test-retest reliability, concurrent validity (correlations with actigraphy) and discriminative validity were then analyzed. The cut-off values for both the full scale and subscale scores were generated by receiver operating characteristics.

The final version of the CSHQ-CH-R included 30 items and 8 subscales. Three subscales (bedtime resistance, sleep duration and daytime sleepiness) were revised, and the other five subscales were unchanged from those of the CSHQ-CH. The internal consistency for the full scale was 0.83 for the community sample, and 0.73 for the clinical sample. The internal consistency ranged from 0.47 (sleep duration) to 0.72 (bedtime resistance) for the community sample, and 0.38 (sleep duration) to 0.83 (sleep-disordered breathing) for the clinical sample. The interobserver reliability for the full scale was 0.55, and ranged from 0.38 (parasomnias) to 0.74 (bedtime resistance) for the subscales. The test-retest reliability was 0.70 for the full scale, and ranged from 0.54 (sleep duration) to 0.78 (bedtime resistance) for the subscales. Concurrent validities against actigraphy were 0.62 and 0.33 for "time in bed" and "night sleep duration", respectively. The clinical sample scored higher than the community sample did for the full scale and subscales (P < 0.01). A cut-off total score of 44 was determined, with an area under the curve of 0.76, a sensitivity of 0.64, and a specificity of 0.76. To facilitate cross-cultural comparisons, psychometric properties and cut-off values for the CSHQ-CH with the original 33 items were also provided.

The CSHQ-CH-R exhibits satisfactory to good psychometric properties, indicating its suitability for identifying sleep problems among Chinese preschool- and school-aged children.