Purpose Sleep disturbances are highly prevalent in neurorehabilitation inpatients and impede functional recovery. Clinically feasible sleep assessments are essential for individuals with communication or cognitive impairments, yet few appropriate instruments exist. We aimed to identify suitable proxy tools by evaluating the psychometric performance, interrater reliability, agreement, and validity of patient-reported (RCSQP) and nurse-rated (RCSQN) proxy Richards–Campbell Sleep Questionnaire. Our aim was to determine their relationships with subjective sleep measures Pittsburgh Sleep Quality Index (PSQI), and objective rest–activity rhythm (RAR) parameters derived from actigraphy. Methods Adults admitted to a neurorehabilitation ward completed the RCSQP and PSQI at admission and discharge; night-shift nurses independently completed the RCSQN. A subgroup underwent wrist actigraphy to obtain RAR parameters (M10, L5, RA, IS, IV). Interrater reliability, internal consistency, agreement, concurrent validity, ROC-derived optimal cutoffs were analyzed. Results 88 participants were enrolled; 36 underwent actigraphy. Both RCSQP and RCSQN showed excellent internal consistency (α = 0.94–0.98) and good interrater reliability (ICC = 0.74–0.86), though nurses consistently reported better sleep. RCSQP correlated strongly with PSQI, whereas RCSQN showed moderate correlations. ROC analyses demonstrated excellent discrimination for RCSQP (AUC 0.94–0.95) with newly identified cutoffs of 68–72, and good discrimination for RCSQN (AUC 0.78–0.85) with higher cutoffs (88–94). Good sleepers exhibited higher M10 at admission and lower L5 at discharge. Conclusions Both RCSQ versions demonstrated strong psychometric properties in neurorehabilitation inpatients. With appropriate cutoff adjustment, RCSQN provides a viable proxy for patients unable to self-report and may improve identification of sleep disturbances during neurorehabilitation. Brief summary Current Knowledge/Study Rationale Sleep disturbances are highly prevalent in neurorehabilitation inpatients and can impede functional recovery. Since few appropriate and feasible assessment tools exist for patients with communication or cognitive impairments, proxy measures, such as the nurse-rated Richards–Campbell Sleep Questionnaire (RCSQN), require validation in this population. Study Impact The study confirmed proxy RCSQ demonstrated strong psychometric property with moderate correlation with the Pittsburgh Sleep Quality Index; RCSQN showed good discrimination (AUC 0.78–0.85) with newly identified cutoffs. With appropriate cutoff adjustment, the RCSQN is validated as a viable proxy for neurorehabilitation patients unable to self-report, facilitating the early identification and management of sleep disturbances to optimize neurorehabilitation outcomes.

Application of the patient-reported and nurse-rated Richards-Campbell Sleep Questionnaire in neurorehabilitation inpatients

期刊论文
发表日期：

2026年4月9日

查看全文

Shao-Yu Chen,
Yu-Hsuan Lin,
Huey-Wen Liang

摘要

Purpose

Sleep disturbances are highly prevalent in neurorehabilitation inpatients and impede functional recovery. Clinically feasible sleep assessments are essential for individuals with communication or cognitive impairments, yet few appropriate instruments exist. We aimed to identify suitable proxy tools by evaluating the psychometric performance, interrater reliability, agreement, and validity of patient-reported (RCSQ_P) and nurse-rated (RCSQ_N) proxy Richards–Campbell Sleep Questionnaire. Our aim was to determine their relationships with subjective sleep measures Pittsburgh Sleep Quality Index (PSQI), and objective rest–activity rhythm (RAR) parameters derived from actigraphy.

Methods

Adults admitted to a neurorehabilitation ward completed the RCSQ_P and PSQI at admission and discharge; night-shift nurses independently completed the RCSQ_N. A subgroup underwent wrist actigraphy to obtain RAR parameters (M10, L5, RA, IS, IV). Interrater reliability, internal consistency, agreement, concurrent validity, ROC-derived optimal cutoffs were analyzed.

Results

88 participants were enrolled; 36 underwent actigraphy. Both RCSQ_P and RCSQ_N showed excellent internal consistency (α = 0.94–0.98) and good interrater reliability (ICC = 0.74–0.86), though nurses consistently reported better sleep. RCSQ_P correlated strongly with PSQI, whereas RCSQ_N showed moderate correlations. ROC analyses demonstrated excellent discrimination for RCSQ_P (AUC 0.94–0.95) with newly identified cutoffs of 68–72, and good discrimination for RCSQ_N (AUC 0.78–0.85) with higher cutoffs (88–94). Good sleepers exhibited higher M10 at admission and lower L5 at discharge.

Conclusions

Both RCSQ versions demonstrated strong psychometric properties in neurorehabilitation inpatients. With appropriate cutoff adjustment, RCSQ_N provides a viable proxy for patients unable to self-report and may improve identification of sleep disturbances during neurorehabilitation.

Brief summary

Current Knowledge/Study Rationale Sleep disturbances are highly prevalent in neurorehabilitation inpatients and can impede functional recovery. Since few appropriate and feasible assessment tools exist for patients with communication or cognitive impairments, proxy measures, such as the nurse-rated Richards–Campbell Sleep Questionnaire (RCSQ_N), require validation in this population.

Study Impact The study confirmed proxy RCSQ demonstrated strong psychometric property with moderate correlation with the Pittsburgh Sleep Quality Index; RCSQ_N showed good discrimination (AUC 0.78–0.85) with newly identified cutoffs. With appropriate cutoff adjustment, the RCSQ_N is validated as a viable proxy for neurorehabilitation patients unable to self-report, facilitating the early identification and management of sleep disturbances to optimize neurorehabilitation outcomes.