Non-pharmacological interventions for sleep promotion in hospitalized children

Background: Healthy sleep is an important component of childhood development. Changes in sleep architecture, including sleep stage composition, quantity, and quality from infancy to adolescence are a reflection of neurologic maturation. Hospital admission for acute illness introduces modifiable risk factors for sleep disruption that may negatively affect active brain development during a period of illness and recovery. Thus, it is important to examine non-pharmacologic interventions for sleep promotion in the pediatric inpatient setting.

Objectives: To evaluate the effect of non-pharmacological sleep promotion interventions in hospitalized children and adolescents on sleep quality and sleep duration, child or parent satisfaction, cost-effectiveness, delirium incidence, length of mechanical ventilation, length of stay, and mortality.

Search methods: We searched CENTRAL, MEDLINE, Embase, CINAHL, three other databases, and three trials registers to December 2021. We searched Google Scholar, and two websites, handsearched conference abstracts, and checked reference lists of included studies.

Selection criteria: Randomized controlled trials (RCTs) or quasi-RCTs, including cross-over trials, investigating the effects of any non-pharmacological sleep promotion intervention on the sleep quality or sleep duration (or both) of children aged 1 month to 18 years in the pediatric inpatient setting (intensive care unit [ICU] or general ward setting).

Data collection and analysis: Two review authors independently assessed trial eligibility, evaluated risk of bias, extracted and synthesized data, and used the GRADE approach to assess certainty of evidence. The primary outcomes were changes in both objective and subjective validated measures of sleep in children; secondary outcomes were child and parent satisfaction, cost-effectiveness ratios, delirium incidence or delirium-free days at time of hospital discharge, duration of mechanical ventilation, length of hospital stay, and mortality.

Main results: We included 10 trials (528 participants; aged 3 to 22 years) in inpatient pediatric settings. Seven studies were conducted in the USA, two in Canada, and one in Brazil. Eight studies were funded by government, charity, or foundation grants. Two provided no information on funding. Eight studies investigated behavioral interventions (massage, touch therapy, and bedtime stories); two investigated physical activity interventions. Duration and timing of interventions varied widely. All studies were at high risk of performance bias due to the nature of the intervention, as participants, parents, and staff could not be masked to group assignment. We were unable to perform a quantitative synthesis due to substantial clinical heterogeneity. Behavioral interventions versus usual care Five studies (145 participants) provided low-certainty evidence of no clear difference between multicomponent relaxation interventions and usual care on objective sleep measures. Overall, evidence from single studies found no clear differences in daytime or nighttime sleep measures (33 participants); any sleep parameter (48 participants); or daytime or nighttime sleep or nighttime arousals (20 participants). One study (34 participants) reported no effect of massage on nighttime sleep, sleep efficiency (SE), wake after sleep onset (WASO), or total sleep time (TST) in adolescents with cancer. Evidence from a cross-over study in 10 children with burns suggested touch therapy may increase TST (391 minutes, interquartile range [IQR] 251 to 467 versus 331 minutes, IQR 268 to 373; P = 0.02); SE (76, IQR 53 to 90 versus 66, IQR 55 to 78; P = 0.04); and the number of rapid eye movement (REM) periods (4.5, IQR 2 to 5 versus 3.5, IQR 2 to 4; P = 0.03); but not WASO, sleep latency (SL), total duration of REM, or per cent of slow wave sleep. Four studies (232 participants) provided very low-certainty evidence on subjective measures of sleep. Evidence from single studies found that sleep efficiency may increase, and the percentage of nighttime wakefulness may decrease more over a five-day period following a massage than usual care (72 participants). One study (48 participants) reported an improvement in Children's Sleep Habits Questionnaire scores after discharge in children who received a multicomponent relaxation intervention compared to usual care. In another study, mean sleep duration per sleep episode was longer (23 minutes versus 15 minutes), and time to fall asleep was shorter (22 minutes versus 27 minutes) following a bedtime story versus no story (18 participants); and children listening to a parent-recorded story had longer SL than when a parent was present (mean 57.5 versus 43.5 minutes); both groups reported longer SL than groups who had a stranger-recorded story, and those who had no story and absent parents (94 participants; P < 0.001). In one study (34 participants), 87% (13/15) of participants felt they slept better following massage, with most parents (92%; 11/12) reporting they wanted their child to receive a massage again. Another study (20 participants) reported that parents thought the music, touch, and reading components of the intervention were acceptable, feasible, and had positive effects on their children (very low-certainty evidence). Physical activity interventions versus usual care One study (29 participants) found that an enhanced physical activity intervention may result in little or no improvement in TST or SE compared to usual care (low-certainty evidence). Another study (139 participants), comparing play versus no play found inconsistent results on subjective measures of sleep across different ages (TST was 49% higher for the no play groups in 4- to 7-year olds, 10% higher in 7- to 11-year olds, and 22% higher in 11- to 14-year olds). This study also found inconsistent results between boys and girls (girls in the first two age groups in the play group slept more than the no play group). No study evaluated child or parent satisfaction for behavioral interventions, or cost-effectiveness, delirium incidence or delirium-free days at hospital discharge, length of mechanical ventilation, length of hospital stay, or mortality for either behavioral or physical activity intervention.

Authors' conclusions: The included studies were heterogeneous, so we could not quantitatively synthesize the results. Our narrative summary found inconsistent, low to very low-certainty evidence. Therefore, we are unable to determine how non-pharmacologic sleep promotion interventions affect sleep quality or sleep duration compared with usual care or other interventions. The evidence base should be strengthened through design and conduct of randomized trials, which use validated and highly reliable sleep assessment tools, including objective measures, such as polysomnography and actigraphy.