Sleep is essential for the healing process, especially for infants, children, and adolescents. Sleep supports the immune system and overall growth.1,2 Chronically poor sleepers are more likely to become ill, and they have a slower recovery from illness.3 In children, poor sleep has been associated with excess adiposity, poor emotional regulation, and an overall lower quality of life.4 Despite the important role of sleep, hospitals are notorious for having a poor sleep environment. Parents have identified poor sleep as a significant burden for themselves and their child during hospitalization, with noise cited in adult studies as the main culprit.5,6
A number of factors can affect a child's sleep in the hospital. While variables such as medical treatments are difficult to adjust, other aspects of the patient's environment and care can be modified. The American Academy of Nursing has endorsed the statement, “Don't wake the patient for routine care unless the patient's condition or care specifically requires it.”7(p1) The Centers for Medicare & Medicaid Services also asks specific questions on their Hospital Consumer Assessment of Healthcare Providers and Systems survey regarding the quietness of the hospital environment.8 Pediatric surveys show a top-box score of 63% regarding environmental quietness.9 With recommendations from national nursing organizations, the significance of patient satisfaction, and the effect on hospital reimbursement, it is clear that nursing administrators must heed the consequences of poor hospital sleep.
Most adult sleep research has been conducted in adult critical care units.7 There is minimal general research focusing on hospitalized children's sleep and even less research on the sleep of very young children or children with diverse health problems.10,11 Pediatric sleep studies have surveyed only specific age groups or diagnoses or used subjective measures. Some research has used surveys to compare hospitalized sleep to home sleep for children.12 Participants in studies that used actigraphy to objectively measure sleep were school aged and adolescents with cancer.13-15 Studies describing hospitalized children's sleep have included ages as young as 12 months but without an objective measure.11
Comprehensive research using actigraphy with subjective measures for all pediatric age groups and diagnoses was needed to describe the sleep environment for hospitalized children. This study is a 1st step in describing the environmental factors that may contribute to sleep interruptions for hospitalized children. Many sleep disruptions are amendable, but a comprehensive descriptive study was needed to describe the problem.
The Human Response Model for Sleep and Fatigue in Hospitalized Pediatric Patients–Revised was the theoretical model used.13 The model, originally developed for children with cancer, was revised to represent all diagnoses (Supplemental Digital Content 1, http://links.lww.com/JONA/A691). The model includes the person factors of age, gender, and race/ethnicity and environmental factors of sleep conditions or interruptions. In theory, changing environmental factors can improve sleep duration.
The purpose of this foundational study was to investigate the amount and type of sleep interruptions experienced by hospitalized children at nighttime to determine sleep efficiency. The secondary objectives were: 1) to compare children's sleep efficiency to demographics; 2) to describe the typical hospital sleep environment over a 1-month period; and 3) to evaluate the perception of the sleep environment by hospitalized children 7 years or older or by their parents as proxy if younger than 7 years.
This was a single-site quantitative, descriptive noninterventional prospective study. The hospital's institutional review board gave permission to conduct the study prior to any research activities.
The Actiwatch (PHILIPS, Amsterdam, Netherlands) is an activity/sleep monitor that is worn like a wristwatch. Actiwatches monitor a person's movements in every direction and store this information as awake, rest, or sleep activity states. The data are comparable to polysomnography data on sleep efficiency, duration, and awakenings16 with a sensitivity of 0.97 and accuracy of 0.86.17 An awakening is defined as any awake period lasting more than 5 minutes.18 As is standard with Actiwatch studies, times that the watch was applied or removed were recorded.
The Quietyme system (Quietyme Inc, Neshkoro, Wisconsin) is designed to measure noise, light, humidity, and temperature in hospital environments, with measurements recorded every second they are in use. Sensors are mounted in electrical outlets in patient rooms, hallways, and common areas to collect data on intensity and type of noise.19
Nursing diaries were used to document common interruptions, such as vital signs (VSs), weights, and turning between 10:00 PM and 10:00 AM. In addition, the electronic medical record was used to complete missing diary entries.
Sleep in the Children's Hospital Survey
Three versions of a self-report sleep survey,13 originally developed at the Children's Hospital of Philadelphia, were used in this study. The Sleep in the Children's Hospital (SinCH-SR) was designed for school-age children and older. The young child version (SinCH-YC) was for younger children, with parents helping to complete the survey. The parent version (SinCH-P) allowed parents to report about their night of sleep in the hospital. The surveys elicited responses about typical sleep at home and the last night's sleep in the hospital. Participants were asked what noises, discomforts, or worries made it hard to sleep on a 4-point scale (response options: not at all, a little, somewhat, or a lot). Finally, respondents were queried about whether pain, noises, lights, or nursing interventions bothered them at bedtime, middle of the night, or this morning (response options: yes, no, or don't know). In each section, a category for “other” comment was included. Children's Hospital of Philadelphia used these surveys in an evidence-based practice project. The instrument has face validity, but other psychometric testing is currently underway.13 All authors provided permission for use of the surveys in the current study.
Demographic information collected included the time of admission, admission diagnosis, date of birth, surgeries during current admission, intravenous (IV) fluid infusions, and parent sleeping-in status. These data were used for descriptive comparisons.
Pediatric patients with any diagnosis from 1 Midwestern pediatric hospital were the sample population. Inclusion criteria were as follows: hospitalized on 1 of 2 designated medical/surgical units, between 1 month and 17 years old, admitted by the recruitment day before 4:00 PM, and parents were able to read and speak English. The exclusion criteria were patients with a procedure requiring sedation or surgery within the last 24 hours, in the foster child system, or with parents younger than 18 years. To obtain a proportionate mix of child ages, a stratified sampling was employed using 3 strata: 1 month to 4 years, 5 to 12 years, and 13 to 17 years. Each age stratum was assigned to each study number and logged into a secure master list. Limitations for the number in each age bracket were based on the unit census data from the previous year.
Prior to enrolling patients, Quietyme environment readers were installed on each hospital unit in all patient rooms, hallways, computer rooms, and break rooms for 1 month to collect surrounding noise (decibel levels) and light level data. These data were not linked to patients.
After the Quietyme data collection was complete, patient data collection was begun on 2 inpatient units. Candidates were screened for inclusion/exclusion criteria by reviewing the daily nursing report and communicating with the unit nurses. Potential participants and parents were approached between 4:00 PM and 7:00 PM on study days to explain the study and obtain permission and assent, if applicable. Six eligible study candidates were identified by room numbers each evening, Sunday through Thursday, for 6 weeks on the 2 units. The study protocol specified up to 6 weeks of data collection, with a maximum of 120 participants per unit. Philips Respironics loaned 12 Actiwatches for the study; however, only 6 could be in use at 1 time because of daily data downloads and recharging.
Data collection was performed through: 1) completing the demographic form; 2) placing an Actiwatch on the child's leg or arm (if <1 year of age) and gathering 1 night/day of data; 3) observing and documenting sleep interruptions on a standardized sleep diary completed by the child's nursing staff and/or from the electronic medical record; and 4) the appropriate survey (SinCH-SR or SinCH-YC) on the following day. The SinCH-P was also completed, if possible, for use in a future psychometric study.
The serial number of the Actiwatch and the patient's study identification number were recorded to match readings to the patient's diary. Between 4:00 PM and 7:00 PM the following evening, the investigators removed the Actiwatch and downloaded the data into actigraphy software on a password-protected computer. The investigators also collected the demographic surveys, sleep diaries, and SinCH surveys. Study personnel entered data from the electronic medical record regarding nursing functions/disruptions onto the sleep diary for those with incomplete data. All data were entered into Research Electronic Data Capture (version 8.5.2; REDCap, Nashville, Tennessee).
Sleep efficiency is defined as the number of minutes of sleep divided by the number of minutes in bed.15 Awakenings during the night were measured as 5 continuous minutes or greater.18 On SinCH surveys, children were asked which activities made it hard to sleep. Responses were dichotomized into “not at all/a little” versus “somewhat/a lot.” For what bothered participants at 3 time points, responses were dichotomized one category for “yes” and another for combined “no/don't know.” Continuous measures such as time slept were not normally distributed, so medians, interquartile ranges (IQRs), and nonparametric tests were used, specifically Wilcoxon signed rank, Wilcoxon rank sum, and Kruskal-Wallis. Spearman correlations were used to determine relationships between diary interruptions and Actiwatch data.
The loudest time on both units was 7:00 PM, the time of evening report. Talking was the predominant cause of noise greater than 65% of the time. The average decibel level for the 2 units was 56.62 and 59.49, respectively. Graphics of 1 week's sound summaries for 1 week are provided in Figure 2.
Ninety-five children had diaries, 85 had Actiwatch readings, and 66 completed SinCH surveys. The median age of the participants was 6.5 years (IQR, 2–12 years), with the full range being 1 month to 17 years. For the age ranges of 1 month to 4 years (group A), 5 to 12 years (group B), and 13 to 17 years (group C), the percentages of participants by age group were 39%, 38%, and 22.8%, respectively (Figure 1). Fifty-eight percent of the participants were male. The proportion of white/Caucasian participants was similar to the hospital's annual report (67.17% vs 65.32%), whereas black/African American, Hispanic/Latino, and multiracial participants varied. χ2 Analyses showed no significant statistical differences for race/ethnicity between the sample and the population of children admitted to the hospital (P = .089). Forty percent of child participants were in the hospital for the 1st time. Participants' diagnoses by percent were respiratory (27%), gastrointestinal (24%), infectious (24%), renal (15%), and other (10%).
More than 83% of children reported getting enough sleep at home, and 80% were normally good sleepers. There was a significant difference (n = 53) in the median number of awakenings between home and hospital sleep (1 [IQR, 0–2] vs 2 [IQR, 0–4], P < .0001) and time slept (9.25 hours [IQR, 8.2–10.3 hours] vs 8.5 hours [IQR, 6.8–9.3 hours], P = .0308). Sixty percent of subjects reported that while in the hospital they were awakened in the morning by another person, only 9.23% said that noise woke them. Almost 60% of the participants reported being “very” to “somewhat” sleepy during the day, and 52% stated they had taken a nap.
The median sleep efficiency was 82.13% (IQR, 73.5%–87.7%). The median for total sleep time was 7 hours 31 minutes (IQR, 6.2–9.3 hours). The median of the longest sleep episode was 5 hours 24 minutes (IQR, 3.4–6.8 hours), and the median number of awakenings over 5 minutes was 2 (IQR, 1–4). There were no significant differences by age group, gender, or diagnosis category (P > .5). Young children had a median of 3 (IQR, 1–5) awakenings per night compared with the other 2 age groups at 2 per night (IQR, 0–4 and 1–3, respectively).
Nursing care occurred in more than half of the hours (52%) between 8:00 PM and 8:00 AM. Vital signs and assessments were the most frequent nighttime interruptions (median of 3/night [IQR 2–4]). On a few occasions, children had VS performed 6 to 9 times/night and assessments 6 to 10 times/night. The median number of discrete interventions was 14/night, with a range of 5 to 37 times/night, although some of these were bundled care (eg, VS, assessment, and medication in the same hour). Almost 54% of children were weighed during the diary hours (10:00 PM to 10:00 AM). More than 31% of children had laboratory draws, and 48% had feedings during the night. Most children (85%) received medications, and 62% had their IV lines checked.
The number of interruptions was negatively correlated with the longest sleep episode (r = −0.28, P = .01). Correlations between sleep efficiency, sleep time, or awakenings over 5 minutes were not statistically significant. The mean for nighttime interruptions as measured by actigraphy was 2.7 (median was 2.0) and was 2.9 by the SinCH surveys (median, 3.0).
Noises disturbances are shown in Figure 2, and nonnoise types of disturbances in Figure 3. Other comments (mostly from parent respondents) included the following: screaming baby next door, starting feeds, getting transferred, cabinet squeak, dry air, nurse banging trash can, cleaning staff, pillow, hard bed, cost of hospital stay, IV machine clicking, breathing treatments, doctors/nurses coming in to talk, closing/opening door, new IV blood transfusion, generators, bedside report, and cleaning of the central IV line. Children reported bothersome nighttime events at specified time periods (Figure 4). No differences among age groups were found in the frequencies of these events.
The goal of this study was to determine typical hospital noise levels and the amount and quality of sleep received in a pediatric hospital. This was one of the 1st sleep studies that included children of all ages and diagnoses. The study used a wide array of measurement, such as Quietyme monitors, Actiwatches, and SinCH surveys for breadth and depth. There were no significant differences among gender, age, and diagnosis on sleep measures by actigraphy.
The study found that children were not getting an adequate amount of uninterrupted sleep in this hospital and confirmed findings of previous studies.1 The environmental sensors recorded that talking was the loudest noise; however, patients and families perceived alarms as the loudest. Normal conversation is typically 50 to 60 decibels. Noise could be affected by the difference in the unit layouts racetrack (Supplemental Digital Content 2, http://links.lww.com/JONA/A692) versus pod (Supplemental Digital Content 3, http://links.lww.com/JONA/A693). The recommended noise level for good sleep is 35 decibels or less.20 In the hospital, there were more factors to contend with than at home affecting children's sleep, such as worries, noises, and staff interruptions.
Comparing this study with other pediatric hospital sleep findings, this one was unique in the use of multiple measures. Actigraphy measurements of sleep efficiency showed a median of 82.13%, while Linder and Christian's15 were 89.58% to 92.72% over 3 nights for 15 school-age children. The incidence of room entries in this study was similar to Hinds et al, as their median was 11 and the current study's was 14 room entries/per night.13 In comparison to the studies with actigraphy, the researchers found a lower occurrence of awakenings (median, 2; n = 85) compared with Hinds et al13 at 14 (n = 29) and Linder and Christian15 at 10.5 to 13 (n = 15), who both used 3 nights of actigraphy.13,15 This study used 5 minutes of wake time to determine 1 episode of wakefulness, while the other authors used 1-minute epochs. Vital signs were the most frequent complaint, and the most common worry/discomfort was being homesick, as in the Meltzer and colleagues'12 study.
This study described sleep on only 2 medical/surgical units in 1 hospital in the Midwest over a period of 12 weeks. A longer investigation was not possible because of time constraints, funding, and staff burden. Participants' illness severity was not accounted, which may have predicted the number of sleep interruptions and influenced sleep measures. The SinCH survey has not been fully psychometrically evaluated, although this is underway.
Bias and inaccuracies are possible in all measurements, such as which parents/children opted not to complete the forms (eg, influenced by high stress or social desirability). Parents could interject their opinion of the child's sleep on the SinCH-YC. The study excluded non–English-speaking parents because of lack of research resources; although 95% of admissions were English speaking. Occasionally, nursing staff discouraged researchers from recruiting certain “problem” families. It is possible that such families would have a different sleep experience than those included in the study. A few Actiwatches did not record data, and some could have been removed, which would have recorded a sleep state when the child was awake. Actigraphy is more accurate when used for longer than 24 hours, particularly considering daytime naps. The Hawthorne effect could have predisposed nurses to enter the patient rooms less than normal or leave off notations of interruptions on diaries. Oftentimes diary recordings were not continued with the day shift after 7:00 AM, and it was not possible to discern if each activity listed in an hour time frame occurred during the same room entry.
Implications for Practice, Education, and Research
Using these findings, nurse administrators may influence change to improve sleep in pediatric hospitals. Researchers can perform multisite interventional studies to test possible environmental changes. Most research on pediatric sleep has been completed in medical/surgical units; therefore, studies in pediatric intensive cares should be conducted to determine how that unique environment impacts sleep. Educational interventions, such as teaching parents to help their children use diaphragmatic breathing, have been shown to help improve sleep in the hospital.21 Because noise was often the main cause of sleep interruptions, interventions to reduce noise are paramount. A 1-hour quiet time on an adult acute care unit demonstrated that patients felt more restful all day.6 Periods of enforced quiet, shut doors, and dimmed lights could also be an intervention on pediatric units. Acuity assessments, such as the Pediatric Early Warning Score, could be used to determine if unneeded interventions could be eliminated at night. Further studies could assess cost/benefit through evaluation of whether better sleep decreases length of stay or usage of pain medications. Improving patient sleep in the hospital is imperative to healing, and nurse administrators play the leading role.
The authors are grateful to Philips Respironics for loaning them Actiwatches and software to complete their study. This study would not have been possible without the support of Children's Mercy Hospital including Susan McElroy, director of Patient Care Services Research, and grant funding from Children's Mercy Hospital's Patient Care Services Research Department.
1. Bisogni S, Chiarini I, Giusti F, Ciofi D, Poggi GM, Festini F. Impact of hospitalization on the sleep patterns of newborns, infants, and toddlers admitted to a pediatric ward: a cross-sectional study. Minerva Pediatr
2. Cohen S, Doyle WJ, Alper CM, Janicki-Deverts D, Turner RB. Sleep habits and susceptibility to the common cold. Arch Intern Med
4. Chaput JP, Gray CE, Poitras VJ, et al. Systematic review of the relationships between sleep duration and health indicators in school-aged children and youth. Appl Physiol Nutr Metab
. 2016;41(6 Suppl 3):S266–S282.
5. Stickland A, Clayton E, Sankey R, Hill CM. A qualitative study of sleep quality in children and their resident parents when in hospital. Arch Dis Child
6. Applebaum D, Calo O, Neville K. Implementation of quiet time for noise reduction on a medical-surgical unit. J Nurs Adm
8. Centers for Medicare & Medicaid Services. Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS). 2018. http://www.hcahpsonline.org
. Accessed April 18, 2018.
9. Toomey SL, Elliott MN, Zaslavsky AM, et al. Variation in family experience of pediatric inpatient care as measured by Child HCAHPS. Pediatrics
10. Pilkington S. Causes and consequences of sleep deprivation in hospitalised patients. Nurs Stand
11. Herbert AR, de Lima J, Fitzgerald DA, Seton C, Waters KA, Collins JJ. Exploratory study of sleeping patterns in children admitted to hospital. J Paediatr Child Health
12. Meltzer LJ, Davis KF, Mindell JA. Patient and parent sleep in a children's hospital. Pediatr Nurs
13. Hinds PS, Hockenberry M, Rai SN, et al. Nocturnal awakenings, sleep environment interruptions, and fatigue in hospitalized children with cancer. Oncol Nurs Forum
14. Linder LA, Christian BJ. Characteristics of the nighttime hospital bedside care environment (sound, light, and temperature) for children with cancer. Cancer Nurs
15. Linder LA, Christian BJ. Nighttime sleep characteristics of hospitalized school-age children with cancer. J Spec Pediatr Nurs
17. Marino M, Li Y, Rueschman MN, et al. Measuring sleep: accuracy, sensitivity, and specificity of wrist actigraphy compared to polysomnography. Sleep
18. Meltzer LJ, Montgomery-Downs HE, Insana SP, Walsh CM. Use of actigraphy for assessment in pediatric sleep research. Sleep Med Rev
20. Oliveira L, Gomes C, Bacelar Nicolau L, Ferreira L, Ferreira R. Environment in pediatric wards: light, sound, and temperature. Sleep Med
21. Papaconstantinou EA, Hodnett E, Stremler R. A behavioral-educational intervention to promote pediatric sleep during hospitalization: a pilot randomized controlled trial. Behav Sleep Med