Article Text
PDF
Abstract
Objectives Family caregivers of children who depend on medical technology (CMT) provide highly skilled care up to 24 hours per day. Sleep disruption places family caregivers at risk for poor health and related outcomes that threaten their long-term caregiving capacity. Few studies exist that have measured sleep in family caregivers, and most have relied entirely on subjective measures.
Methods In a prospective cohort study, family caregivers of CMT (n=42) and caregivers of healthy children (n=43) were recruited. Actigraphy data and a concurrent sleep diary were collected for 6 days/7 nights. Measures of sleep quality, depression, sleepiness, fatigue and quality of life were also administered.
Results Family caregivers of CMT averaged fewer hours of sleep per night (mean (SD)) (6.56 ± 1.4 vs 7.21 ± 0.6, p=0.02) of poorer quality (7.75 ± 2.9 vs 5.45 ± 2.8, p<0.01) than the control group. Three times as many family caregivers of CMT scored in the range for significant depressive symptomatology (12(33%) vs 4(10%), p=0.01) and experienced excessive daytime sleepiness (16(38%) vs 5(12%), p<0.01). Fatigue was also more problematic among family caregivers of CMT (22.12 ± 9.1 vs 17.44 ± 9.0, p=0.02).
Conclusions Family caregivers of CMT are at risk of acute and chronic sleep deprivation, psychological distress and impaired daytime function that may threaten their capacity for sustained caregiving. Family caregivers of CMT may be important targets for screening for sleep disorders and the development of novel sleep-promoting interventions.
- caregivers
- children with medical complexity
- home care
- sleep deprivation
- depression
Statistics from Altmetric.com
What is already known on this topic?
Evidence suggests that sleep disturbance is a ubiquitous complaint among family caregivers of children who depend on medical technology.
Sleep is a complex biobehavioural process and decrements in sleep can have negative influences on physical and mental health, daytime function and safety of family caregivers.
Existing studies of sleep in this diverse population are limited by reliance on self-report, homogeneous samples and failure to control for confounding variables.
What this study adds?
Describes that sleep disturbance is associated with sleepiness, fatigue and depression among family caregivers.
Findings indicate that an integrated approach to caregiver health with development of sleep-promoting interventions seems likely to help these families.
Introduction
The population of children who depend on the use of medical technology (CMT) is growing worldwide.1 2 They are a diverse group with various diagnoses, uses of technology and a high utilisation of healthcare resources.3 Their family caregivers provide vigilant and skilful care every day. Even with homecare supports, sleep disturbance is a ubiquitous complaint among family caregivers of CMT,4 threatening their health, daytime function and capacity for quality care giving at home.5–7
Of the few studies that have examined sleep in this population, most have relied on subjective measures,4 and are limited by their reliance on self-report, diaries focused on single diseases or types of technology and failure to control for key confounders including child’s age. Objective characterisation of sleep in a diverse sample of family caregivers of CMT is necessary for the development of sleep-promoting interventions. The aim of this study was to compare sleep in family caregivers of CMT with that of family caregivers of healthy, same-aged children.
Methods
Setting
Family caregivers of CMT were recruited from March 2013 to September 2013 from tertiary-care clinics at the Hospital for Sick Children in Toronto, Canada. Control subjects were concurrently recruited from primary care clinics in the Toronto area.
Subjects
Eighty-five subjects were included. Eligible subjects were self-identified primary caregivers of a child aged 12 months–18 years (one per household) who lived at home. Included in the CMT group were family caregivers of children who used medical technologies at night due to a complex chronic condition. Control subjects were eligible if they had children with no known developmental or chronic health conditions. We excluded from both groups any caregiver with a diagnosed sleep disorder or who did not speak or read English, and those with children under 12 months of age in the home.
Outcome measures
Sleep: to allow for estimates of wake and sleep, Octagonal Basic Motionlogger (60 g, Ambulatory Monitoring, Ardsley, New York, USA) actigraphs were set for 6 days and 7 nights to capture activity data at 1 min epochs and the Cole-Kripke scoring algorithm (zero crossing mode) aided in analysis.8 9 Congruence between actigraphy and polysomnography (not feasible in this study) indicates 88% agreement between the two methodologies.10 A sleep diary was used according to sleep medicine and research best practices.11 The diary collected details of the unique contributors to sleep in this population. Substitutions of sleep times were made from the sleep diary to complement actigraphy data and calculate sleep variables if a participant had ≥4 and <7 complete nights of actigraphy. Subjective sleep quality was measured using the Pittsburgh Sleep Quality Index (PSQI), a validated self-rated questionnaire comprising 19 items.12 The measure distinguishes between good and poor sleepers with a higher score more indicative of sleep problems (>5 cut-off).12
Health-related quality of life (HRQoL): the SF-12 Health Survey (V.1; short form-12, 12-item)13 was used to measure overall physical and mental health outcomes. Physical component summary and mental component summary were standardised, with scores above and below 50 representing above and below average, respectively. Higher scores indicate a better HRQoL.
Depressive symptomatology: the Centre for Epidemiologic Studies Depression Scale (CES-D Scale; 20-item) measured the presence of depressive symptomatology.14 A score ≥16 has been established as the cut-off suggestive of psychological distress, while >21 suggests clinically meaningful depression.14
Sleepiness and fatigue: the Epworth Sleepiness Scale (8-item) rates the likelihood of falling asleep in specific situations. Scores >10 indicate clinically significant sleepiness.15 Fatigue severity, distress, degree of interference in activities of daily living and timing was measured using the previously validated multidimensional assessment of fatigue (16-item).16 Scores for the test ranged from 1 to 50 (non-severe fatigue).
Procedures
On day 1, study participants were enrolled, sociodemographics and clinical data collected and family caregiver sleep measurement began via actigraphy and concurrent sleep diary. The study was completed with the administration of standardised questionnaires at a home visit on day 7 (see figure 1). Study recruitment occurred 1 day per week in the hospital, coinciding with set clinic days for CMT and in the community on the days having peak clinic volumes. Stratified sampling by child’s age (in years) enabled control of this confounder on family caregiver’s sleep.
Schema for cohort study recruitment. CMT, children who depend on medical technology; FC, family caregivers.
Sample size considerations and statistical analysis
At a significance level of 5%, 23 subjects per group were needed to provide at least 80% power to detect a clinically important difference (a mean difference in total sleep time-nocturnal (TST-N) of 1 (SD=1.2) hour; suggested by Meltzer et al. 17 The sample was increased from 23 to 40 in each cohort to account for the potential influence of under-reported variation in this previously reported TST-N and allow for a 20% rate of missing data or loss to follow-up.
All statistical analyses were completed using SPSS V.22 (Armonk, New York, USA) software. A two-sided significance level of 0.05 was set a priori. For sample characteristics, descriptive statistics were reported as means (SD) for continuous data or counts (percentages) for categorical data. Between-group differences were examined using two-sided independent samples t-tests for continuous variables and χ2 tests for categorical variables (or their respective non-parametric tests as indicated). Fisher’s exact test was used when cell frequencies were under 5. As a measure of night-to-night variability, the average number of consecutive nights per participant that had ≥30% difference in TST-N was calculated. Univariate linear regression screened the relationship between sociodemographic variables known from the literature to influence TST-N. Multiple linear regression analysis determined the most salient subset of predictors.
Institutional Review Board approval was obtained from the Hospital for Sick Children and The University of Toronto (IRB#: 1000034944).
Results
Forty-two family caregivers of CMT (83.3% mothers, n=35) and 43 control subjects (88.4% mothers, n=38) were recruited, with 40 family caregivers per cohort having complete actigraphy data available (figure 1). Missing actigraphy data were few, with sleep-wake diary substitutions made for <2% of all nights and only when at least of three of seven nights of actigraphy were available. The recruitment rates among eligible family caregivers of CMT and control subjects were 84% and 76%, respectively. Subjects differed along the lines of socioeconomic status. For example, family caregivers of CMT were more likely to be single, to have completed less education, to be employed less than full-time and to have a lower household income than controls (table 1).
Baseline demographic characteristics of family caregivers
The family caregivers of CMT had children with varied diagnoses who depended on the use of at least one, and half on more than one (50%, n=21), type of medical technology at night (table 2). An enteral feeding device was most commonly used (70%, n=29), often with the use of another technology. More than 70% of the subjects’ children used respiratory technologies. More than half (54.8%) of the subjects used oxygen saturation monitors as an adjunct to caring for their children (table 2). Most the children had caregiver-reported neurocognitive delay or impairment (61.9%).
Clinical characteristics of children dependent on medical technology
Over a third of the subjects (67.0%) used homecare-nursing services with a wide range in the number of hours/week (0–105 hours). Use of homecare-nursing service at nighttime averaged 18.6±22.6 hours/week.
Sleep
Between-group differences were found on all measured sleep outcomes (p<0.02 for all outcomes; table 3). For the primary outcome, family caregivers of CMT achieved on average 40 min less sleep per night than controls. The difference between the cohorts in the number of consecutive nights±30% TST-N was threefold (1.33±1.5 vs 0.45±0.7, p=0.001) and family caregivers of CMT averaged more than three times as many sleep-deprived (<6 hours) nights (83 nights) during the study period compared with the control group (24 nights) (p<0.001). Nocturnal awakenings were more common among family caregivers of CMT compared with the control group both when measured with actigraphy (8.00±3.9 vs 6.01±3.7, p=0.014) and when self-reported (2.35±2.0 vs 1.42±1.5, p=0.005). Family caregivers of CMT also slept more than twice as long from 09:00 to 21:00 hours (21.57±24.0 vs 9.14±9.9 min, p=0.004).
Between-group differences on sleep and related outcomes
Sleep quality was poorer among family caregivers of CMT compared with the control group (PSQI; 7.75±2.93 vs 5.45±2.8, p=0.001) with twice as many family caregivers of CMT classified as ‘poor sleepers’ (73.8% vs 39.5%). Self-reported sleep-onset latency also differed, with twice as many family caregivers of CMT documenting that it took >30 min to fall asleep (χ2 = 15.36, p=0.001).
Very few caregiver factors linked to sociodemographic status were independently associated with poorer sleep outcomes. In family caregivers of CMT, being born outside Canada was predictive of shorter TST-N (unstandardised coefficient 79.69, SE 22.93, p=0.001), and lower levels of education (high school vs college/university; B 2.88, SE 1.43, p=0.052) and household income (<US$25 000 vs ≥US$100 000; B 4.29, SE 1.78, p=0.021) were associated with more sleep fragmentation.
Depression, daytime sleepiness, fatigue and HRQoL
Family caregivers of CMT had higher global depression scores (11.70±8.7 vs 6.95±6.0, p=0.008) with more than three times as many scoring in the range for clinically meaningful depression (33.3% vs 9.3%) (table 3). Three times as many family caregivers of CMT scored in the range for excessive daytime sleepiness (38.1% vs 11.6%, p<0.001), and scored higher on the measure of fatigue (22.12±9.1 vs 17.44±9.0, p=0.021). There were no HRQoL differences found on the physical (p=0.647) or the mental functioning (p=0.072) component scores.
Factors that predict sleep
A generalised linear model was used to assess the influence of sociodemographic factors on TST-N among family caregivers. On univariate analysis, only caregiver group, age, gender, place of birth and education met the screening criterion (p<0.25) for inclusion into the multivariate analysis. When these variables entered a multiple regression model, caregiver group, gender and place of birth were significant predictors at the 5% level. Age was significant at the 0.09 level and was left in the final model because of its potential clinical significance. The percentage of variance explained was 31% for the final model.
Discussion
Using objective measurement, this study has confirmed that family caregivers of CMT do not achieve adequate sleep and are at risk for negative outcomes associated with acute and chronic sleep deprivation. Findings also highlight that depression, excessive daytime sleepiness and fatigue are prevalent in this population. Additionally, these outcomes are worse when family caregivers assume responsibility for a child with extraordinary care needs, compared with those with healthy same-aged children, even when homecare-nursing services are available. Group differences included select sociodemographic variables that predicted sleep outcomes (in the expected direction).
Family caregivers of CMT achieved, less sleep than the recommended 7–9 hours per night for middle-aged adults.18 In earlier studies, sleep disturbance has predicted obesity, diabetes, hypertension, cardiovascular disease19 20 and mood disturbances,21 and in health professionals, impaired neurocognitive and on-the-job performance.22 23 Although the 40 min difference we detected was smaller than our a priori estimate, it was statistically significant and clinically important, exceeding the 30 min improvement in TST-N considered a minimal clinically important difference in intervention studies among adults with insomnia.24 Moreover, family caregivers of CMT achieved less sleep per night than reported in population-based studies of adults with and without children.25
Wide individual variability in TST-N for family caregivers of CMT between nights may indicate problems with disorganised circadian rhythms26 and poor sleep hygiene.27 In earlier studies, variability in sleep duration among family caregivers of critically ill children has been associated with daytime sleepiness and fatigue.28 Also documented are instability in sleep times among family caregivers of ventilator-assisted children living at home.17 It is widely recognised that unstable sleep patterns are firmly linked to poor health and safety concerns.29
The rate of poor sleep quality was consistent with those reported among earlier caregiver studies with more homogeneous samples.17 30 31 Even brief arousals can negatively influence sleep quality32 and as found in this study, subjective recall of awakenings can be unreliable. Family caregivers may not recognise the effects of fragmented sleep and subsequent decrements in functioning and performance of caregiving responsibilities.33
Our study suggests that family caregivers of CMT may be at risk for the negative effects of depression.34 35 Similarly elevated scores on the CES-D have been reported among family caregivers of ventilator-assisted children.17 Longitudinal data suggest that the relationship between sleep and depression is likely bidirectional34 36 thus if left untreated, each has the potential to influence the other to become debilitative.
Moreover, this study has documented excessive daytime sleepiness which further raises concerns for the safety of family caregivers of CMT and the safety of their children who depend on them to remain vigilant. Elevated levels of fatigue among family caregivers of CMT may also interfere with caregivers’ motivation and ability to care for their entire family unit.
Our findings of no significant difference in HRQoL differ from other studies, perhaps because both groups reported low ratings slightly below the expected population mean (50 (10)).13 Thus, while there are descriptions of positive outcomes of caregiving5 commonly used measures of HRQoL have not appeared to be consistently sensitive to presence of sleep disturbance.
Lower socioeconomic status was found among the family caregivers of CMT. Although not found to be predictors, between-group differences in education and income may be attributed to increased costs, relative underemployment and heightened stress among family caregivers of CMT.6 Caregiver place of birth was a predictor of shorter TST-N. Population-based data suggest that people who have immigrated experience conditions that include working longer hours, for less money, which may negatively affect how much time they have for sleep.37
Strengths and limitations
The methods used in this study addressed the limitations in previous studies. Internal validity of this study has been bolstered through stratified sampling of a comparator group to account for the confounder of child’s age. Furthermore, a diverse sample of caregivers of CMT and use of an objective sleep measure are also strengths. Differences observed between the cohorts on sociodemographic variables were addressed through examination of these as potential confounders. Finally, the nature of the study methods prohibits any conclusions about causality.
Recommendations for practice, policy and research
Clinicians are encouraged to identify ways to address the sleep and respite needs and better health and related outcomes for family caregivers of CMT. The inclusion of screening questions for sleep disturbance in caregiver assessments is recommended. Moreover, it is timely to consider if actigraphy, gaining acceptance as a clinical tool38 may be used to assess the appropriate amount of night nursing to achieve improved sleep and related outcomes in this group. The provision of high-quality information about sleep and sleep hygiene, such as that made available to the public by the National Sleep Foundation (https://sleepfoundation.org/) is recommended for family caregivers.
Studies with strong family engagement are needed to test the feasibility of sleep-promoting interventions among family caregivers of CMT. Future research using psychomotor performance tests to characterise improvements or decrements in daytime function may be informative. Longitudinal data are required to better define key transitions when sleep disturbances may be heightened, and importantly, identify what factors across the life-span and among marginalised groups, including those foreign born, contribute to sleep disturbance and related outcomes in family caregivers of CMT.
Conclusions
This study fills a gap by describing what is known about the sleep and related outcomes in family caregivers of CMT, who provide the most complex care of all community-dwelling individuals.2 It is the first study to report on objectively measured sleep in a diverse sample of family caregivers of CMT.
These findings underpin concerns for the negative relationship between caregiving, sleep, mental health and daytime function in a growing paediatric population. Sleep disturbance is among many of the challenges encountered in caring for CMT and their families, and there is need for quality evidence to improve its understanding and inform targeted interventions and integrated health policy.
Acknowledgments
The authors would like to thank the primary care teams at Village Park Paediatrics and Maple Kidz Clinic and the Hospital for Sick Children’s Long Term Ventilation and Complex Care Teams. The authors would like to thank project assistants Michelle Ho, Lyndsey McRae, Kim Nelson and Nicole Sidhu for their work. The authors would also like to thank Stephanie Gee and Tina Yan for analysis and manuscript support, respectively.
References
Footnotes
Contributors KK conceptualised and designed the study, drafted the initial manuscript, reviewed and revised the manuscript and approved the final manuscript as submitted. EC and KS conceptualised and designed the study, reviewed and revised the manuscript and approved the final manuscript as submitted. RS conceptualised and designed the study, reviewed and revised the initial and subsequent drafts of manuscript and approved the final manuscript as submitted. EP advised on the study design and data analysis, reviewed and revised the manuscript and approved the final manuscript as submitted.
Funding Krista Keilty was supported by The Canadian Lung Association; The Ontario Lung Association; The Janis Rotman Fellowship for Innovation in Paediatric Homecare, SickKids Foundation; Canadian Institutes for Health Research- Team Grant: Better Nights and Better Days and -Team Grant: Sleep and Circadian Rhythms. Eyal Cohen was supported by grant funding from the Canadian Institutes for Health Research (FDN-143315). Robyn Stremler was supported by a Canadian Institutes of Health Research New Investigator Award and an Ontario Ministry of Research and Innovation Early Researcher Award.
Competing interests None declared.
Ethics approval Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Editorial