BACKGROUND: Pediatric sleep disturbances are regularly diagnosed on the basis of parental reports. However, the impact of parental sleeping problems on parental perceptions and reports of their child’s sleep has not yet been studied. We hypothesized that poor parental sleep decreases the parent-reported child sleep quality.
METHODS: A 1-week actigraph recording was performed in 100 children aged 2 to 6 years recruited in 16 day care centers. Their biological parents completed a sleep diary and a Sleep Disturbance Scale for Children (SDSC) on children’s sleep. The parents also completed the Jenkins’ sleep scale on their own sleep, the 12-item General Health Questionnaire, and questions on demographic factors. Linear regression analyses were performed to study the association of the parental Jenkins’ score on their child’s total SDSC score. Analyses were also performed for 3 of the subscales of the SDSC: disorders of excessive somnolence, disorders of initiating and maintaining sleep, and sleep-wake transition disorders.
RESULTS: Parental sleeping problems were associated with more frequent reporting of children’s sleeping problems. This association was unexplained by the actigraph measures of children’s sleep, such as actual 24-hour sleep time and sleep efficiency, parental mental health problems, or any other tested potential confounder or mediator. Similar correlations were seen for the 3 analyzed subscales.
CONCLUSIONS: Parental sleep quality was associated with overreporting of sleep problems in their children. This finding emphasizes the importance of considering parental sleep quality in the diagnosis, treatment, and research of pediatric sleeping problems.
- CI —
- confidence interval
- GHQ-12 —
- 12-item General Health Questionnaire
- SDSC —
- Sleep Disturbance Scale for Children
What’s Known on This Subject:
Previous studies have suggested that pediatric sleep problems impair parental sleep. However, many of these studies used parental reports instead of objective methods to analyze the child’s sleep, thus neglecting parental sleep-related perceptions as a source of bias.
What This Study Adds:
Parents who slept poorly overestimated the sleeping problems of their children when compared with objective sleep analyses. Parental sleeping problems seem to decrease parent-perceived child sleep quality, which verifies parental sleep as an influential factor in pediatric sleep disturbances.
Pediatric medicine is characterized by the use of second-hand information when inquiring about the patient’s medical history and current symptoms. Parental reports are a convenient source of information, but the use of second-hand data allows for the possibility of bias. For example, in comparison with the child’s self-report and the observations of the child’s teacher, depressive mothers report more behavioral problems in their offspring than do nondepressive mothers.1 This kind of distortion can become problematic in medical conditions in which the diagnoses rely heavily on the description given by the parents, such as pediatric sleep disturbances. Many of these disorders are diagnosed on the basis of parental reports and the use of objective methods is rare.2 Nevertheless, the effect of parental sleep quality and well-being on the accuracy of their reports on their child’s sleep has, to our knowledge, not been studied.
Consequently, the question arises whether parents who also suffer from sleep problems tend to overestimate their child’s sleeping difficulties, thus distorting the identification of the causalities involved in the family’s nightly struggles. The aim of the current study was to analyze whether parental sleep quality is associated with the parent-reported sleep quality measures in their children when the effect of the objective sleep quality of the children was taken into account. Whether poor parental sleep correlates with objective child sleep quality and whether parental psychiatric symptoms correlate with parent-reported child sleep quality were also analyzed. We hypothesized that parents with poor sleep quality would report more sleep-related difficulties in their children than suggested by an objective assessment.
Study Design and Sample
The FinAdo 2 study is an ongoing follow-up study that investigates the health and well-being of internationally adopted children <7 years old in Finland. The participants in this study were parents and their biological children (aged 2 to 6 years) who acted as the controls in a study on adoptee sleep. The study was approved by the Ethics Committee of the Hospital District of Southwest Finland. A total of 1560 information letters were distributed to the families in 16 day care centers in the cities of Turku and Kaarina, Finland. Parents of 117 children returned their informed consent to participate. Seven children who moved away from the area or could not be contacted and 2 adopted children were excluded from the study; thus, the remaining 108 children were enrolled. The evaluations were performed between January 2014 and February 2015.
The enrolled families were contacted and an appointment was made at which the parents were provided with questionnaires and instructed on the use of an actigraphy bracelet (GeneActiv Original; Activinsights Ltd, Kimbolton, United Kingdom), which was worn on the child’s nondominant hand. The actigraph event button was to be pressed when the child went to sleep and woke up, and the device was to be taken off before bathing and contact sports. Seven days later at the second appointment the questionnaires and actigraph were returned. The parents kept a sleep diary of their child’s sleep for the duration of the actigraphy recording. In the diary, they also recorded times when the actigraphy bracelet was removed and reasons for removal.
Both of the parents were also asked to individually complete questionnaires on their socioeconomic status and well-being. They also completed questionnaires on the child’s well-being and illnesses. If the parents completed the child’s questionnaires together, the mean scores of the parental scales were used in the analyses, except for education, for which the maternal response was always used. Lower maternal education has been shown to be associated with a child’s shorter time in bed, but this connection has not been identified for paternal education.3
An actigraph was used as the means of objective assessment of the child’s sleep.4,5 The device is a small wristwatch-like band attached to the examinee’s wrist, hip, or ankle. The method examines the amount of movement acceleration in time. It can estimate periods of sleep by using a threshold value for the lack of movement, because nonmoving periods of time represent sleep. However, it is unable to separate different stages of sleep, because it records only physical activity, unlike electroencephalography.
The movements recorded by the actigraph enable an estimation of sleep-onset time, wake-up time, sleep efficiency, sleep fragmentation, actual 24-hour sleep duration, and sleep-onset latency; the sleep-onset latency depicts a child’s resistance to falling asleep. The sleep efficiency index represents the time spent sleeping compared with the total time spent in bed and is influenced by both the resistance to falling asleep and the restlessness of sleep. The sleep fragmentation index shows the amount of physically active time during the whole sleep phase. It also includes all of the motionless sequences of <1 minute that are situated between 2 physically active periods. Thus, the sleep fragmentation index represents the restlessness of sleep.
These actigraphy variables can provide an estimate of the quality of the child’s sleep. Due to the restless nature of children’s sleep, however, the variables are not as reliable in the pediatric field as in adult studies6,7; however, the accuracy can be enhanced by using an appropriate algorithm.8
The GeneActiv actigraph has been proven to be reliable in examining sleep in adults9 and its validity for studying movement in children has also been documented.10 The device records and sums up gravity-subtracted acceleration vector magnitudes in 3 spatial directions by using a microelectromechanical sensor with a range of ±8 g.11 The measurement frequency used was 50 Hz with an epoch length of 1 minute. The threshold for interpreting the sum of the signals as sleep was 8 g per epoch. A research assistant analyzed the gathered actigraph data aided by the sleep diaries. In problematic cases, both the research assistant Lassi Sahlberg and an experienced clinical neurophysiologist (I.V.) analyzed the data together.
Parent-Reported Sleep Duration and Sleep Quality of Their Children
The children’s sleep as reported by the parents was assessed with the sleep diary and the Sleep Disturbance Scale for Children (SDSC). The SDSC is a 30-item questionnaire that evaluates, along with the total score, 6 different sleep domains. In addition to the total score, this study examined the score for disorders of excessive somnolence (5 items) and disorders of initiating and maintaining sleep (7 items), the latter considered to depict the bed resistance of the child. The score for sleep-wake disorders (6 items) was also studied because this score contains the items of bruxism, sleep talking, and other restless behavior during sleep and the act of falling asleep. Such behavior can disturb parental sleep by causing worry and interruptions. The SDSC was originally validated for school-aged children,12 but has since also been proven reliable in examining preschool-aged children.13 The Cronbach’s α for the total scale in the current sample was 0.76.
Parental Sleeping Problems
Parental sleep quality was evaluated with the Jenkins’ sleep scale. The scale consists of 4 questions, which are answered on a rating scale scored from 0 to 5. The scale has been proven reliable for examining the quality of sleep in adults14 and is widely used. The Cronbach’s α in the current sample was 0.78.
Parental Mental Health
The parents also completed the 12-item General Health Questionnaire (GHQ-12). The questionnaire is a valid and reliable measure to assess parental psychiatric morbidity, symptoms of anxiety, and depression.15 The scale uses a 12-item rating scale from 1 to 4. The second item on the questionnaire enquires about lost sleep due to worry. This item is not valid when screening for sleeping difficulties,16 nevertheless,to avoid overlap, the item was omitted in the analyses. The Cronbach’s α in the current sample, both with and without the sleep-related question, was 0.87.
The actigraph data were downloaded to the GeneActiv software, which converted the information for analysis by a trained research assistant to obtain the required information: (1) actual 24-hour sleep duration and (2) sleep efficiency. In step 1, the association between the total SDSC score and Jenkins’ score of the respondent was calculated by using linear regression analyses. In the subsequent steps, the impact of the analyzed potential confounders was investigated, in addition to the analyzed variables in the previous steps. Therefore, in step 2, the influence of the actigraphy measures (actual 24-hour sleep time and sleep efficiency) on the association between the total SDSC score and the respondent’s Jenkins’ score was investigated. The third step consisted of considering the influence of parental stress and psychiatric morbidity, represented by the GHQ-12 (with the sleep item removed), on this association. The child’s age and gender and the existence of long-term illnesses and medication as confounding factors were adjusted for in the fourth step. Finally, the model was adjusted for the time of year (summer or winter), number of siblings, parental marital status, parental age, and maternal education. The same procedures were then repeated for the subscale scores of disorders of excessive somnolence, disorders of initiating and maintaining sleep, and sleep-wake disorders. As additional sensitivity analyses, all analyses were repeated without the responses from those families whose parents completed the child’s questionnaires together. The statistical analyses were performed by using SAS for Windows version 9.3 (SAS Institute, Cary, NC).
The characteristics of the sample are shown in Table 1. Of the 108 children enrolled, 8 either declined to wear the actigraph or the recording could not be completed for other reasons. Of the parents, 18 completed the child’s questionnaires together.
The mean age of the participating children was 4 years, and the age and gender distribution was even. Thirteen percent (n = 13) of the children were reportedly taking long-term medication, 1 of whom (8%) was taking a systemic medication, thyroxin. However, the number of children taking medication at the time of the study was slightly higher, 18% (n = 18); of these children, 10 (56%) used a systemic medication, such as antibiotics, nonsteroidal antiinflammatory medicines, paracetamol, laxatives, or antihistamines. One child was using a first-generation antihistamine at the time of the study. None of the children used melatonin during the study or had been diagnosed with a sleep disturbance.
High levels of parental sleeping problems were associated with more frequently reported children’s sleeping problems (Table 2). This association was not attenuated after the model was adjusted for the actigraph measures of the children’s sleep, such as actual 24-hour sleep time or sleep efficiency. The association was also robust to adjustment for parental mental health problems or any other tested potential confounder or mediator. The association was unchanged after the removal of the 2 children with only 4 nights of actigraph data from the analyses (data not shown).
The parental Jenkins’ score was not associated with the objective measures of the child’s sleep. The association between the parental Jenkins’ score and the actual 24-hour sleep time was not statistically significant (estimate: −0.000601; 95% confidence interval [CI]: −0.00168 to 0.000479; P = .27). It was also not significant for sleep efficiency (estimate: −0.184; 95% CI: −0.383 to 0.0152; P = .07).
Similar associations as those between parental sleeping problems and the total SDSC score were also found for the 3 subscales (Tables 3, 4, and 5): disorders of excessive somnolence, disorders of initiating and maintaining sleep, and sleep-wake transition disorders. The only significant background variable in the final model for the subscale for disorders of initiating and maintaining sleep was parental age. The number of the disorders in question seemed to decline as parental age increased. However, this finding did not decrease the correlation between the parental Jenkins’ score and the SDSC subscale score.
The parental GHQ-12 score, with the sleep-related item omitted, correlated with the child’s SDSC score (estimate: 0.363; 95% CI: 0.533 to 0.673; P = .022). This association was also identified for the subscale of disorders of initiating and maintaining sleep (estimate: 0.171; 95% CI: 0.0261 to 0.316). The association was absent for the analysis of the 2 remaining subscales. The estimate of the subscore for the disorders of excessive somnolence was 0.0271 (95% CI: −0.0627 to 0.117; P = .55); and the subscore estimate for the sleep-wake transition disorder was 0.0511 (95% CI: −0.0511 to 0.153; P = .32). Nevertheless, the parental GHQ-12 score, with the sleep-related item excluded, did not diminish the association between the parental Jenkins’ score and the total, or the subscale, score of the SDSC as shown in Tables 2 and 4. The GHQ-12–related associations disappeared when the parental Jenkins’ score was also considered in the analyses.
All of the analyses were also repeated without the responses from the families whose parents completed the child’s questionnaires together. In these analyses, the associations between the parental Jenkins’ score and the total SDSC score and the parental Jenkins’ score and the subscale of disorders of initiating and maintaining sleep were, in fact, slightly stronger (estimate [95% CI]: 1.173 [0.712 to 1.633] and 0.461 [0.260 to 0.656], respectively; P < .001 for both) than in the original analyses. The other associations remained unchanged.
Consistent with the hypothesis, the current study showed that parents who reported sleeping difficulties of their own also reported more sleep-related problems in their offspring. They experienced their children as having more difficulties of initiating and maintaining sleep, more sleep-wake transition difficulties, and more excessive somnolence than did parents who slept better. This association could not be explained by the children’s sleeping problems as identified with an objective measure, namely the actigraph. Child’s age, gender, number of siblings, existence of chronic illnesses or medication, or the existence of current medication also did not attenuate the association. Furthermore, this finding was not altered by parental psychiatric symptoms, investigated with the GHQ-12, parental education, marital status, socioeconomic status, or time of year. Parental sleeping problems did not correlate with the objective child sleep variables.
A child’s sleeping difficulties have been shown to be associated with parental well-being; Boergers et al17 observed that parents of children with >1 sleep disorder reported more daytime tiredness than did parents of children with only 1 sleep problem. Other studies have also proposed that a child’s sleeping problems impair the sleep of the parents.18–20 On the other hand, parental depressive symptoms correlate with actigraph-measured sleeping problems in the child,21 and studies also showed that marital aggression predicts objective pediatric sleep difficulties.22 Thus, the connection between parental well-being and pediatric insomnia may be reciprocal.
It is worth noting that many of these studies used parent-response questionnaires instead of objective methods to examine the sleep of the child. Such surveys are a viable method, especially when examining large samples, but can lead to unconscious bias from the respondent. Therefore, the direction of the association between the child’s reported sleeping problems and the impaired sleep of the parent is inconclusive.
The current findings are in line with previous studies, which have shown that sleep loss causes an attention bias toward negative events and enhances memory building of negative details.23,24 Sleeping difficulties have also been shown to cause an attention shift toward sleep-related stimuli.25 These observations explain why parents who sleep poorly notice and remember their children’s bed resistance and night waking so well. Hence, in the clinical field, it is of paramount importance to aim interventions in a child’s sleeping difficulties toward the well-being of the whole family instead of only the child. The diagnostic measures should consider not only the sleeping habits of the child but also parental aspects. Tired parents can unconsciously exaggerate their child’s sleeping difficulties, which could lead to misplaced interventions.
The strengths of this study include the utilization of both objective and subjective measures. For 98% of the children, the actigraphy lasted 5 days or longer, which, according to previous research, is a reliable duration of analysis.5 Despite the recruitment in day care centers and the low recruitment rate, the parental employment rate corresponded with the unemployment rate in Finland at the time of the study26; 7 mothers and 7 fathers reported being currently unemployed or retired. In addition, the families who participated in the study well represented the average population compared with a written communication from the Official Statistics of Finland: In 2014, the mean age of a mother of a 4-year-old child in Finland was 34 years and that of a father was 37 years, which corresponds to the ages in the study population. On average, a 4-year-old had 1.5 siblings <18 year old.
However, the limitations of the study need to be discussed. Although the families were recruited in a nonclinical setting, the fact that the sample included mostly white, highly educated families can hinder the generalization of the results.
Sleeping in the parents’ bedroom was relatively common for the children in the study population (26%). In a large multicultural study, <10% of predominantly white (Australia, Canada, New Zealand, United Kingdom, and the United States) children aged 24 to 36 months slept in the same room as their parents.27
Moreover, the age range of the participating children was fairly wide due to the nature of the adoption study, in which the children of the current study acted as controls. In addition, in Finland, children start school at 7 years old, which influenced the choice of the inclusion criteria.
The GeneActiv actigraph has not been validated for pediatric sleep studies. However, it has been validated in a pediatric population10 and its accelerometer has been shown to be as reliable as the accelerometer in the ActiGraph (ActiGraph, Pensacola, FL),28 which is widely used in studies on children’s sleep. Furthermore, the study did not aim to diagnose sleep disturbances but to compare individuals within a group.29
Our study shows that parents who sleep poorly report more sleeping difficulties in their children than do parents who sleep well. This association could not be explained by an objective measure of the child’s sleep. It seems that tired parents present an observation bias with regard to their child’s sleep compared with parents not suffering from sleeping problems. Therefore, in the field of pediatric sleep disorders, future diagnostic methods and treatments should take into consideration not only the examined child but also the whole family, including the parents. Furthermore, the results presented highlight the importance for future pediatric sleep research to take into account parental sleep quality as a confounding factor and to examine whether poor parental sleep also affects the sleep of the child.
We thank the research assistants Anniina Mäkinen, Miia Salokannel, and Lassi Sahlberg for their work in executing the actigraph recordings and the Clinical Neurophysiology Department at Turku University Hospital for supplying the facilities for the actigraph recordings. We also thank the FinAdo research group for their collaboration in the study.
- Accepted January 15, 2016.
- Address correspondence to Helena Lapinleimu, MD, PhD, Department of Pediatrics, University Hospital of Turku, Kiinamyllynkatu 4-8, 20521 Turku, Finland. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: All phases of this study were supported by the Foundation for Pediatric Research, Finland, the EVO (State funding for university-level health research) grant from Turku University Hospital, and the Signe and Ane Gyllenberg Foundation. Dr Rönnlund was supported by a grant from the Foundation of Turku University Central Hospital. Dr Elovainio was funded by the Academy of Finland (265977). Dr Lapinleimu received the EVO grant from Turku University Hospital.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- American Psychiatric Association
- ↵GeneActiv instruction manual. Version 1.2. 2012. Available at: www.geneactiv.org/wp-content/uploads/2014/03/geneactiv_instruction_manual_v1.2.pdf. Accessed November 3, 2015
- Martin J,
- Hiscock H,
- Hardy P,
- Davey B,
- Wake M
- Official Statistics of Finland
- Copyright © 2016 by the American Academy of Pediatrics