Objective. Sleep disorders can cause substantial morbidity but often remain undiagnosed among adults. We identified a series of children with sleep-related symptoms and reviewed medical chart notes for the previous 2 years to determine how often sleep problems had been addressed.
Setting. Two university-affiliated but community-based general pediatrics clinics.
Patients. Children, ages 2.0 to 13.9 years, with clinic appointments.
Measures. Parental and child responses to a validated Pediatric Sleep Questionnaire (PSQ) were used to identify patients at risk for chronic sleep-disordered breathing, periodic leg movements during sleep, insomnia, or excessive daytime sleepiness. Chart notes written within the previous 2 years were searched for sleep-related symptoms, diagnoses, or treatments.
Results. A total of 830 questionnaires were completed; 1395 chart notes of 86 symptomatic participants (mean age: 6.6 ± 3.1 years; 51% male) with 103 identified sleep problems were reviewed. Fewer than 15% of patients had current chart notes that mentioned any of the PSQ-defined sleep problems; diagnoses were mentioned for 2 of 86 patients and no treatments were discussed. Among the 103 sleep problems, only 16 received mention in any past or current note; 10 had led to a diagnosis; 4 had led to intervention; and 3 were treated in a manner likely to be effective. Seventy-four of the sleep problems (72%) occurred in children whose charts did mention something about sleep, but such notations rarely related to concerns uncovered by the PSQ.
Conclusions. Children with PSQ-identified sleep problems at 2 general pediatrics clinics seldom had these problems addressed, diagnosed, or treated, despite discussions about some aspect of their sleep in the large majority of cases. These findings support expansion of clinician and parent education about sleep disorders in children.
- sleep disorders
- obstructive sleep apnea
- sleep-disordered breathing
- disorders of excessive somnolence
- Pediatric Sleep Questionnaire
Epidemiologic and school-based studies suggest that sleep disorders are common among children. Difficulty with sleep onset or sleep maintenance occurs in ∼10% to 20% of 8- to 9-year-old children,1 sleep-disordered breathing (SDB) occurs in ∼1% to 3% of school-aged children,2,,3 and excessive daytime sleepiness (EDS) seems to cause significant problems in at least 10% of school children.4 Among children seen in general pediatric practices, the frequency of sleep disorders is likely to exceed that seen in the community, perhaps because respiratory, behavioral, and other disorders that commonly prompt pediatric visits are often associated with sleep disorders. For example, one sleep disorder, periodic leg movements during sleep (PLMS), is believed to be rare among all children but quite frequent among those with attention-deficit/hyperactivity disorder, a common reason for visits to a pediatrician.5,,6
Pediatric sleep disorders cause significant consequences that remain incompletely defined but include effects on growth, cardiovascular health, cognitive function, and daytime behavior.7–9Studies suggest that disruptive behavior disorders, such as attention-deficit/hyperactivity disorder, can sometimes be caused by undiagnosed sleep disorders, treatment for which improves the behavior.10,,11 Academic grades of students at a wide range of ages are reduced by unrecognized sleep disorders and insufficient sleep.12–14
Although the impact of occult sleep disorders has become increasingly clear, little work has been done to determine how effectively pediatricians identify and treat sleep disorders in children with suggestive symptoms. Chart reviews of adult patients in 1989 and 1990 suggested that the large majority of sleep complaints remained unaddressed and fewer still were effectively diagnosed or treated.15 Since that time, sleep disorders have received much publicity and identification of them may have improved, but information relevant to children is still lacking. We therefore identified a series of children with sleep-related symptoms in 2 general pediatric clinics and then reviewed medical records to determine whether those symptoms had been addressed, diagnosed, and treated within the previous 2 years.
Parents who accompanied their children to appointments at either of 2 general pediatrics clinics were recruited to sign an informed consent for this protocol, which was a component of a larger institutional review board-approved study of sleep and behavior in children. Children older than 9 years of age or old enough to understand the study were asked to sign an informed assent. Criteria for enrollment included: 1) age between 2.00 and 13.99 years, 2) parental ability to read and write, 3) absence of mental or physical impairment severe enough to preclude interpretation of behavioral information collected, and 4) presence of symptoms of SDB, PLMS, insomnia (INS), or EDS, as defined below. Participants were recruited between February 6, 1998 and July 20, 1999 by research assistants who attempted to approach all child–parent pairs at the clinics on designated days.
Both clinics are owned by the same large public university health system. Clinic A is staffed by 5 full-time faculty who finished residency training 3 to 15 years before the beginning of this study. Clinic B was recently bought by the university, without substantial change in that site's practicing physician staff, which includes 4 pediatricians who finished training 3 to 35 years earlier. The clinic population at both sites is primarily community-based. The university owns a large health care maintenance organization, which accounts for ∼35% of all visits to clinic A and 30% of all visits to clinic B. Residents see ∼28% of the patients at clinic A and ∼20% of patients at clinic B, but always with attending physicians. Notes from all such patient visits are signed by both residents and attending physicians.
Parents waiting for their children's appointments completed the Pediatric Sleep Questionnaire (PSQ), which we had previously developed and validated.16,,17 Children's help in completing the PSQ was encouraged. The PSQ contains ∼70 closed question items that request yes, no, or don't know responses. Instructions ask that questions be answered with reference to behavior in general rather than recent days, which may have been unusual if the child had been acutely ill. Several additional open-ended items ask about past medical history, medications, and family history. Question items for symptoms of SDB provide a validated screen for polysomnographically defined obstructive sleep apnea and upper airway resistance syndrome.17 Items about EDS also have been validated among children with and without these 2 diagnoses.17Polysomnographic data have not been used to validate the question items about INS; assessment of INS is generally based on the history rather than on polysomnography or other recordings. Finally, PLMS can be described to parents and observed by some of them.5,,18Although the absence of such an observation in no way excludes the diagnosis, the relatively rare parental affirmation that PLMS are present, based on a precise description of the movements, is likely to carry reasonable specificity and positive predictive value.
Table 1 lists the combinations of PSQ items used to identify children with symptoms of SDB, PLMS, INS, and EDS. For SDB and PLMS, the aim was to identify children who clearly merited inquiry into the possibility that these disorders might be present. For INS and EDS, the aim was to identify children whose symptoms were severe. In this study, identification of a group of participants at high risk for sleep disorders was considered more important than detection of every potentially affected individual.
Investigators contacted parents of symptomatic children by telephone to obtain permission to review medical records. All dictated, handwritten, and telephone notes written between the date of questionnaire completion and 2 years before were reviewed by one investigator (K.M.H.). Notation was made whenever a clinician mentioned: 1) a symptom related to the symptom-complex identified by the PSQ, 2) a diagnosis that might explain the sleep-related symptoms, 3) a treatment designed to address those symptoms, 4) a treatment likely to address those symptoms effectively, and 5) any explicitly sleep-related problem. Table 2 lists specific criteria by which patients' notes were judged to contain each of these 5 elements. In contrast to PSQ-defined case ascertainment, chart review criteria were purposefully liberal to give the benefit of the doubt whenever possible.
Investigators also extracted, from clinic notes, lists of current or ongoing medical problems for each child. These were then coded into the following categories: respiratory (in no cases SDB); cardiovascular; developmental, psychiatric, or behavioral; neurologic; infectious disease; trauma; obesity; and no ongoing medical problems.
Within each group of patients affected by a given symptom-complex, we calculated the frequencies with which current or past medical notes mentioned the relevant symptom, diagnosis, treatment, effective treatment, or any sleep-related issue. Age and gender were tested, within each symptom-complex group, for any increased likelihood that current or past chart notes would mention relevant sleep symptoms. The presence of current or ongoing medical problems was similarly tested when at least 5 children within the symptom-complex group had the medical problem. All analyses were performed with SAS, Version 6.12 (SAS Institute Inc, Cary, NC). The level of significance was set atP < .05.
Approximately 1150 children of appropriate ages came with their parents to appointments at either of the 2 clinics while the PSQ was being distributed; 830 questionnaires were completed; 100 participants were symptomatic for SDB, PLMS, INS, or EDS; and 86 children had parents agreeable to review of located medical charts. The mean age of the 86 participants whose data form the basis for this report was 6.6 ± 3.1 years, and 44 were boys (51%). Numbers of participants, mean ages, and gender for each symptomatic group are given in Table 3. Some participants had >1 symptom-complex; a total of 103 symptom-complexes were identified among the 86 children. Children's current or ongoing medical problems were classified as respiratory (n = 29); cardiovascular (n = 2); developmental, psychiatric, or behavioral (n = 9); neurologic (n = 7, all with headache); infectious disease (n = 4); trauma (n = 1); obesity (n = 2); or no medical problems (n = 40).
The total number of chart notes reviewed was 1395 for the 86 participants and between 2 and 85 for each participant. Fewer than 10% of patients had a chart note, from the day on which the PSQ was completed, that addressed symptoms relevant to PSQ-determined SDB, PLMS, or EDS (Table 4). Only 2 of 15 participants with INS had a current note that mentioned the problem. Even fewer of the children (only 2 of 86) had any diagnosis mentioned in a current note, and none had any treatment mentioned. However, 14% to 20% of the SDB, PLMS, and INS children had a current note that did mention something about their sleep; in the EDS group, no note mentioned sleep. Notes from the previous 2 years, in aggregate, sometimes mentioned symptoms relevant to PSQ-identified sleep problems—and then often contained diagnoses—except that in the PLMS group, no diagnoses were made for the symptoms. In all groups, treatments were rarely mentioned in past notes. The majority of patients with each symptom-complex had past notes that did mention some aspect of sleep but said nothing related to the PSQ-defined problem.
Content of current and past notes, considered together, resembled that of past notes alone (Table 4). Among the 103 sleep problems identified by the PSQ in the 86 children, 16 of the problems (16%) had corresponding symptoms mentioned in any note, 10 (10%) had diagnoses mentioned, 4 (4%) had treatments listed, and 3 (3%) had a treatment likely to be effective listed. Although 74 (72%) of the 103 problems were associated with at least 1 notation of sleep information during the past 2 years, most notations were completely unrelated to the sleep problem identified by the PSQ.
Age, Gender, Clinic Site, and Medical Problems
For each type of PSQ-defined sleep problem, simple logistic regression models of current or past mention of relevant symptoms on age and then gender showed no associations (P > .05). Similarly, Fisher's exact test revealed no statistically significant differences between clinics A and B in the frequencies with which relevant symptoms were mentioned.
Among the 44 SDB children, 18 had respiratory problems and 5 of these (28%) had SDB symptoms mentioned in their charts, whereas none of the 26 children without respiratory problems had SDB symptoms mentioned (Fisher's exact test, P = .008). In contrast, absence of medical problems (n = 22) was always associated with absence of SDB symptoms in charts, whereas 5 (23%) of SDB children with medical problems did have mention of such symptoms in their charts (P = .048). Other medical problems were too infrequent to be tested within the SDB group.
Among the 33 PLMS children, respiratory problems and absence of medical problems showed no association with chart mention of symptoms related to PLMS (P > .05). Among the 15 INS children, absence of medical problems showed no association with mention of INS-related symptoms. Among the 11 EDS children, no single medical problem was common enough to permit statistical testing with reasonable power, but the 2 children with respiratory problems were the only 2 who had sleepiness mentioned in their charts, and this result was significant (P = .018).
Chart Mention of Sleep Symptoms Other Than Those Identified by PSQ
Although only 5 (11%) of 44 patients whose PSQ results suggested SDB had symptoms of SDB discussed in their medical notes, 12 (27%) had INS discussed. Similarly, although only 5 (15%) of 33 patients (15%) with PSQ results suggestive of PLMS had directly related symptoms discussed in their notes, 18 (55%) had INS discussed and 6 (18%) had EDS discussed.
This review of medical records for 86 children with prominent symptoms of chronic sleep disorders suggests that pediatricians rarely discuss such symptoms, establish an underlying diagnosis, or institute needed treatment. Items from a validated questionnaire were used to detect children at substantial risk for SDB or PLMS, or to identify children with particularly problematic EDS and INS. In contrast, charts were reviewed with liberal interpretation of information as sleep-related. Despite this conservative study design, notes written during a 2-year interval indicate that the pediatricians at the 2 clinics discussed only 16 of 103 sleep problems and effectively treated only 3.
Data for comparison to these findings are scarce. In 1989 and 1990 the National Commission on Sleep Disorders Research searched records of over 10 million patients in large US primary care practice databases for codes that represented sleep disorders.15 They found only 73 such occurrences, all of which were codes for obstructive sleep apnea, despite a prevalence among adults of ∼3% for this disorder and much higher rates for conditions that cause INS.19,,20The Commission then examined in detail full written records for over 10 000 patients in family practice settings. Among 50 000 physician–patient contacts, notes that mentioned sleep were found for only 123. In none of these cases did the notes discuss an effective response to the patient's complaint. The commission summarized their findings in this way: “The conclusion is inescapable. At the present time, most Americans with sleep disorders have no recourse but to suffer from their disorder. Unless fortunate enough to know about a sleep disorders center or to see a physician who does, the vast majority of these people will continue to suffer from the symptoms and other consequences associated with their specific illness.”15
During the last decade, much progress has been made in both public and professional education about sleep disorders. Obstructive sleep apnea, in particular, has been the subject of many newspaper articles, magazine features, talk shows, and TV news journals. A recent epidemiologic study suggested that nearly 20% of middle-aged men and 10% of middle-aged women with symptomatic obstructive sleep apnea have been diagnosed.21 However, data for children remain largely unavailable. Many pediatricians expect questions from tired parents about sleep in infants, and many are familiar with common behavioral sleep disorders that emerge in subsequent years. However, the results of the current study suggest that other primary sleep problems may remain unaddressed in the large majority of affected children.
One problem likely to cause underrecognition of sleep disorders in children may be lack of physician training in this area. A 1978 survey found that 46% of medical schools offered no education or training in sleep disorders and that another 38% provided only occasional or fragmented information.22 A 1990–1991 survey of all 126 accredited US medical schools found that on average <2 hours of total teaching time was allocated to sleep and its disorders; 37 schools reported no structured teaching time in sleep medicine.23 More than two thirds of survey respondents thought that additional time should be spent, but reported obstacles included unavailability of qualified faculty, lack of curriculum time, and the need for additional clinical and educational resources. A 1992 survey of professional training programs found that, on average, pediatrics programs offered only ∼5 hours of didactic instruction in sleep disorders, and >40% of programs offered none.24,,25An assessment of 88 pediatricians' knowledge about sleep disorders suggested that they knew most about developmental issues and sleep hygiene and least about more specific sleep disorders.25
Another problem is that patients may not know that they should mention sleep problems to their physicians.15 Parents of young children seem to be more concerned about sleep problems than about difficulties with language development, motor development, toileting, and teething.25 However, adults often suffer from sleep symptoms for many years before they mention anything to their own physicians, and parents may not act differently on behalf of their children.
Failure to recognize childhood sleep problems may be exacerbated by a dearth of pediatricians who are fellowship-trained in sleep medicine. Many sleep specialists see both adults and children with sleep complaints, but the number of pediatric sleep specialists is still miniscule.
In the current study, lack of discussion of important sleep problems cannot be attributed to lack of any discussion about sleep: the large majority of children had notes about their sleep somewhere in their charts. Common chart comments such as “sleeps well” or “cough wakes her during sleep” suggest that discussions about sleep may be too brief or too focused on potential effects of known medical problems. At the same time, our findings also suggest that pediatricians may uncover SDB and EDS in some children precisely because a respiratory disorder other than SDB leads to consideration that the SDB might also be present. In other cases, pediatricians may uncover important sleep symptoms, such as INS, without knowing to ask about potential specific causes, such as PLMS.18
The current study is limited in several respects. It focused on only 2 general pediatrics clinics and is not an epidemiologic sample. The numbers of participants who qualified for this study reflect the protocol used to define sleep symptom-complexes rather than the prevalence of particular sleep problems. Although a number of pediatricians were involved with the 1395 patient contacts that were reviewed, these clinicians and patients may not adequately represent medical practices or problems at other sites. The results derived from the 2 clinics do not imply that care delivered there was in any way substandard; on the contrary, the academic affiliations of these clinics may have resulted in better identification of sleep problems than might be observed at sites without ready access to a sleep disorders center. This study used chart notes as proxy measures of discussions between caregivers, patients, and their parents. In some cases, such notes may inadequately document sleep-related inquiries, diagnoses, and recommendations that did take place. Finally, PSQ-identified sleep problems, such as SDB, PLMS, EDS, and INS, are often but not always chronic. In many cases, chart notes from as far as 2 years before PSQ completion may have omitted mention of sleep problems because no such problems existed at the time.
The apparent neglect of important pediatric sleep problems deserves study among additional practices, definition of underlying causes, and development of strategies to address the problem. The National Sleep Disorders Center, within the National Heart, Lung, and Blood Institute, has sponsored an initiative to improve sleep medicine education, and 20 groups are now able to develop and test tools that will be useful at several different stages of training. The first of these tools are now available for free at the MEDSleep website,http://www.asda.org/medsleep/home.htm. Wider dissemination of such strategies will have to be coupled with a growing understanding among program directors that time for education in sleep medicine must be inserted into already crowded curricula. Finally, education of the lay public should not be neglected in an effort to end needless sleepiness, behavioral problems, academic underachievement, and other morbidity associated with occult sleep disorders in children.
This work was supported by Grant K02-NS02009 from the National Institute of Neurological Disorders and Stroke, by Grant K07-HL03645 from the National Heart, Lung, and Blood Institute, and by Grant MO1 RR00042 to the University of Michigan General Clinical Research Center.
We thank the pediatricians who, committed to the improvement of child health care, permitted these chart reviews despite potential findings; the parents and children who took the time to complete questionnaires so that other children might benefit; and the graduate and undergraduate research assistants who patiently recruited participants at the 2 clinics.
- Received August 24, 2000.
- Accepted October 30, 2000.
Reprint requests to (R.D.C.) Michael S. Aldrich Sleep Disorders Laboratory, 8D8702, University Hospital, 0117, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0117. E-mail:
- SDB =
- sleep-disordered breathing •
- EDS =
- excessive daytime sleepiness •
- PLMS =
- periodic leg movements during sleep •
- INS =
- insomnia •
- PSQ =
- Pediatric Sleep Questionnaire
- Ali NJ,
- Pitson DJ,
- Stradling JR
- Picchietti DL,
- England SJ,
- Walters AS,
- Willis K,
- Verrico T
- Gozal D
- ↵National Commission on Sleep Disorders Research. Wake Up America: A National Sleep Alert. Washington, DC: Government Printing Office; 1993
- Costa e Silva JA,
- Chase M,
- Sartorius N,
- Roth T
- Moline ML,
- Zendell SM
- Mindell JA,
- Moline ML,
- Zendell SM,
- Brown LW,
- Fry JM
- Copyright © 2001 American Academy of Pediatrics