The international Concussion Group (CISG) has emphasized that there are several
modifying factors that have the potential to impact the management of sports-related concussion.
One of these factors is the presence of sleep problems. There is little published research
investigating the relationship between sleep problems and subsequent management of youth
following a concussion. Thus far, only one study has used a validated sleep measure and the
remaining studies have used the PCSS sleep items to form a “sleep problems” composite scale to
assess sleep. The PCSS was not designed to assess sleep functioning and to date no empirical
research has evaluated whether the PCSS sleep items correspond with other validated measures
of sleep functioning. Although the three measures of sleep (i.e. sleep diary, actigraphy and PSG)
are considered to be standards of care, they are more time consuming, costlier than rating scales,
and are often used in sleep-specific clinics studies. In contrast, the the PCSS is extensively used
in the adolescent population, and may serve as a potential screening measure when specific sleep
measures are not available or possible. The primary goal of this study was to examine the
psychometrics of the PCSS sleep-related items on the Child-SCAT3 to examine its clinical
utility. Lastly, if sleep problems alter baseline assessments of cognition and symptom reporting,
this could in turn affect the assessment and management of athletes following a concussive
injury. Thus, the secondary goal of this study was to examine how self-reported sleep problems,
as measured by the PCSS sleep-related items impact baseline performance on the Child-SCAT3.
50
Taken together, this study is important as it provides descriptive data on the PCSS sleep-related
items and PCSS sleep composite in a pediatric population, and evidence of psychometrics of the
PCSS sleep items. It also extends the utility of the PCSS sleep-related items for identifying sleep
problems in adolescents, and expands the research on baseline performance.
Summary and Discussion of Findings: Objectives 1 & 2
Convergent validity analysis. To assess the psychometrics of the PCSS sleep-related
items on the frequently used Child-SCAT3 the internal reliability (i.e. the consistency) and the
convergent validity were examined. The internal reliability for the PCSS-sleep related items and
Sleep/Wake Problems Behaviour scale was found to be poor, which was most likely due to the
restricted range and homogeneity of test items. Despite poor internal consistency for the scales,
as hypothesized, the PCSS sleep composite was significantly associated with the validated sleep
scale, the Sleep/Wake Problems Behaviour composite. With respect to the individual PCSS
sleep-related item analysis, all three PCSS sleep items were significantly associated with the
SWPB items on the validated sleep scale. Although small-to-medium correlations were observed
between the PCSS items, pairs of PCSS items were found to be associated with several SWPB
items.
The above differential finding is particularly noteworthy as it demonstrates the ability of
the PCSS sleep-related items to delineate between tiredness linked to different sleep complaints,
such as excessive daytime sleepiness and delayed sleep onset, a distinction that has important
implications for assessment and management for youth with concussions.
It could be argued that, the PCSS item “I get tired easily” and “I get tired a lot” and the
items it is associated with on the SWPB could be related to “hypersomnia,” or “excessive
sleepiness.” While “I daydream too much” could be related to either “Behavioural insomnia” or
51
“Delayed Sleep Onset.” Delayed sleep onset, otherwise known as “Delayed Sleep Phase
syndrome” (DSPS), is a common form of insomnia found in adolescents that cannot fall asleep
until midnight or later, and subsequently have difficulty awakening for school or staying awake
in early morning classes. This is often influenced by poor sleep habits or sleep hygiene. Students
with DSPS are likely to sleep through early morning classes (Wolfson et al., 2003) and one study
(Thorpy, Korman, Spielman, & Glovinsky, 1988) found that adolescents with DSPS reported
“daydreaming” and had “maximal sleepiness” mostly during the morning classes with a tendency
for greater alertness as the day progressed. Whereas “Hypersomnia” is more associated with
chronic sleepiness that is found in adolescents who take naps at inconvenient times (i.e. such as
during school) and experience trouble waking from long sleeps (Wolfson & Carskadon, 2003).
In some individuals, the sleep disturbance may contribute to impulsivity and risk-taking
behaviour (i.e. the SWPB item “done dangerous things without thinking”). Similarly, Becker &
colleagues (2011) found positive associations between CBCL sleep items and sleep disorder
diagnoses. For example, “sleeps more” was positively associated with hypersomnia and
“overtired” was positively associated with psychological insomnia. These distinctions are
important as they have important implications for assessment and treatment (Bodkin &
Manchanda, 2011).
A current challenge within the sleep disorder literature is evaluating a patient with a
symptom of “tiredness” and determining whether it means “tired or sleepy,” “fatigue or lack of
energy,” or “weak” which are often used interchangeably (Bodkin & Manchanda, 2011).
Understanding the distinction between the terms is important as each term refers to a different
type of sleep problem, and the corresponding evaluation and treatment differ according to what
is meant (Bodkin & Manchanda, 2011). For example, one patient may use the term “tired” to
52
describe their fatigue, while another patient may be describing hypersomnia. Determining what
is actually meant by the sleep complaint is a crucial step to the evaluation and management of
the patient.
That being said, this is speculative as a comprehensive factor analysis for the SWPB has
never been conducted. While more research is needed to tease a part the differences between the
PCSS sleep items “I get tired easily,” “I am tired a lot” and ‘I daydream too much,’ this
preliminary differential diagnosis is notable because it demonstrates the ability of the PCSS sleep
items to potentially distinguish different insomnias which is important for assessment.
In addition to the individual PCSS sleep items, the PCSS sleep composite was
significantly related to total scores of the SWPB. It should be noted almost all of the correlations
were medium-to-large, and only one association emerged between an item not included in the
SWPB total score and PCSS sleep-related items and PCSS sleep composite. For example, no
significant associations were found between the PCSS sleep-related items and the SWPB items a
(felt satisfied with your sleep?), e (awakened too early in the morning and couldn’t get back to
sleep?), l (had nightmares or bad dreams during the night?) and o (had a good night’s sleep?),
with the exception of n (done dangerous things without thinking?). This provides support for the
primary hypothesis that the sleep items on the PCSS are a valid measure of sleep problems for
children and adolescents.
Summary and Discussion of Findings: Objectives 3 & 4
Cross-validity analysis. The secondary aim of this project was to investigate the
influence of sleep problems in a pediatric population by comparing baseline self-reported
symptoms (total and severity), self-reported symptom clusters (eg, cognitive and somatic),
cognitive performance and BESS scores on the Child-SCAT3 across two distinct groups defined
53
by sleep problems as measured by the PCSS sleep-related items. No significant differences were
found between the athletes with and without sleep problems and the SAC-C total and BESS total
scores. As hypothesized, significant differences were observed between groups for Symptom
Total and Symptom Severity, with the sleep problems group reporting more symptoms and
indicating greater symptom severity (Mihalik et al., 2013; Silverberg et al., 2016; Sufrinko et al.,
2016). Lastly, with respect to the PCSS cluster item analysis, the sleep group reported higher
scores in both PCSS clusters, cognitive and somatic; however the effect sizes were small.
Likewise, positive correlations (medium effects) were found between the PCSS sleep problems
composite and the PCSS symptom clusters, cognitive and somatic.
The results reflect the current literature on sports-related concussions, that athletes who
report sleep disturbances during baseline testing endorse more post-concussion-like symptoms.
Moreover, the present findings are similar to Mihalik et al. (2013) who found athletes with low
sleep quality endorsed more of the somatic items on the GSC. However, they did not find a
significant difference between low sleep quality and the cognitive items on the GSC. That being
said, the present findings are in agreement with the current literature that sleep difficulties
impacts cognitive and somatic symptoms. For example, the results are similar to Brooks et al.
(2016), who found a small effect size between male athletes with attention problems and those
without. Brooks et al. (2016) found that boys with attention problems reported more symptoms
than boys without attention problems in the sleep-arousal domains. Examples of symptoms
reported more commonly by boys with attention problems were trouble falling asleep (31% in
boys with attention problems, 21% in boys without attention problems). They also found that
boys with attention problems had more difficulty concentrating (38% in boys with attention
problems, 16% in boys without attention problems) (Brooks, Iverson, Atkins, Zafonte, &
54
Berkner, 2016). The significant effect of sleep disturbance on number and severity of cognitive
symptoms are common in children and adolescents who experience sleep deficiency (Kostyun et
al., 2015). Furthermore, the symptoms in the somatic category, specifically headache, nausea and
faint, are commonly associated with concussions (Howell et al., 2016), was also affected by
sleep disturbance. This is a significant finding because clinicians often question more about
common somatic symptoms such as these during an initial concussion assessment.
In summary, the sleep problems impacted the PCSS symptom scores (eg, total and
severity) and symptom clusters (eg, cognitive and somatic); however it has no discernable effect
on cognitive and balancing performances (Sufrinko, Pearce, Johnson, et al., 2015). While this
was hypothesized, Sufrinko et al. (2016) did observe differences between sleep problem groups
but only after combing both sleep duration and sleep problems.
“Sleep disturbance” can be assessed by either sleep quality or sleep quantity. The current
study only assessed sleep quality, which might explain why these differences were not observed
for cognition and balancing performances. McClure et al., (2014) found that athletes who slept
fewer than 7 hours the night prior to testing performed significantly worse on 3 of the 4 ImPACT
composite scores (i.e. reaction time, visual memory and verbal memory). Athletes who
performed the best on reaction time slept 7 to 9 hours per night, and athletes who slept fewer
than 7 hours had slower reaction times than athletes who slept greater than 9 hours. The majority
of athletes in this study reported getting at least 9 hours of sleep per night within the last two
weeks. Another possible explanation as to why differences were not observed could be that the
sample size (n=80) was not large enough. McClure et al. (2014) used a large sample (n=3686)
and Sufrinko et al. (2016) argued that while there were differences, these differences were small
(Cohen’s d=.10) and may have no practical or clinical significance. That being said, meaningful
55
differences did emerge when reduced sleep duration was considered in concert with self-reported
sleep symptoms on the ImPACT (Sufrinko et al., 2015). Sufrinko et al. (2016) compared athletes
with both reduced sleep duration and concurrent sleep-related symptoms to athletes reporting
optimal sleep and revealed meaningful differences in neurocognitive test performance and
symptoms. They found this was particularly significant among female athletes. The sample for
this project was all male athletes, and the literature has noted that males are not as affected by
sleep on neurocognitive tests and often perform better than female athletes on psychomotor
vigilance tasks, under equivalent sleep conditions (Ballester, Huertas, Yuste, Llorens, &
Sanabria, 2015; Sufrinko, Pearce, Johnson, et al., 2015) This could also be why no differences
were observed on the Child-SCAT3 psychomotor task, the BESS. In addition, Sufrinko et al.
(2016) only used athletes that endorsed both or none of the PCSS sleep-related items on the
ImPACT. The majority of athletes in this project’s sleep problems group endorsed one PCSS
sleep-related item, with the remaining endorsing two and then three. This finding is similar to
what prior researchers have found, that is, there is individual variability in sleep need (Jenni &
Carskadon, 2007) . Or this demonstrates the absence of perceived problem with sleep problems
among children and adolescents (Klein & Wilson, 2002). While the sample size for this study
was moderate, the majority of athletes included in the SLEEP SX group endorsed only one of the
PCSS sleep-related items and reported getting adequate sleep, this could be why this study and
Sufrinko et al. (2015), who used similar criteria for their groups, did not observe differences in
cognitive or balancing performances. As previously mentioned, Sufrinko et al. (2016) only
observed differences when they combined both sleep problems and sleep duration. While not
conclusive, this suggests that both sleep duration and sleep problems need to be considered as a
combined modifying factor. Overall the results are in agreement with previous studies, that sleep
56
quality influences symptoms but had no significant effect on the cognitive measures or BESS
scores. That being said, this comparison must be interpreted with caution as the previous studies
that assessed baseline functioning utilized different concussion tools. While the ImPACT is
similar to the SCAT there are differences between the cognitive measures. The SCAT mainly
assesses long-term and short-term memory using word lists, while the ImPACT uses both words
and shapes for it’s memory tasks (i.e. they are tested both on their verbal memory and spatial
memory). In addition, the sleep-related items used on the ImPACT PCSS differ from the PCSS
sleep-related items used on the Child-SCAT3.
Implications for Research and Practice
The results herein support the need for understanding and evaluating possible preexisting
sleep problems among adolescents, which was discussed in a recent short
communication article “The Emering Importance and Evaluation of Pre-Injury Sleep
Characteristics among Adolescents with Traumatic Brain Injury.” Therefore, this study has
implications for both research and practical domains.
Practical implications. The authors of the short communication, “The Emerging
Importance of Evaluation of Pre-Injury Sleep Characteristics among Adolescents with Traumatic
Brain Injury” proposed adding another instrument, the short version of the Adolescent Sleep
Wake Scale (ASWS). While beneficial, this might not be an option to healthcare providers or
other professionals who work with athletes. This study, albeit a pilot study, provides validity for
the Child-SCAT3 sleep items and sleep items composite for measuring sleep in children and
adolescents. For adults, Lau et al. (2012) reports that the symptoms “sleep more” vs. “less than
usual” on the PCSS falls into different symptomatic clusters: cognitive and sleep, respectively
(Lau, Collins, & Lovell, 2012). Interestingly, in a supplementary analysis (not included in the
57
results section) it was found that the symptoms “I daydream too much” and “I am tired easily”
on the PCSS fell into different symptomatic clusters: cognitive and somatic, respectively, with
medium effect sizes. Overall this study is the first to provide evidence for parsing the effects of
the PCSS-sleep related items on the Child-SCAT3.
Secondly, given the results of this study, it might be wise to consider sleep as a single
modifying factor. In doing so, we can then speculate that return to play and return to learn
decisions could be complicated by perceived differences in baseline and post-concussive sleep
scores. Post-concussive athletes often report more sleep items than their recorded baseline levels,
particularly when rest is recommended as a first-line therapy (Sufrinko et al., 2016). Moreover,
sleep symptoms such as longer sleep onset latency, difficulty maintaining sleep, and increased
daytime sleepiness, are some of the most common sleep disturbances that persist following a
concussion.
Likewise, if baseline levels of sleep and its adverse effects on related symptoms are not
accounted for, post-concussive symptoms may appear to be solely a result of the injury rather a
partial function of premorbid status. Sufrinko et al. (2015) observed that athletes with preinjury
sleep difficulties following an injury had worse sleep-related symptoms when compared to
controls. This is important for clinicians to be aware of, as this may potentially inflate the overall
symptom score when managing these athletes. Lastly, given the current results, athletes that
show specific sleep difficulties on the PCSS at baseline may benefit from education regarding
proper sleep, for example those with poor behavioural sleep habits, may benefit from education
on proper sleep hygiene (i.e., regulated sleep schedule, sleep rituals, and no naps) (Sufrinko et
al., 2015).
58
Research implications, limitations & future directions. The strengths of this study are
the use of a pediatric population and was the first to evaluate the effect of sleep problems on
baseline tests. Nonetheless, several limitations should also be noted. First, objective measures of
sleep were not used. Assessing sleep exclusively using the PCSS-sleep related items is far from
optimal; however, this measure may be useful for understanding sleep problems and how they
influence the assessment and management of youths with concussions.
It might be wise to also use another validated sleep scale, such as the Pittsburgh Sleep
Quality Index or the Epworth Sleepiness Scale, which have both been validated against objective
measures of sleep in patients with mTBI. The sleep scale used in this study, the Sleep/Wake
Problems Behaviour scale is a part of the Sleep School Survey, and while widely used and has
moderate reliability, no empirical research has been conducted on the validity or a factor analysis
on what factors the Sleep/Wake Problems Behaviour scale measures. Another useful instrument,
such as the short version of the Adolescent Sleep Wake Scale (ASWS) could be used as well.
The ASWS is a 10-item scale with three factors including (a) Falling Asleep and Reinitiating
Sleep-Revised, (b) Returning to Wakefulness-Revised, and (c) Going to Bed-Revised.
Second, the generalizability of this study should also be considered a limitation. The
current sample included children of ages 10 – 12 years who are ice hockey players in Edmonton,
Alberta and therefore may not be generalizable to youths in different sports. Moreover, this study
used a moderate sample size, future research should use a larger sample size, ideally with
athletes who play high-contact sports, as these individuals are more likely to have a sportsrelated
concussion. With a larger sample size, it is advisable to use athletes who endorse all of
the sleep-related PCSS items, similar to Sufrinko et al. (2016). Third, the reporting time period
on the PCSS is the day of the assessment while the reporting time on the SWPB measure was
59
within the last 2 weeks. It is recommended that subsequent studies follow-up with a clinical
interview with families to clarify any inconsistencies across the measures and corroborate the
athlete’s self-reported sleep duration. The data included in this study was mainly child-reported
sleep data, future studies should examine parent-reported sleep functioning in relation to the
PCSS sleep-related items/composite.
Conclusion
In conclusion, although the PCSS is not a sleep-specific measure and was not intended to
be developed to assess sleep problems in children and adolescents, the correspondence between
the PCSS sleep-related items and the PCSS sleep composite score reasonably well with a
validated measure of sleep functioning in school-aged children. While more research is needed,
this study does provide support for the use of the PCSS sleep related items to be used in research
studies, as well as a potential screening measure for sleep problems in young athletes when
specific sleep screening is not available or possible. Lastly, the results herein support the current
findings on sleep problems impact on baseline performance. That is, athletes that self-report
sleep problems as measured by the PCSS sleep-related items endorse more symptoms on the
PCSS. Although this is a pilot study, this is the first project to examine the Child-SCAT3’s
potential at assessing sleep difficulties and the influence of sleep problems on baseline measures
for this tool in school-aged male ice hockey players.