The Social Epidemiology of Sleep 1St Edition Dustin T Duncan Full Chapter PDF
The Social Epidemiology of Sleep 1St Edition Dustin T Duncan Full Chapter PDF
The Social Epidemiology of Sleep 1St Edition Dustin T Duncan Full Chapter PDF
https://ebookmass.com/product/epidemiology-of-diabetes-1st-
edition-m-d-moini/
https://ebookmass.com/product/epidemiology-of-thyroid-
disorders-1st-edition-m-d-moini/
https://ebookmass.com/product/the-pyramid-of-lies-duncan-mavin-2/
https://ebookmass.com/product/the-pyramid-of-lies-duncan-mavin/
Britain and the Arctic 1st Edition Duncan Depledge
(Auth.)
https://ebookmass.com/product/britain-and-the-arctic-1st-edition-
duncan-depledge-auth/
https://ebookmass.com/product/sustainable-energy-transitions-
socio-ecological-dimensions-of-decarbonization-1st-ed-edition-
dustin-mulvaney/
https://ebookmass.com/product/critical-epidemiology-and-the-
peoples-health-1st-edition-jaime-breilh/
https://ebookmass.com/product/oxford-handbook-of-sleep-medicine-
consultant-neurologist-professor-of-neurology-and-sleep-medicine-
guy-leschziner/
The Social
Epidemiology of Sleep
The Social
Epidemiology of Sleep
Edited by
Dustin T. Duncan
Ichiro Kawachi
and
Susan Redline
1
1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
This material is not intended to be, and should not be considered, a substitute for medical or other
professional advice. Treatment for the conditions described in this material is highly dependent on
the individual circumstances. And, while this material is designed to offer accurate information with
respect to the subject matter covered and to be current as of the time it was written, research and
knowledge about medical and health issues is constantly evolving and dose schedules for medications
are being revised continually, with new side effects recognized and accounted for regularly. Readers
must therefore always check the product information and clinical procedures with the most up-to-
date published product information and data sheets provided by the manufacturers and the most
recent codes of conduct and safety regulation. The publisher and the authors make no representations
or warranties to readers, express or implied, as to the accuracy or completeness of this material.
Without limiting the foregoing, the publisher and the authors make no representations or warranties
as to the accuracy or efficacy of the drug dosages mentioned in the material. The authors and the
publisher do not accept, and expressly disclaim, any responsibility for any liability, loss or risk that
may be claimed or incurred as a consequence of the use and/or application of any of the contents of
this material.
1 3 5 7 9 8 6 4 2
Paperback printed by Marquis, Canada
Hardback printed by Bridgeport National Bindery, Inc., United States of America
Contents
Part II: Sleep Health Over the Life Course and Among Special Populations
3. Sleep in Pregnancy 49
Louise M. O’Brien and Galit L. Dunietz
Glossary 431
Index 435
About the Editors
Ichiro Kawachi, MBChB, PhD, is the John L. Loeb and Frances Lehman Loeb
Professor of Social Epidemiology at the Harvard T. H. Chan School of Public Health,
where he has taught since 1992. Dr. Kawachi received both his medical degree and his
PhD (in epidemiology) from the University of Otago, New Zealand. Dr. Kawachi is the
coeditor (with Lisa Berkman) of the first textbook on social epidemiology, titled Social
Epidemiology (Oxford University Press, 2000; new and revised edition published in
2014 with Lisa Berkman and Maria Glymour). His other books include Neighborhoods
vii
viii About the Editors
and Health with Lisa Berkman (Oxford University Press, 2003; new and revised edition
published in 2018 with Dustin Duncan); Globalization and Health with Sara Wamala
(Oxford University Press, 2006); Social Capital and Health with S. V. Subramanian
and Daniel Kim (Springer, 2008); the Oxford Handbook of Public Health Practice with
Charles Guest, Walter Ricciardi, and Iain Laing (Oxford University Press, 2013); Global
Perspectives on Social Capital and Health with Soshi Takao and S. V. Subramanian
(Springer, 2013); and Behavioral Economics and Public Health with Christina Roberto
(Oxford University Press, 2015). His current NIH-funded R01 project is focused on
the longitudinal impacts of community social cohesion/social capital on functional
recovery after the March 11, 2011, Great Eastern Japan earthquake and tsunami. In
2013, he launched a massive, open online course (MOOC) through HarvardX called
“Health and Society” (PHx 201), which is a version of a class he has taught for 20 years
at Harvard. In 2013–2014, 35,000 participants from throughout the world enrolled in
the MOOC course. Dr. Kawachi is the coeditor in chief (with S.V. Subramanian) of
the international journal Social Science & Medicine. He is an elected member of the
Institute of Medicine of the U.S. National Academy of Sciences.
Susan Redline, MD, MPH, is the Peter C. Farrell Professor of Sleep Medicine at
Harvard Medical School. She directs programs in Sleep and Cardiovascular Medicine
and Sleep Medicine Epidemiology at Brigham and Women’s Hospital and Beth Israel
Deaconess Medical Center. Dr. Redline’s research includes epidemiological studies and
clinical trials designed to (a) elucidate the etiologies of sleep disorders, including the
role of genetic and early life developmental factors, and (b) understand the cardiovas-
cular and other health outcomes of sleep disorders and the role of sleep interventions in
improving health. She has led sleep assessments made in numerous community-based
cohort studies and clinical trials, as well as national initiatives aimed at identifying the
genetic and environmental contributors to sleep disorders. She has published over 450
peer-reviewed articles and has served the sleep research community in a number of
capacities, including as a member of the board of directors for the American Academy
of Sleep Medicine and the Sleep Research Society, the NIH’s Sleep Disorders Research
Advisory Board, and the Institute of Medicine’s Committee on Sleep Medicine and
Research; she is also deputy editor for the journal Sleep and for Annals of the American
Thoracic Society. She received her BS and MD degrees from Boston University and an
MPH degree from the Harvard T. H. Chan School of Public Health and completed
internal medicine and pulmonary and critical care medicine training at Case Western
Reserve University and a research fellowship in respiratory epidemiology at Harvard
Medical School.
Contributors
ix
x Contributors
An Introduction to
Sleep Epidemiology
1
Sleep Epidemiology
A Social Perspective
Dustin T. Duncan, ScD, Ichiro Kawachi, MBChB, PhD, and
Susan Redline, MD, MPH
When we sleep, where we sleep, and with whom we sleep are all important markers or
indicators of social status, privilege, and prevailing power relations.
Simon J. Williams, 2005
University of Warwick
3
4 An Introduction to Sleep Epidemiology
and family economic pressures by increasing the numbers of work hours and duration
of work shifts. In contrast, “exempt” workers have work hour flexibility and can have
economic value not directly related to hours worked but level of complexity, creativity,
or innovation of their work contributions. Differences in economic resources provide
higher income individuals the resources to assist with home and family obligations,
while lower income working individuals need to find time to address a myriad child
and elder care and household obligations in a finite number of nonoccupational work
hours. These days, the ability to get a decent night’s sleep has become a symbol of
power, prestige, and status.
What are the consequences of poor sleep for population health? Answering this
question using an epidemiological framework has been limited by the relative dearth
of sleep data included in community studies and the late emergence of sleep epidemi-
ology as a discipline, as discussed in Chapter 2. When we sift through the surveys of
early cohort studies (like the Framingham Heart Study, established in 1948), we can
see that the other health behaviors are all represented on early questionnaires (i.e.,
smoking, drinking, exercise, diet) but not sleep, which was not included in surveys
until the 1980s. In 1977, the National Health Interview Survey included a single
question on sleep duration, and, in 1982–1984, the National Health and Nutrition
Examination Survey included seven items on sleep quality, sleeping medications, and
sleep duration. The Nurses Health Study (NHS) began in 1976 but did not include
questions regarding sleep until 1986. No objective measurements of sleep were in-
cluded in any NHS examination, although a pilot study of sleep apnea screening is
planned for 2020. Objective measurements of sleep were not included in large cohort
studies until the launch of the Wisconsin Sleep Cohort in 1993 and the Sleep Heart
Health Study in 1994.
As epidemiologists started to ask about sleep, it became quickly apparent that poor
sleep health contributes to a broad range of health outcomes across populations, in-
cluding premature death, chronic disease, and mental health conditions—not to men-
tion accidents/injuries, (un)happiness, marital discord, learning impairment, and
cognitive decline (see Chapter 2). Despite the strong evidence linking poor sleep to
the population’s leading causes of death and disability, poor sleep is not identified as a
risk factor in the World Health Association’s Global Burden of Disease. The U.S. gov-
ernment has an Office on Smoking and Health at the Centers for Disease Control and
Prevention and a National Institute of Alcohol Abuse and Alcoholism; the Surgeon
General has issued Reports on Physical Activity and Health across the Atlantic; and the
UK government recently even appointed a Minister of Loneliness—but sleep has yet
to reach a comparable level of policy attention. However, early success in recognizing
sleep as a public health concern is reflected in the recent inclusion of sleep health as
an objective in the United States’ Healthy People 2020 goals. For the first time, the
U.S. Department of Health and Human Services specifically prioritized the goal to
“Increase public knowledge of how adequate sleep and treatment of sleep disorders
improve health, productivity, wellness, quality of life, and safety on roads and in the
workplace.” Moreover, internationally, the highly influential Wellcome Trust convened
5 Sleep Epidemiology: A Social Perspective
economic production in society. Karl Marx said as much, commenting on the inven-
tion of night work and shift work during the Industrial Revolution:
The prolongation of the working day beyond the limits of the natural day, into the
night, only acts as a palliative [to the extraction of maximum surplus value from
labourers]. It only slightly quenches the vampire thirst for the living blood of la-
bour. Capitalist production therefore drives, by its inherent nature, towards the
appropriation of labour throughout the whole of the 24 hours of the day. (Marx,
2018, p. 367)
And even though nine-year-old children are no longer put to work in the steel furnaces
and cotton mills of wealthy countries in the 21st century, the social patterning of sleep
still persists due to disparities in people’s employment contracts, working conditions,
and commuting patterns as well as a host of other social dimensions such as dealing
with family/home demands and stress related to being from a low socioeconomic status
background and experiencing discrimination across contexts, including at work.
When we broaden the lens to encompass broader upstream forces, Uri
Bronfenbrenner’s social ecological model demonstrates that individual-, network-,
community-, and public policy–level factors independently and synergistically influ-
ence health and well-being (Duncan, Al-Ajlouni, & Chaix, 2019). From the lens of
the social ecological model, social factors at these multiple levels can influence sleep.
Thus, for example, an immigrant in America working at a minimum wage night shift
job in the service sector not only has to endure the disruption of his or her circa-
dian rhythm through the employment contract but may additionally experience the
daily challenges of workplace sexual harassment, stresses related to loss of autonomy,
discrimination against minorities and immigrants in society, doubling up in crowded
housing conditions, and living in a noisy, unsafe neighborhood. Addressing these
broader social forces will require more than educational tips on sleep hygiene (“avoid
coffee before going to bed”) or the occasional dose of a hypnotic medication.
and therapies for poor sleep health, which this book hopes to inspire. Accordingly
this book is targeted to (a) social epidemiologists who wish to study sleep as a health
behavior and health outcome; (b) sleep epidemiologists who want to learn about the
social determinants of sleep; and (c) other scholars working in the intersection of sleep
health, social epidemiology, and health disparities.
sleep disruption (Eshkoor, Hamid, Nudin, & Mun, 2014). Religion/spirituality and
sleep is also not explicitly focused on in the book. No chapters explicitly focus on psy-
chological factors, such as stress, coping, or resilience. However, psychological factors
(especially stress) are highlighted as a pathway linking social determinants to sleep
across chapters, including in Chapter 14 on housing and sleep. We did not include a
separate chapter on gender and sleep, given the large and consistent body of research
documenting gender differentials in sleep: women tend to report more sleep dura-
tion but poorer sleep quality as compared to men (Burgard & Ailshire, 2013; Fatima,
Doi, Najman, & Mamun, 2016; Mallampalli & Carter, 2014). Instead, there is men-
tion of this important component of sleep health throughout the book. Among other
sociodemographic factors, gender is thought to modify existing relationships between
social determinants and sleep; this is discussed in several of the ensuing chapters and is
essential to the discussion in the ensuing chapters on life course issues in sleep.
Because sleep exists in a social context, we hope that this book encourages new
research on sleep science and sleep disorders within the social context. The emerging
science of the social epidemiology of sleep health offers many opportunities for social
justice and population health improvement.
ACKNOWLEDGMENTS
We thank Orfeu Buxton, William Goedel, Dayna Johnson, and Byoungjun Kim for
providing comments on an earlier version of this chapter. We also thank Yazan Al-
Ajlouni for completing the chapter formatting.
Dr. Dustin Duncan was funded by grants from the National Institute on Mental
Health (Grant Number R01MH112406), National Institute on Minority Health and
Health Disparities (Grant Number R01MD013554), and the Centers for Disease
Control and Prevention (Grant Number U01PS005122).
REFERENCES
Altevogt, B. M., & Colten, H. R. (2006). Sleep disorders and sleep deprivation: An unmet public
health problem. Washington, DC: National Academies Press.
Avidan, A. Y., & Zee, P. C. (2011). Handbook of sleep medicine. Philadelphia, PA: Lippincott
Williams & Wilkins.
Berkman, L. F., & Kawachi, I. (2000a). A historical framework for social epidemiology: Social
epidemiology. New York, NY: Oxford University Press.
Berkman, L. F., & Kawachi, I. (Eds.). (2000b). Social epidemiology. New York, NY: Oxford
University Press.
Berkman, L. F., Kawachi, I., & Glymour, M. M. (Eds.). (2014). Social epidemiology. 2d ed.
New York, NY: Oxford University Press.
Bowleg, L. (2012). The problem with the phrase women and minorities: Intersectionality—
An important theoretical framework for public health. American Journal of Public Health,
102(7), 1267–1273.
Burgard, S. A., & Ailshire, J. A. (2013). Gender and time for sleep among U.S. adults. American
Sociological Review, 78(1), 51–69.
9 Sleep Epidemiology: A Social Perspective
Cappuccio, F., Miller, M. A., & Lockley, S. W. (2010). Sleep, health, and society: From aetiology
to public health. New York, NY: Oxford University Press.
Duncan, D. T. (2015). Putting health and behavior in context: Current knowledge and fu-
ture directions on research embracing ecological perspectives. Behavioral Medicine, 41(3),
172–176.
Duncan, D. T., Al-Ajlouni, Y. A., & Chaix, B. (2019). Multilevel perspectives on urban health.
In S. Galea, C. K. Ettman, & D. Vlahov (Eds.), Urban health (p. 230). New York, NY: Oxford
University Press.
Eshkoor, S. A., Hamid, T. A., Nudin, S. S. H., & Mun, C. Y. (2014). Importance of hyper-
tension and social isolation in causing sleep disruption in dementia. American Journal of
Alzheimer’s Disease & Other Dementias, 29(1), 61–66.
Fatima, Y., Doi, S. A. R., Najman, J. M., & Mamun, A. A. (2016). Exploring gender difference
in sleep quality of young adults: Findings from a large population study. Clinical Medicine
& Research, 14(3–4), 138–144. doi:10.3121/cmr.2016.1338
Julia, C., & Valleron, A.-J. (2011). Louis-René Villermé (1782–1863), a pioneer in social ep-
idemiology: Re-analysis of his data on comparative mortality in Paris in the early 19th
century. Journal of Epidemiology & Community Health, 65(8), 666–670.
Krieger, N. (2001). A glossary for social epidemiology. Journal of Epidemiology & Community
Health, 55(10), 693–700.
Lichstein, K. L., Durrence, H. H., Riedel, B. W., Taylor, D. J., & Bush, A. J. (2013). Epidemiology
of sleep: Age, gender, and ethnicity. New York, NY: Psychology Press.
Mallampalli, M. P., & Carter, C. L. (2014). Exploring sex and gender differences in sleep
health: A Society for Women’s Health research report. Journal of Women’s Health, 23(7),
553–562. doi:10.1089/jwh.2014.4816
Marx, K. (2018). The working day. In K. Marx, Capital (Vol. 1).
Maume, D. J. (2013). Social ties and adolescent sleep disruption. Journal of Health and Social
Behavior, 54(4), 498–515.
Oakes, J. M., & Kaufman, J. S. (2006). Methods in social epidemiology (Vol. 16). Hoboken,
NJ: John Wiley.
World Health Organization. (n.d.). Commission on social determinants of health. 2005–2008.
Retrieved from https://www.who.int/social_determinants/thecommission/en/
2
Sleep Epidemiology
An Introduction
Susan Redline, MD, MPH, Brian Redline, BA, and Peter James, ScD
INTRODUCTION
Sleep is that golden chain that ties health and our bodies together.
Thomas Dekker (playwright; 1572–1632)
11
Publication Count for Articles on “Sleep” AND “Epidemiology” OR “Community”
OR “Population” in Pubmed since 1980
1000
900
800
700
600
500
400
300
200
100
0
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1900
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
Sleep Quality
Sleep Disorders
Note: EEG = electroencephalogram; REM = rapid eye movement; OSA = obstructive sleep apnea;
RLS = restless legs syndrome.
16 An Introduction to Sleep Epidemiology
CONCEPTUALIZING SLEEP
IN EPIDEMIOLOGICAL STUDIES
Social epidemiological studies of sleep generally address a set of questions that relate
to understanding the upstream determinants and/or downstream effects of healthy
and disturbed sleep. These questions and associated research approaches are described
in Chapter 1 and in the forthcoming chapters of this book. The complex interplay of
social, environmental, and behavioral factors and their variation across the lifespan
is summarized in Figure 2.2. In this framework, social, behavioral, and environ-
mental factors influence sleep partly via epigenetic processes. Sleep health interrelates
bidirectionally with physical health, mental health, and well-being. A challenge for the
sleep epidemiologist therefore is to (a) identify and understand the sleep dimensions
relevant to given hypotheses; (b) adopt appropriate sleep measurement strategies;
(c) collect necessary data on relevant social, behavioral, environmental, and biological
factors; and (d) integrate these measurements into a relevant conceptual framework
and apply appropriate causal modeling strategies. A number of conceptual frameworks
Life Course Factors
Childhood Puberty &
In-utero Adulthood Aging
development adolescence
• Pregnancy • Menopause
Social Environmental
Determinants Factors Sleep-related
behaviors
Sleep Health
a deficit in the quantity or quality of sleep obtained versus the amount needed
for optimal health, performance, and well-being; sleep deficiency may result from
prolonged wakefulness leading to sleep deprivation, insufficient sleep duration,
sleep fragmentation, or a sleep disorder, such as in obstructive sleep apnea, that
disrupts sleep and thereby renders sleep non-restorative.
MEASURING SLEEP
Sleep, sleep-related symptoms, and sleep disorders can be characterized using a variety
of tools, such as validated questionnaires, paper or electronic sleep diaries, actigraphy/
accelerometry, and polysomnography (PSG).
19 Sleep Epidemiology: An Introduction
Questionnaires
Questionnaires are used to elicit information on perceived sleep quality, daytime im-
pairment (alertness, sleepiness), and functional impairment (sleep-related limitations
of daily activities); screen for sleep disorders or estimate their likelihood (sleep apnea,
insomnia, periodic limb movement disorder, hypersomnolence); assess sleep dura-
tion and timing; characterize sleep habits and sleep-related behaviors; and describe
chronotype. Questions assessing sleep are more readily incorporated into large-scale
studies than objective measurements due to their relatively low cost and participant
burden. However, questionnaire-based sleep assessments are subject to reporting
errors and biases, which can be particularly large for behaviors that occur during the
sleep period (when the individual is poorly aware of his or her behaviors), or that result
in chronic symptoms that individuals come to accept over time as “normal.” A sum-
mary of commonly used questionnaires is provided in Table 2.2.
Instrument Name Full Version Length Short Forms Subdomains/ Response Format Time Period Assessed Internal Test–Retest
Categories Consistency Reliability
Measures of sleep quality, symptoms, and daytime functioning
Pittsburgh Sleep Quality 19 items - 7 clinical domains 4-point scales, time entry Prior month 0.83 0.85
Index
Epworth Sleepiness Scale 8 items - - 4-point scales In recent times 0.73–0.86 *see note
PROMIS Sleep Disturbance CAT 27-item bank 4-, 6-, - 5-point scales Prior 7 days
instrument 8-item
PROMIS Sleep-related CAT 16-item bank 4- and - 5-point scales Prior 7 days
Impairment instrument 8-item
Functional Outcomes of 30 items 10-item 5 factor subscales 5-or 7-point scales In general 0.90 0.95
Sleep Questionnaire (SF = 0.87)
Insomnia Questionnaires
Women’s Health Study 5 items – – 4-and 5-point scales Prior 4 weeks 0.70–0.85 0.96
Insomnia Severity Scale
Obstructive Sleep Apnea Questionnaires
Berlin Questionnaire 10 items – 3 categories Yes/no, 4- and 5-point Current and lifetime 0.68–0.98
scales
Calgary Sleep Apnea 35 items – 4 categories + 1 7-point scales Prior 4 weeks 0.88–0.92 0.92
Quality of Life Index treatment-related
RLS Questionnaires
International Restless Legs 10 items – 2 subscales 5-point scales Prior week 0.76
Scale
Note. CAT = computerized adaptive test; RLS = restless legs syndrome.
22 An Introduction to Sleep Epidemiology
item, precluding the ability to analyze a discrete marker of sleep disordered breathing
(which would not include coughing).
The Epworth Sleepiness Scale is a widely used measure of sleepiness in clinical
practice and research. It is a self-administered eight-item questionnaire that asks
respondents to rate their sleepiness using a 4-point scale in reference to different
settings, yielding overall scores of 0 to 24 (Johns, 1991, 1992). Higher scores indi-
cate more severe sleepiness, and scores above 10 are interpreted as indicating exces-
sive daytime sleepiness. A systematic review of its psychometric properties showed
good internal validity (0.73–0.86; Kendzerska, Smith, Brignardello-Petersen, Leung,
& Tomlinson, 2014), but the authors of this review concluded that there was inade-
quate data to make definitive assessments of test–retest reliability. In a clinical sample,
wide variation in scores over an approximately 70-day period was reported (Nguyen
et al., 2006). Use in older individuals appears limited, with 23% of participants in one
study unable to complete all items (Onen et al., 2013). External validity as assessed by
correlations with objective sleepiness measurement (the Multiple Sleep Latency Test)
is low to modest (Kendzerska et al., 2014), which could reflect weaknesses of the in-
strument or the unique information provided by self-appraisal compared to objective
measurements.
The Patient- Reported Outcomes Measurement Information System (PROMIS;
PROMIS® Instrument Development and Validation: Scientific Standards, 2013) is “a set
of person-centered measures that evaluates and monitors physical, mental, and social
health in adults and children” (HealthMeasures, n.d.). This battery of instruments was
developed using item response theory and designed to be applicable across a range of
conditions and populations. The PROMIS system includes two sleep assessments: the
PROMIS Sleep Disturbance instrument and the Sleep-Related Impairment instrument,
which, respectively, ask about self-reported perceptions of sleep quality (assessing
concerns about getting to sleep, maintaining sleep, and sleep satisfaction) and sleep-
related impairment (assessing perceived alertness and functional impairments asso-
ciated with sleep problems or impaired alertness), each asked in reference to the past
seven days (Yu et al., 2011). The full item question bank, consisting of 27 items on sleep
disturbance and 16 on sleep impairment, can be administered as a computerized adap-
tive test (CAT), in which participant responses guide which subsequent questions are
displayed, using one of several online and app-based CAT data collection instruments
(e.g., NIH Toolbox iPad App). Standard four-and eight-item short forms (as well as a
six-item option for Sleep Disturbance) are also available and can be scored using the
HealthMeasures Scoring Service, a free, online resource that accommodates missing
data and enables comparisons across groups of participants, or allows scoring man-
ually using raw score to T score conversion tables. The PROMIS sleep instruments
were calibrated and centered using a sample from the general population and a clinic
sample, with the mean T score of 50 representing the mean for a sample enriched with
individuals with chronic health problems. Higher scores indicate more of the construct
being measured. Although less commonly used than the PSQI and Epworth Sleepiness
Scale, initial data suggest that the PROMIS sleep instruments have greater measure-
ment precision than those instruments (Yu et al., 2011).
23 Sleep Epidemiology: An Introduction
Insomnia
Insomnia is defined on the basis of reported difficulties initiating sleep, maintaining
sleep, early morning awakenings, and associated daytime impairment. It is a dis-
order based on reported symptoms. While a diagnosis of insomnia requires a clin-
ical interview, in epidemiological studies validated questionnaires are used to classify
individuals with and without symptoms of insomnia, as well as to obtain indices that
quantify severity of symptoms. (Notably, neither polysomnography nor actigraphy are
required to diagnose insomnia, but they can identify characteristic findings such as
prolonged sleep latency or decreased sleep efficiency.)
The Insomnia Severity Index (ISI) is commonly used in clinical settings to screen
for insomnia (followed by diagnostic interviews) and in research settings to quan-
tify symptoms of insomnia (Morin, Belleville, Belanger, & Ivers, 2011). It is a seven-
item questionnaire that assesses the nature, severity, and impact of insomnia over the
prior two weeks with questions regarding sleep onset, sleep maintenance, and early
morning awakening problems, sleep dissatisfaction, impact of sleep disturbances
with daytime functioning, observed sleep problems by others, and distress caused
by the sleep difficulties. Responses are ascertained using a 5-point scale, with total
scores range from 0 to 28. A cutoff of 10 has a sensitivity of 86.1% and specificity
of 87.7% for detecting insomnia cases in a community sample (Bastien, Vallieres, &
Morin, 2001; Morin et al., 2011). It has excellent internal consistency (Cronbach’s α
>0.90 ;Morin et al., 2011). A limitation is that the ISI asks the participants to rate “their
sleep problems,” which may be problematic in general population studies where many
participants do not think of themselves as having a sleep problem.
The Women’s Health Initiative Insomnia Rating Index (WHIIRS) is a five-item in-
strument (scores 0 to 20) that asks participants to rate the quality of their sleep and
frequency of the occurrence of the cardinal symptoms of insomnia over the prior four
weeks: trouble falling asleep, nighttime awakenings, and early morning awakenings
(Levine et al., 2003). The instrument was developed and validated in large samples
of postmenopausal women where it was shown to have high internal reliability
(Cronbach’s α = 0.78) and test–retest reliability (0.96 for same-day; 0.66 after one or
more years). A sample of over 400 women studied using actigraphy was used to test
the ability of the WHIIRS to discriminate individuals likely to have insomnia based on
an elevated sleep onset latency (>30 minutes) and low sleep efficiency (<85%; Levine
et al., 2005). In this sample, a threshold WHIIRS score of 9 had a sensitivity of 0.53
24 An Introduction to Sleep Epidemiology
and specificity of 0.67 to distinguish individuals with and without these actigraphic
markers of poor sleep; the modest diagnostic accuracy is likely partly attributable to
the limitations of actigraphy to serve as a gold-standard insomnia outcome. While the
development and validation of this instrument was on populations of older women,
the content reflects the core criteria for insomnia. Nonetheless, additional research
in more general samples, use of other criteria for evaluating diagnostic accuracy, and
further assessment of cutoff values for identifying individuals likely to have insomnia
are needed.
questions on cardinal symptoms of OSA (snoring, tired, observed apneas) and OSA
risk factors and disease correlates (high blood pressure, obesity, older age, increased
neck circumference/collar size, gender). Sensitivity and specificity in relationship
to laboratory-diagnosed OSA varies according to the threshold score considered
to be abnormal and to the pretest probability of the sample (Nagappa et al., 2015).
The Berlin Questionnaire is also commonly used. It includes 10 questions across
three domains: snoring, excessive sleepiness, and comorbidities (hypertension or
obesity). Its psychometric performance varies according to the populations studied
and “gold standard” definition of OSA (i.e., AHI >5 vs. AHI >15; Chiu et al., 2017;
Netzer, Riccardo, Netzer, Clark, & Strohl, 1999). A detailed review of OSA screening
instruments is found in Miller and Berger (2016).
A limitation of many OSA screening instruments is that they were developed and
validated in clinical settings and thus may not generalize well to population samples,
including women and individuals whose symptoms do not rise to threshold levels to
trigger clinical evaluation. In an analysis of 12,158 participants from the Hispanic
Community Health Study, Shah and colleagues (2016) considered whether a broad
array of symptoms and conditions and consideration of gender-specific associations
would improve OSA prediction. Seventeen candidate predictors were considered in
development and validation models, including demographic factors, snoring, BMI,
waist and neck circumferences, hypertension, diabetes, heart disease, and symptoms
related to sleep quality (Shah et al., 2016). Four items—age, BMI, gender, and snoring—
modeled using gender-specific coefficients, were found to provide the overall best pre-
dictive value with a sensitivity of 0.77 and a specificity of 0.75. A rigorous evaluation of
screening questions for OSA in pregnancy similarly found that an equation using only
demographics and snoring data predicted OSA in pregnancy, with C-statistics ranging
from 0.809 to 0.870 (Louis et al., 2018).
The Calgary Sleep Apnea Quality of Life Index is designed to assess OSA-specific
quality of life, including responsiveness to OSA treatment. This instrument correlates
with global quality of life measures, has a high two-week test–retest reliability (r = 0.92),
and a large responsiveness index (1.9) and effect size (1.1) to OSA therapy (Flemons &
Reimer, 1998; Silva, Goodwin, Vana, & Quan, 2016).
Sleep Duration
Self-reported sleep duration is one of the most commonly reported sleep measures
in epidemiological studies. Unfortunately, there is large variation in how sleep du-
ration is ascertained, with no standard, recommended approach, contributing to
noncomparability of findings across many studies. Some questions ask individuals to
estimate their total sleep in the night while others ask about sleep over a 24-hour period.
Some studies estimate usual sleep duration using a single question (e.g., “How many
hours of sleep do you usually get at night [or when you usually sleep]?”). Responses
use categorical response scales (e.g., <6, 7–8, 8–9, >9) or free text, with room to record
hours and minutes in numerical format. Use of categories displayed as whole integers
facilitates data management but reduces resolution, potentially obscuring physiologi-
cally important quantitative differences in sleep duration. Alternatively, some studies
ask participants to record the clock times corresponding to bed or sleep time and wake
time. A problem that frequently arises with those questions is that when individuals
use clock times they may confuse AM and PM, necessitating stringent data clean-
up procedures. Some surveys include separate questions regarding weekday (work/
school) versus weekend (“free,” weekend), which allows derivation of measures of
sleep timing and quantification of variation in weekday and weekend sleep (“social
jetlag” is the difference in sleep timing between weekday and weekend).
Choice of which sleep duration question(s) to use often reflects the trade-offs
investigators make after considering the types of biases of most concern for given
studies, the “space” allotted for sleep questions on a given survey, and whether there
is interest in sleep timing (chronotype, social jet lag) as well as duration. For ex-
ample, a single question regarding habitual sleep duration may be influenced by
reporting of socially desirable sleep durations (if the participant is aware of public
health messages to sleep for 7 to 8 hours). Individuals often know their bed and
27 Sleep Epidemiology: An Introduction
wake times, and reporting these may be less influenced by social desirability bias
and may minimize errors participants make when subtracting their own bed and
wake times to derive sleep duration. On the other hand, reported bedtime does not
reflect actual sleep onset—which can be difficult to estimate—and can be differ-
entially reported by individuals with sleep state misperception, a characteristic of
dyssomnias. A limited amount of research has compared sleep duration estimated
from different questions. Jackson and colleagues recently reported a detailed anal-
ysis comparing self-reported habitual sleep to sleep duration based on differences
between self-reported bed and wake times in the Jackson Heart Sleep Study. The
authors found that compared to actigraphy-assessed sleep, self-reported habitual
sleep duration from a single question underestimated sleep duration by an average
of 31 minutes, while the difference between bed and wake times overestimated sleep
duration by an average of 45 minutes. Moreover, biases were greater in individuals
with insomnia. These findings underscore the critical importance of interpreting
sleep duration data in reference to the specific ways this information was ascertained
and considering sources of bias.
Several studies compared self- reported sleep duration to actigraphy- based
estimates. In substudies of large community-based samples, including Coronary Artery
Risk Development in Young Adults (CARDIA), Hispanic Community Health Study
(HCHS), and Multi-Ethnic Study of Atherosclerosis (MESA), correlations between
self-reported sleep duration and actigraphy-based assessments were weak to mod-
erate: 0.28 to 0.47 (Cespedes et al., 2016; Lauderdale, Knutson, Yan, Liu, & Rathouz,
2008; Lauderdale et al., 2006). The strength of these associations varied by a presence/
absence of symptoms of sleep disturbances, gender, and population group. Several
studies showed weaker associations among Black compared to White participants (e.g.,
in CARDIA, rs = 0.56 for Whites, 0.29 for Blacks; in MESA, rs = 0.45 for Whites, 0.28 for
Blacks, 0.38 for Hispanics/Latinos, and 0.35 for Chinese; Chen et al., 2015; Lauderdale
et al., 2008). Differences in associations across racial/ethnic groups may reflect socio-
cultural differences in how questions are interpreted as discussed in chapter 7, as well
as between-group differences in the accuracy of the reference actigraphic measure-
ment (see later discussion). In addition, some groups, such as Blacks, have higher var-
iability in sleep duration across the week, and this variability—which is not captured
by single questions—may account for the poorer ability of a question that asks about
“average” behaviors to quantify a time-variable activity. The direction of biases across
studies also varies, which likely reflects the differences in study-specific instruments
and underlying cohort characteristics. For example, self-reported habitual sleep dura-
tion was overestimated by 48 minutes compared to three-day actigraphy-based meas-
ures in CARDIA, while in the Jackson Heart Study Sleep Study, sleep duration was
31 minutes lower when measured by self-report compared to seven-day actigraphy
(Jackson, in review). When using bed–wake time differences, self-reported sleep du-
ration was overestimated compared to actigraphy by over one hour in both the HCHS
and MESA. These large differences likely reflect the influence of sleep latency, which
are not included in the calculation of bed–wake time differences.
28 An Introduction to Sleep Epidemiology
Sleep Diaries
Daily sleep diaries (using paper or electronic format) allow examination of the timing
and day-to-day variation of sleep, providing estimates of chronotype, sleep duration
and timing variability, and social jetlag. Some diaries also ask about exposures (e.g.,
caffeine, alcohol, hypnotic use), daily activity (work, vacation), and mood, providing
opportunities to explore the influence of substance use, location, mood, and social
surroundings on day-to-day variation in sleep. As an example, Spadola and colleagues
used daily sleep and exposure data from the Jackson Heart Sleep Study to estimate the
impact of nightly tobacco and alcohol use on sleep duration and efficiency. Despite
the temporal resolution possible with use of diaries, using these instruments is more
burdensome that using simpler, one-time administered questionnaires, reflects only
sleep behaviors during the limited period of monitoring, and can require substantive
data cleaning due to missing data and erroneous data entries. Completions rates vary
according to the setting and participant motivation. With increasing use of app-based
health questionnaires, electronic sleep diaries and email prompts appear to be prom-
ising modalities for improving participant response as well as data quality in groups
who have access to this technology.
To improve standardization of sleep diaries, the American Academy of Sleep
Medicine assembled an expert group and conducted focus groups with patients,
exploring which items and formats appeared to be most resonant and acceptable for
characterizing daily sleep. This effort resulted in the Consensus Sleep Diary consisting
of a core of nine items (focused on sleep–wake times and sleep quality), with expanded
versions to capture medications and other exposures (Carney et al., 2012). Additional
research is needed on the role of sleep diaries in research, including understanding
response bias and how to enhance data completion rates, as well as the role of paper
versus electronic versions and ways to optimize the user experience (e.g., use of scroll
wheels vs. “time pickers”).
Actigraphy
Wrist accelerometry (also referred to as “actigraphy”) is increasingly used in epi-
demiological studies to provide low-burden, objective estimations of sleep–wake
patterns. Worn over multiple days (and sometimes months), accelerometers measure
digitized acceleration signals that are analyzed using specialized software to classify
each consecutive time period (or “epoch,” typically 15 to 30 seconds long) as “sleep”
or “wake.” These sleep–wake epochs are summarized to estimate overall sleep and
wake durations (and their derivatives, such as sleep efficiency) across each sleep pe-
riod and recording period. Advantages over questionnaires include (a) elimination
of recall and reporting biases, (b) ability to characterize sleep timing and day-to-
day variability through prospective time-stamped data collection, and (c) ability to
provide estimates of the impact of wakefulness on sleep latency and continuity (i.e.,
sleep onset latency, sleep efficiency, sleep fragmentation, and wake after sleep onset
time). Circadian rest–activity rhythms also can be estimated from multiple days of
29 Sleep Epidemiology: An Introduction
Polysomnography
The gold standard for measuring sleep is with in-laboratory overnight polysomnography,
which records multiple physiological signals simultaneously and includes the sig-
nals needed to (a) stage sleep: EEG, chin electromyograms, and electro-oculogram;
(b) characterize and quantify sleep-disordered breathing: pulse oximetry, respiratory
excursions, and airflow; and (c) identify periodic limb movements: leg electromyog-
raphy or movement. In addition, many montages include electrocardiographic, body
position, snoring, and other sensor data.
PSG enables each 30-second epoch of sleep to be classified into a stage: N1, N2, and
N3 refer to lighter (N1, N2) and deeper (N3, formerly considered “slow wave sleep”)
nonrapid eye movement sleep, and rapid eye movement (REM) sleep. This yields the
ability to characterize
30 An Introduction to Sleep Epidemiology
• AHI: number of apneas plus hypopneas per hour of sleep. This index is based on
tabulation of
• Number of apneas (cessation of airflow for 10 seconds or more), with individual
events classified as “obstructive” (persistent respiratory effort) or “central”
(absence of respiratory effort)
• Numbers of hypopneas (reduction in airflow for at least 10 seconds during
sleep accompanied by desaturation or an arousal)
• Sleep-associated oxygen saturation patterns: minimal, average, time below
threshold levels of saturation, and event-related hypoxic burden
Periodic limb movements are characterized as the numbers of leg movements that
occur in defined clusters, with and without associated EEG arousal.
Although comprehensive, PSG can be burdensome and costly. It is impractical to use
over multiple nights, potentially reducing the reliability of assessments due to single-
night measurement. In addition, “first-night” effects (i.e., poorer sleep occurring on
the first night assessment due to lack of habituation to the sleep laboratory environ-
ment) may result in underestimation of the proportion of “deeper” sleep.
Advances in technology allow both full PSG (referred to as Level 2 PSG) as well
as more limited studies (Levels 3, 4) to be conducted in at-home settings. The Sleep
Heart Health Study was the first major epidemiological study to use Level 2 PSG to
characterize a full range of EEG, breathing, and heart rate signals over one night. In
this study of over 6,000 individuals from across the United States, study failure rates
were low and there was minimal first-night effects with high night to night reliability
(intraclass correlation coefficients [ICCs]: 0.77 to 0.81) for the AHI and most other
sleep parameters, with the exception of time in REM, where the ICC was 0.37 (Quan
et al., 2002). Measurements also agreed well with in-laboratory PSG, with differences
largely attributed to more representative sleep in the home (less time supine, more
time in stage R sleep; Iber et al., 2004).
More limited monitoring devices, such as home sleep apnea tests, which measure a
few channels of respiratory-related data, also have been demonstrated to provide reli-
able estimates of sleep apnea at relatively minimal burden. It is important to note that
there are multiple types of home sleep apnea tests that have varying sensors, hardware,
and software and thus differ in their failure rates, signal fidelity, and content. Single
Another random document with
no related content on Scribd:
quarrelling with the chief, presently accepted the latter's invitation to stroll
round the camp and visit the pickets. For Te Karearea observed all proper
military precautions, and maintained an iron discipline in camp and field.
'If you are hinting at me, I have no intention of trying,' was George's
reply to this suggestive remark. 'But why are you so anxious to detain me?'
'Why are you so anxious to leave me, my friend?' countered the Maori,
and, as George burst out laughing, 'I have not treated you ill, Hortoni,' he
added rather wistfully.
'True. Still, you talk as a fool. Home, friends, duty, inclination, all call
me away from you. You are in arms against the men of my race. Is it any
wonder that I fret in the toils?'
'Yet there are chiefs who have their Pakehas,' urged Te Karearea.
'That is not much to the credit of those Pakehas,' George said loftily; and
to change the subject went on: 'Where is Paeroa?'
'Be wise in time, Hortoni,' the chief urged earnestly. 'You possess,
though you do not realise it, a certain means of attaining greatness. Ascend
the ladder which I am holding for you, and you will be great. Refuse, and
you are doomed, even as your race is doomed. You ask for Paeroa. He is
gone to carry the message of my coming.'
'Say rather, who will not hear my word?' Te Karearea drew himself up
proudly. 'Waikato and Ngatiawa shall hear and flock to my standard.
Taranaki and Wanganui shall lift the spear and shake the tomahawk. Taupo
and Ngaiterangi, Whakatoea and Ngatiporou, Ngatiapa and Ngatihau[1]—
all these and more shall hear and come with club and gun. But Arawa, the
accursed, shall be deaf, and them and the Pakehas shall my legions smite
and slay until the land which has been ours since Maui drew it forth from
the sea, is ours once again. Behold! I, Te Karearea, have sworn it.'
[1] All the Maori tribes named above were in arms against the British at
one time or another during the wars. The Arawas were friendly.
The sonorous cadence of vowels rolled out into the night, and George, to
his surprise, felt a passing throb of sympathy for this uncrowned king. After
all, the land had originally—and not so long ago—belonged to the Maori;
nor could the Pakeha be said to be altogether clean-handed in the matter. It
was a fleeting mood; but it sufficed to induce George to let the chief down
gently, and to refrain from further argument.
Just then the rapid beat of a horse's hoofs was heard, and Te Karearea,
with a word of excuse to George, ran back to the sentry they had just
passed, whispered an order, and at once rejoined his guest, as he was
pleased to style his paroled prisoner.
'During the afternoon I learned that the captain of the force opposed to
me sent to Turanga for reinforcements,' he began, smiling. 'This, in all
probability, is the messenger returning. I am going to catch him.'
'But,' objected George, 'if the messenger recognise that the sentry is not a
"friendly," he will bolt, and then your man will certainly shoot him.'
'It takes some education for a Pakeha to distinguish, let us say, Arawa
from Ngatiawa,' said Te Karearea reassuringly. 'No; there will be no
difficulty—of that sort.' He paused to whisper instructions to a sentry on the
inner ring, and George, glancing back, saw that the messenger was slowly
walking his tired horse towards the picket.
'I must ask you to retire, Hortoni,' said Te Karearea courteously. 'I must
examine this man, and——'
'Oh, quite so,' agreed George. 'The poor beggar little dreams what is in
store for him. When your interrogation is at an end, turn him over to me,
and I will do my best to console him.' He nodded to the chief and turned his
back upon the bivouac, thinking as he went of the grim jest which Fate was
about to play upon the unlucky messenger.
Hoping to get a bit of news on his own account, George strolled towards
the outer picket, and in course of time was challenged by the sentry in the
strictly orthodox manner: 'Halt! Who goes there?'
George explained, and handed the sentry a plug of tobacco, off which the
Maori promptly bit a piece. But he was a surly fellow, and gave a gruff
negative when asked if he happened to know anything of the Pakeha who
had ridden into the camp.
'They will eat the oyster and throw away the shell; that's all I know,' he
growled, his answer showing that he came from the coast.
'Meaning, I suppose, that they will turn him out of the camp when they
have learned all that he has to tell,' commented George. 'I should like a
word with him before he goes. I wonder if he will come this way.'
'Whakatore Atua!' (the gods forbid) ejaculated the sentry, with a nervous
glance over his shoulder. 'Let him take another road to Reinga. I want no
ghosts on my beat.'
'I have come to the wrong place, it seems,' he said cheerfully, little
imagining how true were his words. 'It is Captain Westrupp's bivouac I'm
after. Well, boys, I suppose you licked those rascals?'
'Yes; we licked them,' answered Winata Pakaro in fluent English, while
his leader remained unobtrusively in the background. 'They are now in full
retreat.'
'Hurrah! Well, I must hunt up the captain. Where is his bivouac?' He cast
a longing eye upon the cold viands, scattered about.
'Nay; sit and eat,' invited Winata Pakaro. 'You need food after your long
ride. The captain is not in the camp, nor is it likely that he will return to-
night.'
'Oh, in that case, here goes'; and the young soldier sat down and ate with
appetite, while Winata Pakaro pumped him dry of information as to the
number and disposal of the British and Colonial troops. The meal and the
interrogation ended together.
'Thank you, boys; you are the real old sort,' said the messenger
gratefully. 'Now tell me where my mates are camped. It is odd that none of
them are about; but I suppose they are all dog-tired.'
He turned to go, smiling at them; but at a sign from Winata his arms
were pinioned, and while a couple of Maoris held him in a firm grip, a third
lashed his ankles together.
He was very strong, that was evident; but he was intelligent too, and did
not waste his strength in useless struggles. 'You crafty demons!' he snarled
at them. 'You are Te Karearea's men.'
'Yes,' admitted Winata Pakaro,' and we are also brothers of the men who
died to-day. So there is a blood-feud, and, as we have you, you must die.'
'Oh, we will do all things as they ought to be done, and follow the rules
of war. You come by night into our camp, pretending to take us for
"friendlies," and endeavour to worm information out of us. Thus you are
proved a spy. It is the custom of civilised nations at war to hang spies.
Good! We will hang you, and so escape the vengeance of the Pakeha.' His
saturnine chuckle was echoed by the chiefs who stood in a semi-circle
about the prisoner.
The unhappy soldier looked round despairingly. What hope was there for
him? Before him a crescent of stern-faced men, and all about him men of
the same colour, with faces yet more fierce and horrid. For the rank and file
had gathered to hear the last of the discussion—to see the last of the
Pakeha.
He was a soldier, and he would not show the despair he felt to these
enemies of his flag. 'I warn you that a terrible vengeance will be taken if
you murder me,' he said boldly.
A derisive yell arose among the bystanders, and at a covert sign one of
the executioners drew the rope taut, handing the loose end to the other.
The miserable messenger gave up hope. He was brave, and he did not
mean to go out of the world like a craven. But it was hard, for he was young
and strong, and life glowed in his veins. He cast an agonised glance around,
but only savage, grinning faces met his eyes. He closed them, murmuring a
prayer, when a shout, not far off, struck his tense nerves with such a shock
that they quivered, like harp-strings suddenly smitten, and for the first time
he trembled—not with fear, but with hope.
Again that shout, loud and insistent, crying something in Maori which he
could not understand. Yet when he heard it, he trembled all the more, for
there was something in the voice which rang familiar in his ears. Yet how
could that be?
Once more the frantic appeal: 'Kei whakamate ia koe!—Do not kill him!
Do not kill him!'
Stamping footsteps, crushing down the rustling fern—nearer, louder,
furious at the feeble opposition. And at last a man, panting, sobbing for
breath, burst into the open space illumined by the bivouac fire, gasping as
he came his ever-recurring 'Kei whakamate ia koe!'
For one instant the soldier stared, incredulous. He seemed paralysed. His
eyes started from his head. His limbs shook under him. Suddenly he felt the
tightening noose, stiffened, caught at a hasty breath, and spent it in a
quavering shriek: 'George! Quick! They're murdering me!'
The two Maoris with the rope set off at a run. But ere the cord could
press the swelling throat, George Haughton crashed through the encircling
crowd, tumbling them this way and that; and, as he charged down upon
them, whirling the mysterious mere over his head, the executioners dropped
the rope and fled for their lives, howling.
In an instant George was at his friend, plucked the cruel rope from his
neck, and flung it away. Then pushing Terence behind him against the tree,
he stood on the defensive, eyes glaring, but keen; his chest heaving from his
run; challenge and menace in every line of him.
CHAPTER X
TOGETHER AGAIN
When the Maoris recovered from the shock of his rush, they faced
George as he stood covering his friend's body with his own. There was no
noise, no shouting; but the stern Roman faces looked very grim and
determined. Then Winata Pakaro with oily tongue began an argument, in
the midst of which was heard the click of the hammer of a gun drawn back
to full cock.
But while Winata's smooth periods flowed on, there was a sudden rush, a
scuffle, a shout of wrathful surprise, and there was George back again under
the tree with the rifle in his hand. He had wrested it from the astonished
warrior who had so stealthily—as he imagined—made ready to use it.
In another moment Terence's wrists and ankles were free and the rifle in
his grasp, while George once more flourished his famous club, rightly
judging that its moral effect would be considerable, while as to its physical
possibilities there was no doubt whatever.
In another moment Terence's wrists were free,
and the rifle in his grasp (page 106).
These things done in the space of a second or two, George began to
harangue the Maoris, but Winata Pakaro cut into his first words with:
'Stand aside, Hortoni! We wish not to injure you; but this man must die.'
But a long whistle shrilled from Winata's lips, and he flung himself flat
on the ground as the Maoris made an ugly rush forward and Terence's rifle
spoke.
Fortunately for the friends, the bullet merely startled an elderly chief into
a most undignified caper as it hummed past his ear, and on the instant
Winata leaped from the ground and hurled himself at Terence.
But the great fighter was handicapped by his fear of George, whose own
weapon came more than once so dangerously near his head that he gave
back in alarm; for there was no knowing when that magical piece of
greenstone would spring out of its master's hand and begin a devil's dance
upon its own account.
Still, it might have gone hardly with Terence, but that, as the Maoris
surged about him, a deep voice cried angrily: 'Is the word of Te Karearea of
no weight in this hapu? And you, Hortoni, why do you break faith with
me?'
Where the chief had sprung from George had no idea, but he was
uncommonly glad to see him, and, as the Maoris shrank back, he briefly
explained who Terence was and what had taken place. Thereupon Te
Karearea turned upon Winata Pakaro and rated him viciously, demanding
how he had dared to take so much upon himself. To this verbal castigation
Winata merely opposed a smile of cynical amusement as he walked away.
Then Te Karearea faced George once more and said graciously, 'I give
you, O Hortoni, the life of the friend for whom you would have given your
own. To-morrow you shall tell me the story of your friendship. But he must
give up the rifle.'
With his arm round Terence's shoulders, George piloted his chum
towards the huge fallen tree, beneath which he designed to pass the night.
'You dear old fellow!' he said heartily, drawing Terence to him. 'Who would
have thought of meeting here, and like this? What a mercy I came up in
time!'
'Thank God you did!' replied Terence, unable to repress a shudder; for
when a brave man has stared death in the face, and the grim Gatherer has
passed on, leaving him untouched, he is not, as a rule, flippant about his
experience.
'When I came rushing up, I hadn't the faintest idea that I was to meet
you,' went on George. 'Indeed, I only recognised you when you screamed at
me in that queer, cracked voice. In the first place, I had never seen you in
uniform, and in the—— Hold up, old fellow!'
For Terence staggered and would have fallen, had not the strong arm
around his shoulders slipped to his waist and supported him.
George laid him down and bent anxiously over him, seeing that he had
fainted. The strain had been dreadful, and, brave though he was, his
emotional nature had lent an added poignancy to the sufferings of that
terrible half-hour.
At this Terence gasped in a queer way and collapsed flat upon his face;
but when George, who thought that he had fainted again, was about to rise,
he scrambled to his knees, and catching his friend in a bear's hug, exclaimed
brokenly: 'Oh, thank God! O George, thank God I didn't hit you! Oh!'
He buried his face in his arms, while George patted his broad back,
saying soothingly: 'So it was you after all! Curiously enough, I thought so at
the time; but I did not see how such a thing could be. Cheer up, old fellow!
There's no harm done.'
Terence wrung his friend's hand. There were traces of tears upon his
cheeks, but he did not seem to mind. 'I took deliberate aim at you,' he said.
'We all thought that the white man on the hill must be one of those Pakeha-
Maori rascals; so I let drive and——'
'Missed him! So that's all right,' finished George cheerfully. 'You must
not let out to these people that you were the slayer of their comrades, or we
shall hear a lot about a blood-feud and have endless trouble. By the way,
was Te Karearea present at your court-martial?'
'I saw nothing of him until he stopped our little fight. Why?'
George did not explain. He had reason to know that the chief did not
always choose to appear as the moving spirit in the programme of events.
'No matter,' he said. 'Now, I want to hear all your news. Are you hungry?'
'Oh no; your friends fed me well before turning me over to the hangman.'
'You are yourself again, I see,' said George, laughing. 'Fill your pipe and
let me hear your adventures.'
'I have had none until to-day,' began Terence. 'Colonel Cranstoun was
very kind to me on board; but he and Horn kept me at it with never-ending
drill. By the way, the colonel expressed his regret that he had packed you
off in the tug.'
'Yes. He pulled his long moustache, and observed: "I should have done
better to keep the young scapegrace under my own eye."'
'He is a fine old fellow, George. On the third day out we met a Sydney-
bound brig, which hove to, and the colonel sent a letter to your father. You
saw it, no doubt.'
'No; but I am glad he wrote it. I started on your trail next morning.'
'What a fellow you are!' said Terence admiringly. 'I was sure that you
would lose no time. But next morning!'
George handed over the club, which Terence examined with deep
interest.
'It seems to me,' he said at last, 'that you will do well to take that old
man's advice and hold fast to this club; for——'
'Explain, then, its return after your own eyes had seen it falling into the
sea.'
'Say, rather, that, like all your unimaginative race, you refuse to believe
in anything you cannot understand. If there is nothing exceptional about the
club, why is Te Karearea so anxious to get it?'
'It is, of course, surrounded with traditions,' began George, and suddenly
sprang up and darted round the tree in time to see a dark figure bounding
away into the bush. Pursuit was useless, so George returned to their fire,
expressing his conviction that the eavesdropper had been Te Karearea.
'As I said, he attaches importance to the club, if you don't,' was Terence's
comment.
'More likely he came here to learn what he could about you,' George
argued; 'for I don't believe in his protestation of ignorance of English.'
'All the same, you follow the old man's advice, and never let that club be
far from you,' urged Terence.
'Well, it is a singular fact that the moment of my greatest peril was just
after I had been deprived of the mere,' admitted George.
'Yes; and I do not understand why. From the moment of our meeting, Te
Karearea has treated me with great consideration, and—though it may
sound absurd—has sometimes seemed afraid of me. Not, of course, in a
physical sense. There is something incomprehensible at work.'
'True. The great thing is the plain fact that association with the club has
saved your life so far. I think——' He yawned widely.
'I think that you are more than half asleep,' finished George. 'Your bed is
there, under the tree, and here is a blanket for you.'
'Brute that I am, I have never even thanked you. But you know, old
fellow—my dear old chum, you know——' He paused, blinking hard.
'I should think I did!' cried George, capering with the pain of that
friendly squeeze. 'Brute! You are indeed. A grizzly isn't in it with you.
Away with you to bed, and don't talk any more nonsense.'
'I won't,' said Terence seriously; 'but I will do at last what I ought to have
done at first.' Without a word more he dropped upon his knees and buried
his face in his hands. A few minutes later he rose quietly, and with a nod at
George, lay down upon his fern-bed and prepared to sleep.
CHAPTER XI
'Oh, the soldier fashion in which we do things here would not disgrace
the "old regiment," as you call it,' said George, smiling. 'Your own red coat,
by the way, has a suspicious newness about it. Did you sleep well?'
Soon they were again upon the march, and Te Karearea, who had taken
every precaution against a surprise, jogged peacefully along, smoking a
looted cigar, and listening with interest to the story of the youthful
adventures of George and Terence, whom he addressed as Mura, or The
Blazing One. The name had much the sound of Moore, but it was the
appearance of the Irishman, with his red coat and flaming head of hair,
which had really suggested the title.
'It is good to hear of such friendship,' the chief said, beaming upon the
pair during a pause in their narration. 'Surely Mura will not wish to leave us
now that he has found you, Hortoni. Persuade him to stay, my friend.'
'What did he mean by that remark, George?' inquired Terence as they lay
in the shade during the midday halt.
'I can't say exactly, for one never knows what the crafty beggar is up to.'
He looked cautiously round, but as no one was near, went on: 'He may even
wish you to try and escape, in order to—to——'
'To find an excuse for knocking me on the head,' supplied Terence. 'Then
he'll be disappointed, for I'll not leave you—unless escape meant a good
chance of helping you out of the trap. In that case I'd go this minute.'
'I am sure you would, dear old fellow!' said George affectionately; 'but
we will stick together as long as possible. Only, if the chief does not parole
you, then——' He broke off short, staring up at Te Karearea, who had, as
usual, approached unobserved.
'It looks as if the rascal possesses the power to render himself invisible at
will,' said George disgustedly, when the chief had withdrawn after
informing them that the march was about to be resumed. 'We shall have to
go warily, Terence; for there is no knowing how much he may have heard.'
George smiled. 'Oh, I know what you can do; but a couple of hundred to
one is long odds. Meantime, you must not run the risk of offending him; for,
remember, he is utterly unscrupulous. In some mysterious way I appear to
be necessary to him; but were it otherwise, he would kill me without the
slightest compunction. Of that I feel sure. Come! it is time we joined him.'
Four days later, towards sunset, they debouched from the forest through
which for the last sixty hours they had toiled wearily along a narrow,
difficult track. It had been a terrible journey for the Maoris, but far more so
for the white men, and all alike rejoiced when at last the dreadful bush lay
behind them, and they beheld the river which alone divided them from the
pah which was their goal.
As was usual with the Maori fortresses, the position was one of immense
strength. The island plain, at the back of which rose a considerable hill, was
a swampy area overgrown with flax, and extended for nearly a mile on
every side of the eminence but one, being itself enclosed by a forked ravine,
at the bottom of which the river roared and swirled among giant boulders.
No doubt, at some far-off day this roughly level plain had itself been
covered with forest; but dead and gone generations of Maoris had cleared
away the offending wood, so that no one could now approach the pah
unobserved. The single side of the hill unflanked by the plain was simply a
vast, precipitous rock-face, having for its vis-à-vis the equally precipitous
opposing wall of the ravine, into whose depths it dropped a sheer two
hundred feet, the twin cliffs forming a cañon through which the river raced
on its way to unite again with the main stream.
The place was, indeed, almost inaccessible when once the only approach
from the forest was barred. This was merely a rough bridge across the river
on the side furthest from the hill, and when the tree-trunks forming this
were withdrawn, a handful of men could easily hold the island against an
army.
But even were the bridge to be rushed, the ascent of the hill was made
difficult by carefully laid trenches and rifle-pits, and, finally, the pah was
encircled by a double row of palisades of great height and immense
strength, the chinks between the massive logs being filled with hard-baked
mud and clay. The palisades were loopholed above, and a rude platform ran
along the inner side of each row, where men might lie, secure themselves,
and fire upon an advancing foe.
It would indeed be a desperate and determined foe who would venture to
attack, much more succeed in taking, the Pah O Te Mate—the Pah of the
Slain, the Fortress of Death.
As it happened, the weary travellers were not destined to enter the pah
just yet; for as the vanguard swung out of the forest and prepared to cross
the hundred yards or so of cleared ground between them and the bridge,
they saw a sight which halted them as effectually as though some sudden
stroke had robbed their limbs of all power.
But they could not stand still, for those in the rear pressed them on, and
presently the little clearing became almost blocked with armed men vainly
striving to preserve their customary proud and resolute bearing, and with
trembling women who did not attempt to hide their extreme terror.
In the midst of the confusion the voice of Te Karearea was heard angrily
demanding the cause of the block, but no sooner had the chief forced his
way to the front than he, too, stopped as if compelled, all signs of anger
faded from his face, and he stood meek and inoffensive, his hands crossed
upon his broad chest, his plumed head bowed low.
And what was the cause of all this fear and commotion? Standing alone
at the bridge-head was one old man. His figure was bent, his snow-white
hair fell, a tangle of locks, below his shoulders, and the hand which grasped
the staff upon which he leaned, trembled as it clutched the crook. Yet there
was fire in his rolling eyes, and a hint of mastery, if not of menace, in the
gesture with which he flung up his free hand, forbidding the advance; and
his voice, far from piping in the thin treble of extreme old age, rang stern
and sonorous, as the liquid Maori speech gushed from his venerable lips.
'Which means that the ancient rogue has reasons for keeping us off the
hill to-night.'
'Very likely; but it won't hurt us, fortunately. What do you say to supper
and early bed?'
'I'm with you there,' agreed Terence, 'for I'm dog tired.'
So they hunted about until they found pretty Miss Kawainga, who soon
provided them with an excellent meal, after which they selected a
comfortable spot for their bivouac, spread their blankets on the fern, and
were quickly asleep.
From afar the horrid noises screamed through the shivering forest, mixed
now and again with a singular gabble of words which somehow had the
sound of English, though the distance made it difficult to judge.
George made a fierce effort to collect himself. Terence had suffered
enough already, and for his sake he must not give way. But to his intense
surprise he saw the object of his concern sitting up and listening with an
expression of deep interest on his face.
'Queer row, isn't it?' said Terence. 'Do you see those lights on the hill
behind there? That is where they are. Perhaps this explains the mysterious
confab between the chief and the wizard. I vote we go and have a look at
them; we may never get another chance.'
George could scarcely believe his ears. The noise which now, as before,
so shocked him, was accepted by Terence as something merely interesting.
Still, the sight of his friend's unconcern did much to steady his own jumping
nerves.
Receiving no answer, Terence looked up. The dying fire added to the
ghastliness of George's face. 'Hullo! What is it, old fellow?' he cried, rolling
out of his bag. 'Are you ill?'
'I plead guilty to a bad fit of the horrors,' answered George, 'though your
coolness is rapidly convincing me that my bogy is not so awful as I
imagined it to be. I never was so frightened in my life as when I first heard
those terrible sounds at dead of night aboard the brig. I did not speak of it to
you when we met, because it had nothing to do with my story. If you know
what the noise means, for heaven's sake tell me at once.'
'I thought you knew all about it,' replied Terence. 'The row is horrid, but
simple enough in its origin. It is a part of the religious service, or
incantation, perhaps I should say, of the Hau-haus.'
'Men, distinctly; but with a strong dash of the devil in them, too.'
'Very much so. The same among whom you have been adventuring this
month past. Let us steal back to that hill and lay your ghost for once and all.