Dimensions of Daytimes Sleepiness in the General Population

1
Dimensions of Daytimes Sleepiness in the General
Population (Progress in Measuring Sleepiness
a Multifactorial Condition) Hyon Kim, PhD April
5th, 2004 Population Health Sciences Seminar
2
(No Transcript)
3
Three Mile Island 1979
Bhopal 1984
Exxon Valdez 1989
Chernobyl 1986
4
(No Transcript)
5
Are we getting enough sleep?
National Sleep Foundation, Sleep in America Poll
(2001) Random sample, telephone interview,
n1004, 18 yrs
6
Sleepiness and outcomes?
National Sleep Foundation, Sleep in America Poll
(2001) Random sample, telephone interview,
n1004, 18 yrs
7
Sleepiness
A transitional condition between wakefulness and
sleep
Wakefulness
Sleep
Wake drive
Sleep drive
Modifiers

• Intrinsic
• Sleep Fragmentation
• Sleep Apnea
• Narcolepsy, Idiopathic CNS disorders
• Co-morbid conditions

• Extrinsic
• Voluntary sleep deprivation
• Phase alteration/Circadian rhythms
• Drugs

8
How do we measure of Sleepiness?
Objective sleep/wake behavior
Subjective self-reported somnolence

• Electrophysiological
• Multiple Sleep Latency Test
• Maintenance of Wakeful Test
• Pupillometry
• Behavioral Observations
• Yawning frequency
• Eye lid closures
• Performances tests
• Reaction time tests
• Driver simulators

State (instant)

• Self-reported Rating scales
• Stanford Sleepiness Scale
• Visual analog

• Global level of sleepiness
• Self-reported sleepiness problems
• Self-reported sleep propensity
• Epworth Sleepiness Scale

Trait (stable)
9
Subjective-state sleepiness

• Rating Scales (immediate feelings)
• -Stanford Sleepiness Scale (SSS)
• 1. Alert and wide awake.
• 4. A little foggy.
• 7. Almost asleep.
• -Visual Analog

X
Very alert
Very drowsy

Advantage good for short-term sleep experimental
sleep deprivation studies, Disadvantage Poor
instrument for longitudinal studies
10
Objective-state sleepiness
Psychomotor Vigilance Test (Reaction time)
Multiple Sleep Latency Test (Nap Studies)
Advantages Free from subjective
noise Disadvantage Time consuming expensive
11
Subjective-trait sleepiness

• Global level of sleepiness
• Self-reported sleepiness problems
• Feeling unrest regardless of sleep amount
• Difficulty getting up in the morning
• Feeling excessive daytime sleepiness
• Need for stimulants
• Self-reported sleep propensity
• (Epworth Sleepiness Scale)
• Chance of dozing off in recent times in 8
  situations
• Sitting and reading
• Watching TV
• Sitting inactive in a public place
• Riding in a car
• Resting
• Talking to someone
• Sitting quietly after lunch
• Stopped traffic

Advantages inexpensive and good in large scale
studies
12
Scope of the Problem 1
subjective-state
subjective-trait
Stanford Sleepiness Scale

• Epworth SS
• Sleepiness problems

Weak or no relationships
Multiple Sleep Latency Test
objective-state
13
Scope of the Problem 2
Stanford Sleepiness Scale

Sleep-Disordered Breathing (SDB)
, -

• Epworth SS
• Sleepiness problems

Multiple Sleep Latency Test
, -
14
How are the sleepiness factors differentially
related ?
Research Questions
Is subjective (trait) sleepiness
multi-factorial ?

• Age
• Gender
• BMI
• Education
• Sleep debt
• MSLT- state-objective
• SSS - state-subjective

15
Methods
16

• Data Collection
• From a sleep survey (n2913)
• Self-reported sleepiness problems (7 items)
• Feeling excessively sleepy
• Not feeling rested
• Difficult getting up in the morning
• Need for stimulants to stay awake
• Epworth Sleepiness Scale (ESS) (8 items)
• Covariates
• From daytime study (n145)
• Stanford Sleepiness Scale (SSS)
• Multiple Sleep Latency Test (MSLT)

17
Multiple regression analysisAssess the
relationships between covariates, MSLT, SSS and
the sleepiness factors
Statistical Analysis

• Principal-axis Factor Analysis (FA)
• Identify dimensions or structure of the 15
  subjective-trait sleepiness items

• Age
• Gender
• BMI
• Education
• Sleep debt
• MSLT- state-objective
• SSS - state-subjective

18
Factor Analysis Decathlon example
100 Meters
SPEED
400 Meters
Long Jump
High Jump
STRENGTH
100m Hurdles
Discus
Shot Put
ENDURANCE
Javelin
Pole Vault
1500 Meters
19
Factor Analysis Overview
Original data matrix
Correlation matrix
Factor matrix
Variables
v1 v2 . . . v10
v1 v2 . . . v10
Obs
Correlation coefficients
20
Factor Analysis (FA) Methods Decision

• 1) Research objective?
• Confirmatory FA
• A priori testing
• Structural Equation Modeling
• Exploratory FA
• Identify factors among variables

21
Factor Analysis (FA) Methods Decision

• 2) Factoring method?
• Data summary
• Principal Component Analysis (PCA)
• Analyzes total (common unique error) variance
• Identify structure
• Principal Axis Factor Analysis (i.e, common
  factor analysis)
• Analyze common variance only

C1 b11 (X1) b12 (X2) . . . b1p (Xp)
22
Factor Analysis (FA) Methods Decision

• 3) Number of significant factors to retain?
• Scree plot
• Meaningful of the factors
• Eigenvalue1
• Variance explained

elbow
23
Factor Analysis (FA) Methods Decision

• 4) Rotation method for simple structure?
• Uncorrelated factors (Orthogonal rotation) Vs.
  Correlated Factors (Oblique rotation)

Unrotated Factor Loadings Variable
I II v1 0.50 0.80 v2
0.60 0.70 v3 0.90
-0.25 v4 0.80 -0.30 v5
0.60 -0.50
v1
v2
?
v4
v3
v5
24
Data Analysis

• Exploratory/Common Factor Analysis
• Scree test to determine number of factors
• Guided by meaningfulness of the items to factor
• Oblique rotation for simple structure

25
Characteristics of the study sample (n2913) Age
in years at sleep survey, mean (SD) 46.6
(7.9) Body Mass Index (kg/m2), mean (SD) 27.6
(6.1) Male Gender, n () 1351 (46.4) Highest
Educational Attainment, n () Some college
or less 1674 (57.5) College graduate or
higher 1239 (42.5) Sleep Debt, n () -3.0
to -0.25 144 (5.0) 0.00 837 (28.7)
0.25 to 0.75 230 (7.9) 1.00 836
(28.7) 1.25 to 1.75 274 (9.4)
2.00 388 (13.3) 2.25 to 2.75 76
(2.6) 3.00 to 6.00 128 (4.4)
Self-reported hours of non-workday night sleep
minus workday sleep
58.4
26

• Descriptive statistics of the observed sleepiness
  items from the sleep survey (n2913)
• response
• Item 0 1 2 3
  4
• 1. Not feeling rested during the day no matter
• how many hours of sleep you had 7
  35 36 16 6 (22)
• 2. Need for coffee or other stimulants to stay
• awake during the day 34 28
  21 11 6 (17)
• 3. Great difficulty getting up in the morning
  16 40 30 10 4
  (14)
• 4. Feeling excessive daytime sleepiness
  10 40 36 11
  3 (14)
• 5. Falling asleep or dozing momentarily
• at meetings, church, etc. 45 32
  17 5 1 (6)
• Note. 1 Item responses (frequency per month) 0
  never, 1 rarely (once), 2 sometimes (2-4), 3
  often (5-15), 4 almost always (16-30).

27
Descriptive statistics of the observed sleepiness
items from the sleep survey (n2913) continued .
. . response Epworth Sleepiness
Scale items 0 1 2
3 Lying down to rest in the
afternoon when circumstances permit
9 25 29
37 Watching TV 11
41 33 15 Sitting and
reading 18 44 26 12
As a passenger in a car for an hour
without a break 27
41 20 12
Sitting quietly after a lunch without alcohol
48 36 13 3 Sitting
inactive in a public place 45
40 12 3 In a car while
stopped for a few minutes in traffic 88
10 1 1 Sitting and talking
to someone 85 13
1 1 Epworth Sleepiness Scale
items (likely to doze off in specific
situations) 0 never, 1 slightly, 2
moderate, 3 high.
28
Factor analysis with oblique rotation of the
sleepiness variable (n2913)


Factor Factor grouping/item
I II III I. Dysfunctional
sleepiness 1. Not rested during the day no
matter how many hours of sleep you
had 0.74 -0.04 0.01 3. Feeling of excessive
daytime sleepiness 0.71 0.08 0.03 2. Great
difficulty getting up in the morning 0.62 -0.06 -
0.05 5. Need for coffee, or other stimulants
0.44 0.11 0.02 to stay awake during
the day II. Sleep propensity in active
situations 11. Likely to doze of while
sitting and talking to -0.05 0.68 -0.06
someone 13. Likely to doze off in a car while
stopped 0.01 0.63 -0.12 for a few minutes
in traffic 8. Likely to doze off while sitting
inactive -0.04 0.62 0.24 in a public place
(e.g., a theater or a meeting) 4. Fall asleep
or doze momentarily 0.03 0.61 0.13 -at
meetings, church, etc. 12. Likely to doze off
while sitting quietly 0.05 0.51 0.18
after lunch without alcohol 9. Likely to
doze off as a passenger in a 0.10 0.36
0.22 car for an hour without a break III.
Sleep propensity in passive situations 7.
Likely to doze off while watching TV
-0.05 -0.05 0.71 6. Likely to doze off
while sitting and reading -0.01
0.06 0.69 10. Likely to doze while lying down
to rest in the afternoon when circumstances
permit 0.15 0.14 0.26 Percent of
variance explained 8.7 30.3 3.0 Cronbach
s alpha 0.74 0.81 0.63
29
EFA Results-Factor I
Variance explained 8.7 Cronbachs alpha 0.74
30
Variance explained 30.3 Cronbachs alpha
0.81
31
Variance explained 3.0 Cronbachs alpha 0.63
32
Sleepiness Dimensions
trait-subjective
SLEEPINESS
state-objective
state-subjective
33
How are the sleepiness factors differentially
related ?

• Sleep debt
• Age
• Gender
• BMI
• Education
• MSLT- state-objective
• SSS - state-subjective

34
(No Transcript)
35
Models for subjective sleepiness factors fitted
by gender, age, body mass index,
and educational status (n2913)
Age 5 years change in age. BMI 2 units change
in body mass index. All relationships
statistically significantly, p-value except for educational status-sleepy
behavior-active sleepiness factor relationship.
36
Gender and Daytime Somnolence
Baldwin et al., Sleep 2003 (Multicenter Sleep
Heart Health Study) n6440
respiratory disturbance index, cardiac
respiratory diseases, frequent awakenings,
sedative alcohol use, educational level, race
p 37
Subjective sleepiness dimensions SSS
Multiple regression coefficient adjusting for
gender, age, education status, body mass index,
and sleep debt. p-value indicate 95 confidence interval.
38
Sleepiness Dimensions
trait-subjective
state-objective
39
Subjective sleepiness dimension MSLT (n145)
Multiple regression coefficient adjusting for
gender, age, education status, body mass
index, and sleep debt. p-value Vertical bars indicate 95 confidence interval.
40
Sleepiness Dimensions
trait-subjective
41
ESS and MSLT
Punjabi, et al (Sleep 2002) (Wisconsin Sleep
Cohort Study) n632

• Survival analysis for censored MSLT data
• Proportional hazard model

30 and 69 increase in sleep onset, at any
given point in time
42
Conclusions

• Subjective sleepiness is not a unitary phenomenon
  as supported by . . .
• Finding of 3 factors of SUBJECTIVE, trait
  sleepiness (Factor analysis)
• Differential effects of sleep debt, age, and
  gender on the sleepiness dimensions
• Convergent and divergent validity of the
  SUBJECTIVE, trait sleepiness factors with the
  MSLT and SSS.

43
Conclusions

• Accurate relationships between causes and effects
  of subjective trait sleepiness maybe improved
  with the use of the proper aspect(s) of
  subjective sleepiness.
• ESS, a widely used measure of subjective
  sleepiness, may miss sleepiness in females.
• Differential Item Functioning (DIF) analysis is
  needed to investigate true effect vs. item
  response (gender age).