The Science and Practice of Antifogmatics Gregory Belenky, M.D. - - PowerPoint PPT Presentation

▶
the science and practice of
SMART_READER_LITE
LIVE PREVIEW

The Science and Practice of Antifogmatics Gregory Belenky, M.D. - - PowerPoint PPT Presentation

Apr 06, 2023 8 likes •1.14k views

Lifting the Fog of Fatigue: The Science and Practice of Antifogmatics Gregory Belenky, M.D. Research Professor and Director Sleep and Performance Research Center Spokane, WA The Operational Environment Defined Operational

slide-1
SLIDE 1

Lifting the Fog of Fatigue: The Science and Practice of “Antifogmatics”

Gregory Belenky, M.D. Research Professor and Director Sleep and Performance Research Center Spokane, WA

slide-2
SLIDE 2

Washington State University

The Operational Environment Defined

  • Operational Environment
  • Human performance critical to correct outcome of the system – the outcome

itself is critical

  • There a temporal envelope within which the correct decision must be made or

the system fails

  • John Boyd and the Observe, Orient, Decide, Act (OODA) Loop
  • Most operational settings are complex and tightly-coupled
  • Many operational settings involve 24x7 operations, extended work hours

and shift work

  • High reliability organizations maintain
  • Mindfulness in day-to-day operations
  • Presence of mind in emergencies

Coram – John Boyd: The Fighter Pilot Who Revolutionized War

slide-3
SLIDE 3

The Earth at Night: The Problem of 24/7 Operations

slide-4
SLIDE 4

Washington State University

Sleep: A Fundamental Mystery in Neurobiology

  • Sleep is found humans, mammals, birds, reptiles, fish, insects, and

(perhaps) jellyfish – in any animal with one or more assemblies of nerve cells (neuronal assemblies)

  • After over 100 years of experimental work, we know:
  • Adequate sleep sustains performance
  • Inadequate sleep degrades performance
  • We do not know:
  • Why extended waking degrades performance?
  • How sleep restores performance?
slide-5
SLIDE 5

Washington State University

Why Sleep?

  • Productivity
  • Personal
  • Corporate
  • Safety
  • Personal
  • Corporate
  • Public
  • Health
  • Well-being

“I do not care how much they sleep; I want to know how well they perform.” General Max Thurman Disasters spring from “not a major blunder, but reasoned calculations that slip just a little.” Brigadier General S.L.A. Marshall

slide-6
SLIDE 6

Washington State University

Risks of Sleep Restriction and Sleep Deprivation

  • Short term (operational environment)
  • Minutes, hours
  • Error, accident, catastrophe
  • Mid-term
  • Weeks, months, years
  • Bad planning, inadequate strategizing, poor life decisions
  • Long-term
  • Years
  • Overweight/obesity, Type II Diabetes, Metabolic Syndrome, Cardiovascular

Disease, etc.

  • Triad of factors supporting health, productivity, and well-being
  • Diet
  • Exercise
  • Sleep
slide-7
SLIDE 7

Fatigue: Operationally Defined

  • Fatigue is subjectively defined by

– Verbal self-report – “I am tired.” – Endorsing the high end of a fatigue scale, e.g., Samn-Perelli – Observation by co-workers of fatigue behaviors or degraded performance

  • Fatigue is objectively defined by degraded performance

– Added metrics – Psychomotor vigilance task (PVT) – Embedded metrics

  • Lane deviation – driving
  • FOQA – flying
  • etc….
slide-8
SLIDE 8

Field Measurement of Sleep and Performance

  • Objective Measurement
  • f Sleep

– Electrophysiological recording (PSG) – Activity Monitoring (Actigraph)

  • Objective Measurement
  • f Performance (PVT)
slide-9
SLIDE 9

Actigraph Data

  • Blue boxes indicate

sleep intervals

  • Teal boxes indicate

periods of rest

  • Purple black out

indicates excluded data due to off-wrist detection

slide-10
SLIDE 10

Actigraph Data

  • Blue boxes indicate

sleep intervals

  • Teal boxes indicate

periods of rest

  • Purple black out

indicates excluded data due to off-wrist detection

slide-11
SLIDE 11

The Psychomotor Vigilance Task (PVT): A Sensitive Metric of Vigilance

  • A reaction time test
  • Administered by PC or Smartphone
  • ~10 stimuli present/min for 10 min

– Sensitive to sleep deprivation and sleep restriction – Sensitive to circadian periodicity – Sensitive to time on task (workload)

  • Good psychometric properties

– IQ independent – Virtually no learning involved – Unforgiving of attentional lapses

slide-12
SLIDE 12

Response by Response: Attentional Lapses during Sleep Deprivation

Washington State University

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70

200 0 400 0 600 0 800 0

RESPONSE NUMBER 60 Hours Awake 36 Hours Awake 12 Hours Awake 84 Hours Awake

200 0 400 0 600 0 800 0 200 0 400 0 600 0 800 0 200 0 400 0 600 0 800 0

12 Hours Awake 36 Hours Awake 60 Hours Awake 84 Hours Awake

slide-13
SLIDE 13

From Wake State Instability to Accidents: Performance Lapses Predict Risk (Not Errors)

Washington State University

0 Time on task (min) 10

Probe of brain impairment (PVT RT in ms) Demands

  • f task and

environment Impact

  • f failure

Swiss Cheese Model of Accident Causation (Reason, 2001)

From Van Dongen and Hursh (2010) PPSM 5e

slide-14
SLIDE 14

Washington State University

Fatigue and its Components

  • Fatigue operationally defined
  • Subjectively by self-report
  • Objectively by degraded performance
  • Fatigue is the final common pathway integrating
  • Sleep/wake history (time awake and sleep loss)
  • Circadian rhythm (time of day)
  • Workload (time on task, task intensity, and task

complexity)

  • Individual differences in response to time awake, time
  • f day, and time on task
slide-15
SLIDE 15

Washington State University

Sleep, Fatigue, and Predicting Performance and Fatigue Risk

  • Epidemiological studies associate sleep/wake history,

circadian phase, and workload with accident risk

  • Mathematical models have are being developed to

integrate sleep/wake history, circadian rhythm, and workload to predict individual performance (fatigue- risk) in real-time

slide-16
SLIDE 16

Fatigue as the Integration of Sleep Loss, Circadian Rhythm, and Workload

slide-17
SLIDE 17

Linear Decline with Extended Waking Modulated by Circadian Phase Amplifies Time on Task

slide-18
SLIDE 18

Interaction of Time Awake, Time of Day, and Time on Task

slide-19
SLIDE 19

The Science of Sleep and Circadian Rhythms

slide-20
SLIDE 20

Washington State University

The Sleep/Wake Cycle

0000 1200 0600 1800 NREM REM Waking

slide-21
SLIDE 21

Physiology of Slow Wave and REM Sleep

Slow Wave Sleep REM Sleep Sleep Cycle Kryger, Roth and Dement, Principles and Practice of Sleep Medicine, 2005

slide-22
SLIDE 22

Washington State University

The Internal Body Clock (Circadian Rhythm)

From Kryger, Roth and Dement, 2005

slide-23
SLIDE 23

The Circadian Rhythm in Sleep and Core Body Temperature

  • The clock has an intrinsic periodicity of just
  • ver 24 hours
  • Residing in the suprachiasmatic nucleus
  • Other clocks
  • All cells express the “clock” genes and have an

intrinsic rhythmicity

  • The circadian clock is entrained

(synchronized) to the light dark cycle by light exposure

  • The retina contains specialized (non-visual)

receptors that are sensitive to blue light (blue sky detectors)

  • The circadian rhythm is expressed in a variety
  • f ways
  • Core body temperature
  • Dim light melatonin onset
  • Hormonal rhythms
  • A rhythm in performance
  • A rhythm in sleep propensity

Washington State University Mistlberger and Rusak (2005) In Kryger, Roth, and Dement, Principles and Practice of Sleep Medicine

slide-24
SLIDE 24

The Circadian Rhythm Consolidates Sleep

Edgar, Dement, & Fuller, 1993

slide-25
SLIDE 25

Washington State University

85 Hours of Total Sleep Deprivation: Effect on Performance

Adapted from Thomas, et al., 2000

slide-26
SLIDE 26

Washington State University

Sleep Deprivation and Alcohol Intoxication

Dawson & Reid, 1997

slide-27
SLIDE 27

Washington State University

Consequences of Sleep Restriction and Sleep Deprivation

  • Short term
  • Minutes, hours
  • Error, accident, catastrophe
  • Mid-term
  • Weeks, months, years
  • Bad planning, inadequate strategizing, poor life decisions
  • Long-term
  • Years
  • Overweight/obesity, Type II Diabetes, Sleep Disorder Breathing, Metabolic

Syndrome, etc.

  • Triad of factors supporting health, productivity, and well-being
  • Diet
  • Exercise
  • Sleep
slide-28
SLIDE 28

Washington State University

Total Sleep Deprivation Imaging Studies

Throughput (Percent of Baseline)

120 100 80 60 40 20

Sleep Deprivation (Hours) 0 24 48 72 86

  • Mean Performance (N=17)
  • Cubic Spline
  • Linear Regression

PET Scans

slide-29
SLIDE 29

Washington State University

Brain Metabolism at 24, 48, & 72 Hours

  • f Sleep Deprivation

+ 32 mm AC-PC 24 h SD 48 h SD 72 h SD + 8 mm AC-PC

Z

1.65 2.33 2.58 3.08 > 4.16 N = 17 Thomas, et al., J. Sleep Res. 2000

slide-30
SLIDE 30

Washington State University

Brain Metabolism during Slow Wave and REM Sleep

Frontal areas are deactivated during Slow Wave Sleep; decline in flow of ~30% Frontal areas remain deactivated during REM; increase in flow to waking levels or above except in prefrontal cortex Frontal areas are re- activated only after awakening

Braun et al., Brain, 1997

slide-31
SLIDE 31

Sleep Restriction and Performance

Washington State University

slide-32
SLIDE 32

Washington State University

Effects of Sleep Restriction in Performance: A Sleep Dose/Response Study

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

8 hrs in bed 3, 5, 7, 9 hrs in bed Adaptation Phase Experimental Phase Recovery Phase 4 5 6 7 8 9 10 11 12 13 14 15

Release from study

8 hrs in bed 1 2 3

slide-33
SLIDE 33

Washington State University

Volunteers in the Laboratory

  • Sleep measured with

polysomnography (electrodes, wires, recorders)

  • Performance measured

with computer-based tests.

slide-34
SLIDE 34

Washington State University

1 3 5 7 9 T2 B E1 E2 E3 E4 E5 E6 E7 R1 R2

Day Amount of Sleep (Hrs)

9 HR 7 HR 5 HR 3 HR

Mean Sleep, Baseline, Experimental Days, & Recovery

Mean Sleep Experimental Days 9 hr group – 7.9 hrs 7 hr group – 6.3 hrs 5 hr group – 4.7 hrs 3 hr group – 2.9 hrs

slide-35
SLIDE 35

Washington State University

Psychomotor Vigilance Task

Belenky et al., 2003

slide-36
SLIDE 36

Washington State University

Driving Simulator

slide-37
SLIDE 37

Washington State University

Driving Simulator – Lane Deviation

0.5 1 1.5 2 T1 T2 B E1 E2 E3 E4 E5 E6 E7 R1 R2 R3

3 Hr 5 Hr 7 Hr 9 Hr

Deviation of Lane Position Day

slide-38
SLIDE 38

Washington State University

Individual Variability in Resistance to Sleep Restriction

0.001 0.002 0.003 0.004 0.005 1 2 3 4 5 6 7 8 9 10 11 12 13

Days Speed on PVT

Mean +/- SEM (n = 18) Resistant Subject Sensitive Subjet #1 Sensitive Subject #2

Baseline Recovery 3 Hours Sleep/Night X 7 Days

slide-39
SLIDE 39

Chronic Sleep Loss: Objective and Subjective Effects

Adapted from Van Dongen et al (2003)

slide-40
SLIDE 40

Time on Task and Time Awake Effects on Performance during Sleep Deprivation and Sleep Restriction

  • The Psychomotor Vigilance Task (PVT) is a

sensitive metric of human performance

– A reaction time test – Administered by PC, Palm OS PDA, or Window Pocket PC PDA – ~10 stimuli present/min for 10 min – Sensitive to sleep deprivation and sleep restriction – Sensitive to circadian periodicity – Sensitive to time on task

  • Time on task effects develop over minutes and

are reversed by simple rest (time off task)

  • Time awake effects develop over hours and days

and require sleep to reverse

  • Time on task interacts with time awake
  • The effects of time awake on performance may

be mediated through increasing sensitivity to time on task

  • The PVT may be an excellent task to probe using

new techniques of brain imaging the “use- dependency” of the effects of time awake on performance

1.80 2.20 2.60 3.00 3.40 0800 1200 1600 2000 0000 0400 0800 1200 1600 2000

Time (Hours) Psychomotor Vigilance Task (PVT) Performance Time on Task Effects during 38 Hours

  • f Total Sleep Deprivation

1.5 2 2.5 3 3.5 4 4.5 Baseline E1 E2 E3 E4 E5 E6 E7 R1 R2 R3

3-Hr 5-Hr 7-Hr 9-Hr

Psychomotor Vigilance Task (PVT) Performance Time (Days) Time on Task Effects during 7 Days of Sleep Restriction and Subsequent Recovery N= 16 -18/group N = 49

slide-41
SLIDE 41

Split Sleep and Napping and Sleep

slide-42
SLIDE 42

Newark to Hong Kong – Over the North Pole

Washington State University

slide-43
SLIDE 43

Home, Layover, and In-Flight Sleep in a Boeing 777 Pilot

Belenky, et al., in preparation

slide-44
SLIDE 44

Transmeridian Travel: Actigraph Record

  • Overseas Travel
  • 2/11: Leave Eastern US
  • 2/12 - 2/16: SWA
  • 2/17: Germany
  • 2/18 - 2/19: Hawaii
  • 2/20: Arrive Eastern US
  • Sleep in afternoon (EST) (1)

and some divided sleep (2) during time in SWA and Germany

  • Sleep in mid-morning hours

(EST) (3) during time in Hawaii

  • Sleep during normal sleeping

hours (EST) (4) on return to Eastern US

Eastern Standard Time (EST) Days (1) (1) (1) (1) (1) (1) (2) (3) (4) (4) (4) (4) (4) (4) (2)

slide-45
SLIDE 45

Night Float vs. Day Shift in Physicians in Training

Washington State University

slide-46
SLIDE 46

Physician on Day Shift and Night Float Sequence

Washington State University

slide-47
SLIDE 47

Sleep Off Shift & On Shift / Day Shift

  • vs. Night Float

Washington State University

Day Shift Night Float

slide-48
SLIDE 48

Response Surface Mapping of PVT Lapses in Split Restricted Sleep

slide-49
SLIDE 49

Consolidated vs. Split vs. Fragmented Sleep

  • Recuperative value of sleep depends
  • n total sleep time over 24 hours
  • Consolidated sleep

– Nocturnal (night) – typically 7-8 hours; facilitated by circadian rhythm – Diurnal (day) – typically ~ 5 hours; truncated by circadian rhythm

  • Split sleep

– 5 hours nocturnal / 2-3 hours diurnal

  • Fragmented sleep

– Awakening every 2-3 minutes – Fragmentation to this degree abolishes recuperative value of sleep

  • Sleep interrupted every 20+ minutes

as recuperative as uninterrupted sleep

Washington State University

Bonnet M & Arand D (2003) Clinical effects of sleep fragmentation vs. sleep deprivation. Sleep Medicine Reviews, 7(4) 297-310

slide-50
SLIDE 50

Bed – Flat Sleeperette – 49.5 degrees to the vertical Reclining Seat - 37 degrees to the vertical Armchair - 17.5 degrees to the vertical

slide-51
SLIDE 51

Bed – Flat Sleeperette – 49.5 degrees to the vertical Reclining Seat - 37 degrees to the vertical Armchair - 17.5 degrees to the vertical

slide-52
SLIDE 52

Cockpit Napping

slide-53
SLIDE 53

Washington State University

Other Countermeasures

  • Stimulants on shift
  • Caffeine
  • Other stimulant drugs, e.g., modafinil
  • Stimulants (caffeine, d-amphetamine, modafinil) appear

equivalent for first few hours in clinically acceptable doses

  • Sleep-inducing drugs when sleeping off shift
  • BZD receptor agonists
  • Melatonin and melatonin analogues
  • Naps on shift
  • Bright (blue) light on shift
  • Strict environmental control when sleeping off shift
  • Light and noise while sleeping
  • Commute times to and from work

Wesensten et al., 2005

slide-54
SLIDE 54

Washington State University

50 60 70 80 90 100 110

0800 1600 0000 0800 1600 0000 0800 1600 0000 0800 1600 0000 0800 1600

Time of Day

Mean Relative Speed .

Placebo d-Amphetamine 20 mg Caffeine 600 mg Modafinil 400 mg Drug @ 65 hrs sleep loss

RECOVERY SLEEP

Adapted from Wesensten et al., 2005

Amphetamine vs. Modafinil vs. Caffeine

slide-55
SLIDE 55

Modafinil vs. Caffeine

1.0 1.5 2.0 2.5 3.0 3.5

0800 1200 1600 2000 0000 0400 0800 1200 1600 2000 0000 0400 0800 1200

Time of Day Mean Speed (1/RT * 1000)

Placebo Modafinil 100 mg Modafinil 200 mg Modafinil 400 mg Caffeine 600 mg

Drug or Placebo @ 2355

DAY 2 DAY 3 DAY 4

slide-56
SLIDE 56

Sleep and Performance in Operations

slide-57
SLIDE 57

Washington State University

Acute Total Sleep Deprivation in a Air Cargo Flight Accident: American International Flight 808 18 August 1993

slide-58
SLIDE 58

Washington State University

Guantanamo Bay, Cuba

slide-59
SLIDE 59

Washington State University

All 3 crew members were rescued from the cockpit and survived

Crash Site

slide-60
SLIDE 60

Washington State University

The Approach to Guantanamo

Approach to Guantanamo requires a sharp right bank to avoid Cuban air space

slide-61
SLIDE 61

Crash of American International Flight 808: Sleep Amounts Prior to Crash Landing

0000 0800 1600 0000 0800 1600 0000 0800 1600

|_____|_____|_____|_____|_____|_____|_____|_____|__

0000 0800 1600 0000 0800 1600 0000 0800 1600

|_____|_____|_____|_____|_____|_____|_____|_____|__

0000 0800 1600 0000 0800 1600 0000 0800 1600

|_____|_____|_____|_____|_____|_____|_____|_____|__

Captain Co-PILOT Engineer

16 August 17 August 18 August

= reported sleep time

slide-62
SLIDE 62

Accident Investigation – American International Flight 808 (1993)

Captain: 71% Co-Pilot: 70% Engineer: 77%

slide-63
SLIDE 63

Cockpit Voice Recorder just Prior to Crash

Engineer: Slow, Airspeed Co-Pilot: Check the turn. Captain: Where’s the strobe? Co-Pilot: Right over here. Captain: Where? Co-Pilot: Right inside there, right inside there. Engineer: You know, we’re not gettin’ our airspeed back there. Captain: Where is the strobe? Co-Pilot: Right down there. Captain: I still don’t see it. Engineer: #, we’re never goin’ to make this. Captain: Where do you see a strobe light? Co-Pilot: Right over here. Captain: Gear, gear down, spoilers armed. Engineer: Gear down, three green spoilers, flaps, checklist ???: There you go, right there, lookin’ good. Captain: Where’s the strobe? Co-Pilot: Do you think you’re gonna make this? Captain: Yeah… if I can catch the strobe light. Co-Pilot: 500, you’re in good shape. Engineer: Watch the, keep your airspeed up. Co-Pilot:

  • 140. [sound of stall warning]

???: Don’t – stall warning. Captain: I got it. Co-Pilot: Stall warning. Engineer: Stall Warning Captain: I got it, back off. ???: Max power! ???: There it goes, there it goes! ???: Oh no!

slide-64
SLIDE 64

"The impaired judgment, decision-making, and flying abilities of the captain and flight crew due to the effects

  • f fatigue [sleep deprivation]; the captain's failure to

properly assess the conditions for landing and maintaining vigilant situational awareness of the airplane while maneuvering onto final approach; his failure to prevent the loss of airspeed and avoid a stall while in the steep bank turn; and his failure to execute immediate action to recover from a stall.” _____________________________ From NTSB Report Crash of American International Flight 808: Probable Causes

slide-65
SLIDE 65

Washington State University

The Harvard Intervention Studies: A Simple Case of Fatigue Risk Management

slide-66
SLIDE 66

Washington State University

Traditional vs. Intervention Schedule

Landrigan, et al. (2004) NEJM 351: 18, 1838-1848

slide-67
SLIDE 67

Duration of Work Week and Effect on Sleep

  • Duration of work week decreased from 85

hours to 65 hours

  • Total sleep time/24 hours increased from 6.6

to 7.4 hours

Washington State University

slide-68
SLIDE 68

Washington State University

Limiting Work Hours: Effect on Serious Medical Errors

Landrigan, et al. (2004) NEJM 351: 18, 1838-1848

slide-69
SLIDE 69

Medical Errors, Adverse Events, & Car Crashes

  • Survey of 2737 residents (PGY 1s)

– Extended (≥ 24 hours) shifts vs. normal day shifts

  • Barger et al., 2005

– More crashes, near misses, and fall asleep while driving with extended work hours

  • Barger et al., 2006

– More significant medical errors, attentional failures, and fatigue-related preventable adverse events resulting in a fatality

  • Ayas et al., 2008

– Increased percutaneous injuries

Washington State University

Barger et al., NEJM 2005 Barger et al., NEJM 2006 Ayas et al., PLoS 2008

slide-70
SLIDE 70

Washington State University

Acute Partial Sleep Deprivation in an Air Traffic Control and Pilot Error Accident

slide-71
SLIDE 71

Comair Flight 5191

  • Lexington, KY to Atlanta, GA
  • Take off ~ 0630 hrs
  • Assigned the Runway 22
  • Used Runway 26
  • Pilot took wrong turn onto unlit

Runway 26

  • Neither pilots nor air traffic

controller noticed error

  • Turned aircraft over to First

Officer for take off

  • Crashed just past the end of the

runway

  • Killed all 47 passengers and two
  • f the three crew
  • Similar error in 1993
  • Caught prior to take-off roll
  • By both pilots and air traffic

controller

Runway 22 Runway 26

slide-72
SLIDE 72

Washington State University

Sleep in Air Traffic Controller and Pilots

  • Air traffic controller (a 17-year veteran) working

alone at an airport in Kentucky

  • Worked early day shift from 0630-1430 hours (6:30 AM –

2:30 PM)

  • Had the mandatory by FAA rules 8 hours off
  • Slept ~ 2 hours in the late afternoon
  • Went back to work at 2330 (11:30 PM)
  • Worked through the night until the accident at ~0600 hrs
  • Pilots and co-pilot scheduled for take-off at 0600 hrs
  • Likely in bed no earlier than 2200 hrs (10:00 PM)
  • Awake at 0400 hrs.
  • Both air traffic controller and pilots were sleep

restricted and at low point in circadian rhythm

slide-73
SLIDE 73

Washington State University

Model-Based Accident Reconstruction in a Court-martial on a Charge of Negligent Homicide

slide-74
SLIDE 74

Washington State University

Attend Funeral Leave FTX Return to FTX Accident All-night Duty (no sleep)

|____|____|____|____|____|____|

0000 0000 0000 0000 0000 0000 0000 ~ 6.5 hrs sleep per night

Accident Reconstruction: Timeline

Day 1 Day 2 Day 3

slide-75
SLIDE 75

Washington State University

~ 1 mile Obstacle 1 Obstacle 2

Side Path

Accident Reconstruction: Setting

slide-76
SLIDE 76

Washington State University

  • 24%
  • 10%

Day 1 Day 2 Day 3

Performance Prediction: 8 Hrs Sleep/Night Performance Prediction: NCO’s Sleep/Wake History

The Accident 80 70 60

  • 14%

Accident Reconstruction: Model Predictions

slide-77
SLIDE 77

Individual Differences in Sensitivity to Sleep Loss, Circadian Phase, and Workload

slide-78
SLIDE 78

4 8 12 16 20 24 28 32 36 40

hours awake 5 10 15 20 25 PVT lapses

08 12 16 20 00 04 08 12 16 20 00

time of day

n=8 n=7

Trait Individual Differences in Vulnerability to Performance Impairment from Sleep Loss

cognitive impairment →

Adapted from Van Dongen et al (2004)

slide-79
SLIDE 79

4 8 12 16 20 24 28 32 36 40

hours awake 1 2 3 4 5 S t a n f

  • r

d S l e e p i n e s s S c a l e

08 12 16 20 00 04 08 12 16 20 00

time of day

4 8 12 16 20 24 28 32 36 40

hours awake 5 10 15 20 25 PVT lapses

08 12 16 20 00 04 08 12 16 20 00

time of day

Mismatch between Subjective Sleepiness and Objective Performance Deficits

Adapted from Van Dongen et al (2004)

slide-80
SLIDE 80

Empirical Best Linear Unbiased Predictors

  • 0.8
  • 0.6
  • 0.4
  • 0.2

0.0 0.2 0.4 0.6 0.8 1 2 3 4 5 6 7 8 9 10 11

subjects relative performance

Individual Differences in Active-Duty Air Force Pilots during Simulated F-117 Extended Night Flights

left 720 degrees turn roll performance

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

  • 6
  • 1

4 9 14 19 24

time of day

N=10 flight path deviation performance distribution left 720 degrees turn roll angle performance over time

A B C D E F G H I J

subjects relative performance

(Self-)selection mechanisms do not eliminate individual differences in vulnerability to sleep loss—even in highly specialized professions

From Van Dongen et al (2006)

slide-81
SLIDE 81

Individualized Performance Modeling

Subject A

24 h ahead performance prediction snapshots at 44h awake (time 03:30)

Subject B

Group-Average Model Individualized Model

past performance future performance performance prediction │ 95% confidence interval

impairment → impairment →

Adapted from Van Dongen et al (2007)

slide-82
SLIDE 82

Washington State University Spokane

Predicting Performance from Actigraphically- Derived Sleep Wake History

slide-83
SLIDE 83

1 1 1 1 1 1 1 6 12 18 24 30 36 42 48 54 60

Time awake Process S

  • 3
  • 2
  • 1

1 2 3

Process C

  • 4

4 8 12 16 20 24 28 32 6 12 18 24 30 36 42 48 54 60

Time awake PVT lapses Model fit

Homeostatic Process Circadian Process PVT Lapses Model Fit

N=11

The SPRC Two-Process Model: Components and Fit to Data

measured

  • r predicted

sleep measured or predicted light exposure planned work/rest schedule measured

  • r predicted

sleep measured or predicted light exposure planned work/rest schedule predicted fatigue predicted fatigue sleep inertia sleep inertia homeostatic process circadian process homeostatic process circadian process effect of chronic sleep loss effect of chronic sleep loss

83 Washington State University

slide-84
SLIDE 84

Washington State University

The Internal Body Clock (Circadian Rhythm)

Kryger, Roth and Dement, Principles and Practice of Sleep Medicine, 2005

slide-85
SLIDE 85

Washington State University

Normal vs. Night Shift-Work Sleep

  • Graphs matched on time scale
  • Note naps during work shift and

in late afternoon

  • Note truncated main daytime

sleep

Normal Sleep Shift-Worker Sleep Akerstedt, Occupational Medicine, 2003

slide-86
SLIDE 86

Washington State University

The New Science and Art of Fatigue Risk Management

slide-87
SLIDE 87

Humans and Machines: The Person in the Loop

  • Alternative futures (as

envisioned ~30 years ago):

– Man without computer – Computer without man – Man against computer – Man with computer against man with computer

  • Current state:

– Persons embedded in robotic systems – Monitored, assisted, sustained

… all watched over by machines of loving grace.”

  • Richard Brautigan (1963)
slide-88
SLIDE 88

Integration of Fatigue Risk Management into Rostering and Scheduling Software

  • Personal biomedical status monitoring
  • Sleep/wake history (by sleep watch/actigraph)
  • Circadian rhythm phase (by technology TBD)
  • Predict performance in real time person by person (by

biomathematical performance prediction model)

  • Validate with embedded performance metrics
  • Lane deviation (trucking)
  • Flight performance (commercial aviation)
  • Integrate performance prediction into rostering and

scheduling software

  • Integrate into objective function
  • Optimize along with other constraints
slide-89
SLIDE 89

Example of Actigraph Record

  • An example of an

actigraph record recorded over 6 days.

  • This person slept

from ~ 22:00 to 08:00. Sleeping Waking

slide-90
SLIDE 90

Effect of Sleep Loss on Performance on the Psychomotor Vigilance Test (PVT)

Washington State University

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70

200 0 400 0 600 0 800 0

RESPONSE NUMBER 60 Hours Awake 36 Hours Awake 12 Hours Awake 84 Hours Awake

200 0 400 0 600 0 800 0 200 0 400 0 600 0 800 0 200 0 400 0 600 0 800 0

12 Hours Awake 36 Hours Awake 60 Hours Awake 84 Hours Awake

From Doran et al. (2001) Arch Ital Biol.

slide-91
SLIDE 91

At Home and In-Flight Sleep in a Boeing 777 Pilot

:

Belenky, et al., in preparation

slide-92
SLIDE 92

Flights to Staff Pilots Available

Crew Scheduling in Commercial Aviation is Controlled by Rules & Objectives

Other Rules & Objectives Labor Agreements Flight Time Limits

Rules & Objectives Must be Satisfied

Alertness Model Scheduling Software Optimizes Assignments

From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

slide-93
SLIDE 93

Flight Time Limits Alone Do Not Protect Alertness

From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

slide-94
SLIDE 94

Labor Agreements Add Extra Protection – At A Cost

From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

slide-95
SLIDE 95

A Model Within Existing Constraints Improves Alertness

From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

slide-96
SLIDE 96

Modeling Alone Improves Alertness & Productivity

From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

slide-97
SLIDE 97

Washington State University

Gregory Belenky, MD Research Professor and Director Sleep and Performance Research Center Washington State University P.O. Box 1495 Spokane, WA 99210-1495 Phone: (509) 358-7738 FAX: (509) 358-7810 Email: belenky@wsu.edu

Point of Contact

slide-98
SLIDE 98

Washington State University

The New Science and Art of Fatigue Risk Management

slide-99
SLIDE 99

Humans and Machines: The Person in the Loop

  • Alternative futures (as

envisioned ~30 years ago):

  • Man without computer
  • Computer without man
  • Man against computer
  • Man with computer

against man with computer

  • Current state:
  • Persons embedded in

robotic systems

  • Monitored, assisted,

sustained

Washington State University

… all watched over by machines of loving grace.”

  • Richard Brautigan (1963)
slide-100
SLIDE 100

11/6/2010

Apparent Change in Performance Actual Change in Performance

SAFE UNSAFE SAFE UNSAFE

Sudden vs. Graceful Degradation

  • Sleep deprivation-induced orderly decreases in

performance and productivity precede accidents and catastrophic failures

slide-101
SLIDE 101

The Swiss Cheese Model of Accident Causation

Washington State University

Adapted from Reason, 2000

slide-102
SLIDE 102

Transmeridian Travel: Actigraph Record

  • Overseas Travel
  • 2/11: Leave Eastern US
  • 2/12 - 2/16: SWA
  • 2/17: Germany
  • 2/18 - 2/19: Hawaii
  • 2/20: Arrive Eastern US
  • Sleep in afternoon (EST) (1)

and some divided sleep (2) during time in SWA and Germany

  • Sleep in mid-morning hours

(EST) (3) during time in Hawaii

  • Sleep during normal sleeping

hours (EST) (4) on return to Eastern US

Eastern Standard Time (EST) Days (1) (1) (1) (1) (1) (1) (2) (3) (4) (4) (4) (4) (4) (4) (2)

slide-103
SLIDE 103

Sleep as an Item of Logistic Resupply

  • A Commander manages fuel by knowing:
  • How much the unit has on hand
  • Rates of utilization
  • Anticipated operations
  • With these quantities, the Commander can plan

for timely resupply to sustain operational performance

  • To manage sleep the Commander must know:
  • How much sleep his unit has been getting
  • How long this will sustain acceptable performance
  • Ensure that adequate opportunity for sleep exists

to sustain operational performance

slide-104
SLIDE 104

Integration of Fatigue Risk Management into Rostering and Scheduling Software

  • Personal biomedical status monitoring
  • Sleep/wake history (by sleep watch)
  • Circadian rhythm phase (by technology TBD)
  • Predict performance in real time person by person (by

biomathematical performance prediction model)

  • Validate with embedded performance metrics
  • Lane deviation (trucking)
  • Flight performance (commercial aviation)
  • Integrate performance prediction into rostering and

scheduling software

  • Integrate into objective function
  • Optimize along with other constraints
slide-105
SLIDE 105

Fatigue Risk Management & Safety Management

  • Embed within corporate safety management system (SMS)
  • Move fatigue issues from labor/management to safety
  • Safety enhances productivity (and the reverse)
  • SMS has built-in structure, yields economies of scale
  • Fatigue risk management systems (FRMS)
  • Multi-layered defense against fatigue-related error, incident, and

accident

  • Each layer “sloppy” but in the Swiss cheese model highly efficient at

preventing fatigue-related errors

  • Current examples are Union Pacific Railroad and easyJet Airlines
slide-106
SLIDE 106

At Home and In-Flight Sleep in a Boeing 777 Pilot

:

slide-107
SLIDE 107

Flights to Staff Pilots Available

Crew Scheduling in Commercial Aviation is Controlled by Rules & Objectives

Other Rules & Objectives Labor Agreements Flight Time Limits

Rules & Objectives Must be Satisfied

Alertness Model Scheduling Software Optimizes Assignments

Adapted from Romig & Klemets (2010)

slide-108
SLIDE 108

Flight Time Limits Alone Do Not Protect Alertness

Adapted from Romig & Klemets (2010)

slide-109
SLIDE 109

Labor Agreements Add Extra Protection – At A Cost

Adapted from Romig & Klemets (2010)

slide-110
SLIDE 110

A Model Within Existing Constraints Improves Alertness

Adapted from Romig & Klemets (2010)

slide-111
SLIDE 111

Modeling Improves Alertness & Productivity

Adapted from Romig & Klemets (2010)

slide-112
SLIDE 112

Washington State University

Summary: Shift Work and Sleep

  • Working night and early morning shifts is

associated with disrupted and truncated sleep

  • In some individuals this leads to insomnia during

available sleep opportunity and excessive sleepiness while awake

  • A function of circadian influences on sleep

propensity limiting sleep even when there is adequate opportunity for sleep

  • A variety of behavioral and pharmacological

measures may ameliorate these effects

slide-113
SLIDE 113

Washington State University

Fatigue Risk Management System (FRMS)

  • Five-tiered defense-in-depth to prevent fatigue related

errors, incidents, and accidents

  • Tier 1 – Does system of shift timing and duration allow for

adequate opportunity for sleep?

  • Computer-based rostering
  • Predictive Modeling
  • Tier 2 – Do employees take advantage of the sleep
  • pportunity?
  • Self-report
  • Wrist-worn actigraph (sleep watch)
  • Tier 3 – In the workplace, do they maintain adequate

alertness and performance?

  • Self-report & co-worker report
  • Palm Pilot Psychomotor Vigilance Task (PVT)
  • Embedded performance metrics
  • Tier 4 – Are there errors, near-misses?
  • Tier 5 – Are there incidents and accidents?

Dawson & McCulloch 2005