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

Mean Sleep, Baseline, Experimental Days, & Recovery Mean Sleep Experimental Days 9 hr group – 7.9 hrs 9 7 hr group – 6.3 hrs 5 hr group – 4.7 hrs Amount of Sleep (Hrs) 3 hr group – 2.9 hrs 7 5 9 HR 7 HR 3 5 HR 3 HR 1 T2 B E1 E2 E3 E4 E5 E6 E7 R1 R2 Day Washington State University

Psychomotor Vigilance Task Belenky et al., 2003 Washington State University

Driving Simulator Washington State University

Driving Simulator – Lane Deviation 2 3 Hr Deviation of Lane Position 5 Hr 7 Hr 9 Hr 1.5 1 0.5 T1 T2 B E1 E2 E3 E4 E5 E6 E7 R1 R2 R3 Day Washington State University

Individual Variability in Resistance to Sleep Restriction 3 Hours Sleep/Night X 7 0.005 Days Recovery Baseline Speed on PVT 0.004 0.003 0.002 0.001 Mean +/- SEM (n = 18) Resistant Subject 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Sensitive Subjet #1 Days Sensitive Subject #2 Washington State University

Chronic Sleep Loss: Objective and Subjective Effects Adapted from Van Dongen et al (2003)

Time on Task and Time Awake Effects on Performance during Sleep Deprivation and Sleep Restriction • Time on Task Effects during 38 Hours The Psychomotor Vigilance Task (PVT) is a of Total Sleep Deprivation Psychomotor Vigilance Task sensitive metric of human performance 3.40 – A reaction time test (PVT) Performance – Administered by PC, Palm OS PDA, or Window Pocket PC 3.00 PDA – ~10 stimuli present/min for 10 min 2.60 – Sensitive to sleep deprivation and sleep restriction N = 49 – 2.20 Sensitive to circadian periodicity – Sensitive to time on task 1.80 • Time on task effects develop over minutes and 0800 1200 1600 2000 0000 0400 0800 1200 1600 2000 Time (Hours) are reversed by simple rest (time off task) • Time awake effects develop over hours and days Time on Task Effects during 7 Days of Sleep Psychomotor Vigilance Task (PVT) and require sleep to reverse Restriction and Subsequent Recovery • Time on task interacts with time awake 4.5 • The effects of time awake on performance may 4 Performance be mediated through increasing sensitivity to 3.5 time on task 3 3-Hr • The PVT may be an excellent task to probe using 2.5 5-Hr N= 16 -18/group new techniques of brain imaging the “use - 7-Hr 2 9-Hr dependency” of the effects of time awake on 1.5 performance Baseline E1 E2 E3 E4 E5 E6 E7 R1 R2 R3 Time (Days)

Split Sleep and Napping and Sleep

Newark to Hong Kong – Over the North Pole Washington State University

Home, Layover, and In-Flight Sleep in a Boeing 777 Pilot Belenky, et al., in preparation

Transmeridian Travel: Actigraph Record  Overseas Travel  2/11: Leave Eastern US (1)  2/12 - 2/16: SWA (1)  2/17: Germany (1)  2/18 - 2/19: Hawaii (1)  2/20: Arrive Eastern US (2) (1)  (2) Sleep in afternoon (EST) (1) (1) and some divided sleep (2) (3) during time in SWA and Germany (4) Days  Sleep in mid-morning hours (4) (EST) (3) during time in (4) Hawaii (4)  Sleep during normal sleeping (4) hours (EST) (4) on return to (4) Eastern US Eastern Standard Time (EST)

Night Float vs. Day Shift in Physicians in Training Washington State University

Physician on Day Shift and Night Float Sequence Washington State University

Sleep Off Shift & On Shift / Day Shift vs. Night Float Day Shift Night Float Washington State University

Response Surface Mapping of PVT Lapses in Split Restricted Sleep

Consolidated vs. Split vs. Fragmented Sleep • Recuperative value of sleep depends on 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 Bonnet M & Arand D (2003) Clinical effects of sleep fragmentation vs. sleep deprivation. Sleep Medicine Reviews, 7(4) 297-310 Washington State University

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

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

Cockpit Napping

Other Countermeasures  Stimulants on shift Wesensten et al., 2005  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 Washington State University

Amphetamine vs. Modafinil vs. Caffeine 110 Drug @ 65 hrs sleep loss 100 Mean Relative Speed . 90 80 RECOVERY SLEEP 70 Placebo 60 d-Amphetamine 20 mg Caffeine 600 mg Modafinil 400 mg 50 0800 1600 0000 0800 1600 0000 0800 1600 0000 0800 1600 0000 0800 1600 Time of Day Adapted from Washington State University Wesensten et al., 2005

Modafinil vs. Caffeine 3.5 Drug or Placebo @ 2355 Mean Speed (1/RT * 1000) 3.0 2.5 Placebo 2.0 Modafinil 100 mg Modafinil 200 mg Modafinil 400 mg 1.5 Caffeine 600 mg DAY 2 DAY 3 DAY 4 1.0 0800 1200 1600 2000 0000 0400 0800 1200 1600 2000 0000 0400 0800 1200 Time of Day

Sleep and Performance in Operations

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

Guantanamo Bay, Cuba Washington State University

Crash Site All 3 crew members were rescued from the cockpit and survived Washington State University

The Approach to Guantanamo Approach to Guantanamo requires a sharp right bank to avoid Cuban air space Washington State University

Crash of American International Flight 808: Sleep Amounts Prior to Crash Landing 16 August 17 August 18 August 0000 0800 1600 0000 0800 1600 0000 0800 1600 Captain |_____|_____|_____|_____|_____|_____|_____|_____|__ 0000 0800 1600 0000 0800 1600 0000 0800 1600 |_____|_____|_____|_____|_____|_____|_____|_____|__ Co-PILOT 0000 0800 1600 0000 0800 1600 0000 0800 1600 |_____|_____|_____|_____|_____|_____|_____|_____|__ Engineer = reported sleep time

Accident Investigation – American International Flight 808 (1993) Captain: 71% Co-Pilot: 70% Engineer: 77%

Cockpit Voice Recorder just Prior to Crash ???: There you go, right there, lookin’ good. Engineer: Slow, Airspeed Captain: Where’s the strobe? Co-Pilot: Check the turn. Co-Pilot: Do you think you’re gonna make this? Captain: Where’s the strobe? Captain: Yeah… if I can catch the strobe light. Co-Pilot: Right over here. Co-Pilot: 500, you’re in good shape. Captain: Where? Engineer: Watch the, keep your airspeed up. Co-Pilot: Right inside there, right inside there. Co-Pilot: 140. [sound of stall warning] Engineer: You know, we’re not gettin’ our airspeed back there. ???: Don’t – stall warning. Captain: Where is the strobe? Captain: I got it. Co-Pilot: Right down there. Co-Pilot: Stall warning. Captain: I still don’t see it. Engineer: Stall Warning Engineer: #, we’re never goin’ to make this. Captain: I got it, back off. Captain: Where do you see a strobe light? ???: Max power! Co-Pilot: Right over here. ???: There it goes, there it goes! Captain: Gear, gear down, spoilers armed. ???: Oh no! Engineer: Gear down, three green spoilers, flaps, checklist

Crash of American International Flight 808: Probable Causes "The impaired judgment, decision-making , and flying abilities of the captain and flight crew due to the effects of 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

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

Traditional vs. Intervention Schedule Landrigan, et al. (2004) NEJM 351: 18, 1838-1848 Washington State University

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

Limiting Work Hours: Effect on Serious Medical Errors Landrigan, et al. (2004) NEJM 351: 18, 1838-1848 Washington State University

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 Barger et al., NEJM 2005 Barger et al., NEJM 2006 – Increased percutaneous injuries Ayas et al., PLoS 2008 Washington State University

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

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 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 of the three crew Runway 22  Similar error in 1993  Caught prior to take-off roll  By both pilots and air traffic controller

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 Washington State University

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

Accident Reconstruction: Timeline Return Leave FTX Attend Accident to FTX Funeral |____|____|____|____|____|____| 0000 0000 0000 0000 0000 0000 0000 Day 1 Day 2 Day 3 All-night Duty (no sleep) ~ 6.5 hrs sleep per night Washington State University

Accident Reconstruction: Setting Obstacle 1 Side Path Obstacle 2 ~ 1 mile Washington State University

Accident Reconstruction: Model Predictions Performance Prediction: 8 Hrs Sleep/Night 80 -10% 70 -24% -14% Performance Prediction: NCO’s Sleep/Wake History 60 Day 1 Day 2 Day 3 The Accident Washington State University

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

Trait Individual Differences in Vulnerability to Performance Impairment from Sleep Loss Adapted from Van Dongen et al (2004) time of day 08 12 16 20 00 04 08 12 16 20 00 25 n=7 20 cognitive impairment → PVT lapses 15 10 5 n=8 0 0 4 8 12 16 20 24 28 32 36 40 hours awake

Mismatch between Subjective Sleepiness and Objective Performance Deficits Adapted from Van Dongen et al (2004) time of day time of day 08 12 16 20 00 04 08 12 16 20 00 08 12 16 20 00 04 08 12 16 20 00 25 5 e l 20 a c 4 S s PVT lapses s 15 e n i p 3 e e 10 l S d r o 2 f n 5 a t S 1 0 0 4 8 12 16 20 24 28 32 36 40 0 4 8 12 16 20 24 28 32 36 40 hours awake hours awake

Individual Differences in Active-Duty Air Force Pilots during Simulated F-117 Extended Night Flights From Van Dongen et al (2006) left 720 degrees turn left 720 degrees turn roll performance roll angle performance over time Empirical Best Linear performance distribution Unbiased Predictors 1.8 0.8 1.6 N=10 relative performance flight path deviation 0.6 1.4 relative performance 0.4 1.2 0.2 1.0 0.0 0.8 -0.2 0.6 -0.4 0.4 -0.6 0.2 -0.8 0 1 2 3 4 5 6 7 8 9 10 11 0.0 subjects -6 -1 4 9 14 19 24 subjects A B C D E F G H I J time of day (Self-)selection mechanisms do not eliminate individual differences in vulnerability to sleep loss — even in highly specialized professions

Adapted from Van Dongen et al (2007) Individualized Performance Modeling Group-Average Model Individualized Model impairment → Subject A 24 h ahead performance prediction snapshots at 44h awake (time 03:30) impairment → Subject B past performance future performance performance prediction │ 95% confidence interval

Predicting Performance from Actigraphically- Derived Sleep Wake History Washington State University Spokane

The SPRC Two-Process Model: Components and Fit to Data 0 3 32 0 28 2 0 Homeostatic Process 24 Circadian Process 0 1 20 PVT Lapses Model Fit PVT lapses Process S Process C 1 Model fit 16 1 0 12 1 -1 8 1 4 1 -2 0 N=11 1 1 -3 -4 0 6 12 18 24 30 36 42 48 54 60 0 6 12 18 24 30 36 42 48 54 60 Time awake Time awake circadian circadian measured or measured or process process predicted light predicted light exposure exposure predicted predicted fatigue fatigue sleep sleep inertia inertia homeostatic homeostatic planned planned process process work/rest work/rest schedule schedule measured measured or predicted or predicted sleep sleep effect of effect of chronic chronic sleep loss sleep loss Washington State University 83

The Internal Body Clock (Circadian Rhythm) Kryger, Roth and Dement, Principles and Practice of Sleep Medicine, 2005 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 Washington State University

The New Science and Art of Fatigue Risk Management Washington State University

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, … all watched over by machines of loving grace.” sustained - Richard Brautigan (1963)

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

Example of Actigraph Record • An example of an Sleeping actigraph record recorded over 6 days. • This person slept from ~ 22:00 to 08:00. Waking

Effect of Sleep Loss on Performance on the Psychomotor Vigilance Test (PVT) 800 0 12 Hours Awake 600 0 12 Hours 400 0 Awake 200 0 0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 800 0 36 Hours Awake 600 0 36 Hours 400 0 Awake 200 0 0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 800 0 60 Hours Awake 600 0 60 Hours 400 0 Awake 200 0 From Doran 0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 800 0 et al. (2001) 84 Hours Awake 600 0 84 Hours Arch Ital Biol. 400 0 Awake 200 0 0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 RESPONSE NUMBER Washington State University

: At Home and In-Flight Sleep in a Boeing 777 Pilot Belenky, et al., in preparation

Crew Scheduling in Commercial Aviation is Controlled by Rules & Objectives Scheduling Software Optimizes Assignments Flight Time Limits Labor Agreements Other Rules & Objectives Alertness Model Flights to Staff Pilots Available From Romig & Klemets Rules & Objectives (2010) Presentation to NSF Must be Satisfied Sleep Health and Safety Conference

Flight Time Limits Alone Do Not Protect Alertness From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

Labor Agreements Add Extra Protection – At A Cost From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

A Model Within Existing Constraints Improves Alertness From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

Modeling Alone Improves Alertness & Productivity From Romig & Klemets (2010) Presentation to NSF Sleep Health and Safety Conference

Point of Contact 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 Washington State University

The New Science and Art of Fatigue Risk Management Washington State University

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 … all watched over by robotic systems machines of loving grace.”  Monitored, assisted, sustained - Richard Brautigan (1963) Washington State University

Sudden vs. Graceful Degradation • Sleep deprivation-induced orderly decreases in performance and productivity precede accidents and catastrophic failures Apparent Change Actual Change in Performance in Performance SAFE SAFE UNSAFE UNSAFE 11/6/2010