SAFTE/FAST Evidence-based Aviation Fatigue Risk Management Steven R.

SAFTE/FAST Evidence-based Aviation Fatigue Risk Management Steven R.

SAFTE/FAST Evidence-based Aviation Fatigue Risk Management Steven R. Hursh, Ph.D. President, IBR and Professor, Johns Hopkins University School of Medicine September 1, 2011 Copyright 2010, 2008, IBR Major Fatigue Factors Time of Day: between midnight and 0600 hrs. Recent Sleep: less than eight hours in last 24 hrs. Continuous Hours Awake: more than 17 hours since last major sleep period. Cumulative Sleep Debt: more than eight hours accumulation since last full night of sleep (includes disrupted

sleep). Copyright 2008, IBR An Objective Fatigue Metric No Blood Test for fatigue, yet The conditions that lead to fatigue are well known. A fatigue model simulates the specific conditions and determines if fatigue could be present. The model can estimate the level of degradation in performance and provide an estimate of schedule induced fatigue risk. Copyright 2008, IBR SAFTE The Sleep, Activity, Fatigue, and Task Effectiveness (SAFTE) Model is based on 12 years of fatigue modeling experience. Validated against laboratory and simulator measures of fatigue. Validated and calibrated to predict accident risk by the Department of

Transportation. Peer reviewed and found to have the least error of any available fatigue model. Accepted by the US DOD (Air Force, Army, Navy, Marines) as the common Copyright 2008, IBR SAFTE Model Components Copyright 2008, IBR Least Error for Conditions of Sleep Restriction 2002 Seattle Fatigue & Performance Modeling Workshop, of all models tested against laboratory measurements: SAFTE had least error predicting objective vigilance performance. SAFTE had least error predicting subjective ratings of fatigue. Advances since 2002 have further validated the sleep and performance assumptions and prediction of accident risk and severity.

Von Dongen, Aviation, Space, and Environmental Medicine, March, 2004, vol. 75, no. 3, 6 section II Copyright 2008, IBR Accuracy of Predicting Sleep Pattern and Duration Measure Signalman Maintenance of Way Dispatchers (less night workers) Train and Engine Mean Agreement 92% 92%

90% 88% Daily Sleep (Est.- Log) -24 min -21 min -3 min - 10.8 min 7 Copyright 2008, IBR Economic Risk and Effectiveness 6 Railroad Accident Relative Risk

Economic Risk - Damage & Casualty Cost 5 516% Accident Risk Relative Risk 4 3 2 1 = Unchanged relative risk $ 103%$ # 1

70% 0 142% # 25% Greater than 90 No Fatigue $ # 49% 77 to 90 Less than or equal to 77 Crew Effectiveness Score High Fatigue

Copyright 2008, IBR Validated Fatigue Modeling Tools SAFTE/FAST Fatigue Science has the exclusive license from the US Army to commercialize SAFTE model. Copyright 2008, IBR Practical Software for Implementation Fatigue Avoidance Scheduling Tool (FAST) Fatigue assessment tool using the SAFTE model Developed for the US Air Force and the US Army FAST Features DOT / FRA sponsored work has lead to Sleep estimation algorithm enhancements for transportation Graphical analysis tools applications

Dashboard of fatigue factors Data based of all effectiveness scores Copyright 2008, IBR FAST Aviation Sleep Estimation Accurate estimation of sleep is critical: Measure: actigraphy or log books Estimate: algorithm to simulate sleep behavior Aviation specific estimates that can be refined with actrigraphic measurement. Considers time zone changes and is valid for any city pair. Copyright 2008, IBR Sleep Estimator Tailored to Aviation Environment: Copyright 2008, IBR

FAST Aviation Specific AutoSleep Mimics typical sleep patterns Tailored to workgroup and schedule demands Considers total duty period and commuting Naps prior to anticipated late starts Considers time zones Slits sleep when appropriate Automatically inserts in-flight sleep for augmented crews. User defined parameters: amount of augmentation and quality of sleep environment 13 Copyright 2008, IBR Fatigue RiskImprovement Management Process System Continuous Measure Define the situation

Schedule evaluation Actigraph recordings Monitor Assess operational indicators Individual self-evaluation Feedback to process Model & Analyze FRM Steering Committee Involves all stakeholders at each stage: management, labor, aided by science Model the fatigue problem Analyze sources and Fatigue factors SAFTE/FAST

Modify/Mitigate Manage Shared Responsibility Operating practices Labor agreements Individual life style Collaborate for solutions Obtain commitment to solve problem Enablers Employee training Medical screening Economic analysis Technology aids Copyright 2008, IBR FAST Aviation Fatigue Assessment Process

Airline Specific Schedule Database XML format City-pairs/Trips or 30-day Bids Aviation AutoSleep SAFTE Model Output results to folder Links to Manager Standard FAST schedule is created by FAST Aviation Modeler FAST Aviation Manager FAST Aviation Modeler

Translation Tools available for any scheduling system Sorts by Criterion Displays results Links to Analyzer Fleet level reports FAST Analyzer Individual Schedules Examine schedules Effectiveness Graph

Fatigue Factors What-If Drills Individual reports Modular Process for Speed and Flexibility Copyright 2008, IBR FAST Aviation Modeler Model the schedules: 1. Set AutoSleep parameters if necessary 2. Name the Output folder a unique name 3. Choose either a City Pairs file or a Bid Schedules file for modeling Airline Schedule Database Processed Schedules Modeling Results Copyright 2008, IBR FAST Aviation Manager Sort by: Flight Time Below Criterion Level (FTBCL)

Critical TBCL (30 min associated with takeoff/landing) Average Effectiveness Minimum Effectiveness overall Minimum Effectiveness during critical periods Maximum Workload (high workload score in schedule) Median Workload (central score across schedule) Save table to text file Click on any line in Aviation Manager and schedule opens in FAST for detailed analysis. Copyright 2008, IBR If fatigue is present, what do you do about it? Modeling tools must do more than give you a fatigue score: It must estimate fatigue risk It must show detail of each schedule It must calculate fatigue factors It must suggest conditions that lead to fatigue so mitigations can be implemented by an FRMS

18 Copyright 2008, IBR Detailed Analysis Results Creates detailed database that shows: All duty periods and estimated sleep intervals Effectiveness in each half hour of each duty period Effectiveness at each half hour of the clock Distribution of duty time in effectiveness categories Allows results be sorted based on user defined categories Individual ID reports with 19effectiveness at Copyright 2008, IBR FAST Aviation Analyzer San Francisco to Sydney Pairing 14.5 hr flight Pre-flight nap

Sleep Timing based on both physiological and social cues In-flight sleep Dashboa Fatigue Pre-flight nap Schedules in Aviation Manager link to FAST for detailed analysis. Copyright 2008, IBR Dashboard Information Content based on fatigue analysis workshop hosted by NTSB and conducted by Drs. Mark Rosekind & David Dinges, funded by FRA Office of Safety. Sleep (last 24 Flags are fatigue indicators

hrs) Chronic Sleep Debt Hours Awake Time of Day Out of Phase Performance Values Value at point in schedule Criteria Effectiveness (vigilance) Mean Cognitive Lapse Index Reaction Time

Reservoir Copyright 2008, IBR Schedule Files for Evaluation A: 90 Short haul pairings, 1094 active flights B: 56 Short haul monthly rosters, 3963 active flights C: 47 Long haul pairings, 188 active flights D: 64 Long haul monthly rosters, 1006 active flights Translated the spread sheets using Access database into the required XML file structure. Batch processed through FAST Aviation Used FAST Manager to rank order Pairings and Rosters Used output spread sheet to rank order segments 22 Copyright 2008, IBR Fatigue Metrics Typically, FAST is used to assess the tail of the distribution how much critical duty time is spent at low effectiveness. For this exercise, we were asked to rank order all segments and schedules, not just

the extreme cases. We rank ordered segments by minimum effectiveness at critical times of flight takeoffs and landings. We rank ordered pairings and rosters by minimum effectiveness and critical time 23 below criterion which is more useful for Copyright 2008, IBR SAFTE/FAST - Segment Analysis A SH Pairings Rank 1 2 3 4 5 A10040 A10002

A10045 A10002 A10002 Start End Min Critical Time Minimum E Rank LAX LPB ATL VVI IQQ ATL IQQ SEA LPB SCL 5/15/2010 5:55

2/4/2011 8:15 5/14/2010 1:30 2/4/2011 6:30 2/4/2011 10:10 5/15/2010 10:08 2/4/2011 9:19 5/14/2010 6:57 2/4/2011 7:44 2/4/2011 12:24 70.84 74.64 74.8 74.99 75.65 3 7 8 9 10 ANC PDX

DFW BRU LAX LAX NRT BRU SHJ AMS 9/16/2010 15:05 5/17/2010 21:10 9/23/2010 18:15 9/27/2010 16:40 9/16/2010 23:10 9/16/2010 20:09 5/18/2010 7:59 9/24/2010 3:34 9/27/2010 23:09 9/17/2010 9:49 70.32 70.45 72.06

73.62 73.81 1 2 4 5 6 ARI IQQ ANF ARI IQQ SCL SCL SCL SCL SCL 12/22/2010 5:45 1/10/2011 5:30 1/8/2011 3:40 1/16/2011 2:55

1/10/2011 4:25 12/22/2010 8:09 1/10/2011 7:44 1/8/2011 5:29 1/16/2011 5:19 1/10/2011 6:39 72.59 72.84 73.66 74.17 74.39 6 7 8 9 10 NRT 5/26/2010 5:00 5/26/2010 8:54 SCL 1/4/2011 6:15

1/4/2011 9:34 ANC 5/18/2010 3:35 5/18/2010 8:29 HEL 11/17/2010 17:15 11/18/2010 4:19 ANC 6/2/2010 3:35 6/2/2010 8:29 66.82 67.76 67.78 68.59 68.88 1 2 3 4 5 C LH Pairings 1 2 3 4

5 C10027 C10014 C10033 C10025 C10030 B SH Rosters 1 2 3 4 5 B10049 B10046 B10052 B10049 B10044 1 2 3 4

5 D LH Rosters D10029 GUM D10032 LIM D10024 SLC D10052 HKG D10024 SLC 24 Copyright 2008, IBR SAFTE/FAST - Roster Analysis Rank A SH 1 2 3 4 5

Pairings A10002 A10001 A10008 A10040 A10000 Critical TBCL (77) TBCL Overall Rank 149 2 88 4 61 6 53 9 41 10 C LH 1 2

3 4 5 Pairings C10014 C10033 C10019 C10038 C10027 182 143 77 61 61 1 3 5 6 6 3 4

7 1 2 B SH 1 2 3 4 5 Rosters B10049 B10050 B10046 B10045 B10047 201 194 126 124 123 1

2 4 5 6 1 6 2 8 5 D LH 1 2 3 4 5 Rosters D10024 D10043 D10044 D10022 D10029

180 111 91 85 79 3 7 8 9 10 3 9 6 16 1 Minimum E Rank 2 4 5 1 8

25 Copyright 2008, IBR Minimum Critical Time Effectiveness SEGMENTS 26 Copyright 2008, IBR Short Haul Pairing A 10040 LAX to ATL Segment Early Start Daytime Rest Night flight 27 Copyright 2008, IBR LH Pairing - C 10027 Anchorage-LAX Segment 86% reservoir 5 hrs

28 Copyright 2008, IBR LH Pairing - C 10014 Narita Segment Narita 29 Copyright 2008, IBR Short Haul Pairing - A 10002 Santa Cruz, Bolivia Santiago, Chile 1.5 hr Nap 30 Copyright 2008, IBR A 10002, continued Santa Cruz, Bolivia Santiago, Chile

Possible Mitigation 3 hr Nap 31 Copyright 2008, IBR Long Haul Pairing - C 10014 Honolulu to Salt Lake 93% Res 3:15 Base Narita Honolulu Salt Lake 32 Copyright 2008, IBR C 10014 Honolulu to Salt Lake Altered Sleep Pattern

Altered Sleep Pattern 33 Copyright 2008, IBR Long Haul Pairing C 10033 Kuala Lumpur Based Pilot 34 Copyright 2008, IBR Greatest Time Below Criterion ROSTERS 35 Copyright 2008, IBR Long Haul Roster - D 10029 11 Day - Closer Examination 36 Copyright 2008, IBR

D 10029 Detroit Narita- Guam Segment 37 Copyright 2008, IBR Long Haul Roster D 10029 Guam Narita Segment 38 Copyright 2008, IBR Short Haul Roster - B 10049 92 Segments in 57 Days 25th Ranking Workload of 56 39 Copyright 2008, IBR B 10049 Early Start on 12/19 40

Copyright 2008, IBR B 10049 Multiple Segments at Night on 12/21 starting 1730 to 0510 41 Copyright 2008, IBR B 10049 Multiple Segments at Night 42 Copyright 2008, IBR Long Haul Roster - D 10024 Salt Lake-Anchorage-Minn 43 Copyright 2008, IBR D 10024 Six Day Series Salt Lake-Anchorage-Minn 44

Copyright 2008, IBR D 10024 Salt Lake-AnchorageMinn Explanation Two Early Starts (0700 and 0500) Two Consecutive Night Flights (2235 and 0020) Daytime Recovery 6 hrs 5 hrs 45 Copyright 2008, IBR Workload Factor According to the NTSB: One factor that contributes to self-reported pilot fatigue, especially in short-haul flight operations, is the number of legs flown in a duty period.* The highest workload in a flight occurs at take-off and landing; increasing segments multiplies these

high stress periods. FAST Aviation is the first fatigue *NTSB Safety Recommendation, A-09-61 through -66, August 7, 2009 Copyright 2008, IBR Sample Workload Pattern Workload Calculation 14 12 Maximum 6 Segments me 4 2 Workload

ti er ov o W e as e cr s in ent a d gm o l r k se s ate sip dis ad 8

th klo i w s or W Number or Factor 10 Median 0 1 2 3

4 5 6 7 8 Days from Start of Schedule Copyright 2008, IBR B 10007 Top Ranking Workload 58 Segments in 12 Days 48 Copyright 2008, IBR Advantages of Modeling Approach Validated model with history of outstanding performance under

independent review. Explicit Sleep Estimator (AutoSleep) tailored to habits and policies of the airline. Aviation specific drivers of fatigue. Cognitive fatigue Workload related fatigue Analysis tools that lead to specific fatigue factors and mitigation Copyright 2008, IBR Summary Copyright 2008, IBR Fatigue RiskImprovement Management Process System Continuous Measure Define the situation Schedule evaluation Actigraph recordings

Monitor Model & Analyze Assess operational indicators Model the fatigue problem Involves all Mitigations are stakeholders Proportional to the Risk Individual self-evaluation Analyze sources and at each stage: Feedback to process Fatigue factors Evolutionary,management, Incremental labor,Improvement by science Responsive toaided Changing Circumstances

Modify/Mitigate Manage Shared Responsibility Operating practices Labor agreements Individual life style Collaborate for solutions Obtain commitment to solve problem Enablers Employee training Medical screening Economic analysis Technology aids Copyright 2008, IBR Ev

en tS eq ue nc e Accidents & Incidents D Fa iag tig no ue sis M ea M M od su el re Fa ea in

s s tig ur g & e u B e s& M ar M e Fa a rie B o s d a rs tig ur el rr e i

u i ng er M e s& s e M Fa a B o tig sur de ar ue es lin rie g rs M & od Ba el rr in ie g rs Fatigue Risk Pyramid Fatigue Related

Errors Subjective Awareness Job Performance Changes Level 3 Defense Employee sleep habits, traits, & conditions Level 2 Defense Work demands, schedules, and sleep opportunities Level 1 Defense Based on James Reason, Managing the Risks of Organizational Accidents, Figure 1.6, Stages in the development and investigation of an organizational accident. Copyright 2008, IBR If fatigue is present, what do you do about it? Modeling tools must do more than give you a fatigue score: It must estimate fatigue risk It must show detail of each schedule

It must calculate fatigue factors It must provide context of conditions that lead to fatigue so mitigations can be implemented 53 Copyright 2008, IBR Questions: Steven R. Hursh, PhD and Reid Blank Institutes for Behavior Resources 2104 Maryland Avenue Baltimore, MD 21218 (410) 752-6080 [email protected] [email protected] Chris Hallman Baines Simmons America 17 Greenville St., Suite 221 Newnan, GA 30263 (678) 343-1635 Office (770) 251-5654 Fax [email protected] www.safetyfromknowledge.com 54

Copyright 2008, IBR FS Actigraph Data Processing Individual fatigue risk levels are amalgamated into a group report SAFTE evaluates the fatigue risk and effectiveness of each individual driver Personnel wear the actigraph that measures wrist movements Daily sleep/wake/work times are fed into the SAFTE risk evaluation model Wrist movements are recorded 24/7 and downloaded over the internet Downloaded data are converted to daily sleep/wake/work

times Copyright 2008, IBR

Recently Viewed Presentations

  • Seminario STORIA DEI PAESI ISLAMICI Professoressa Maria ...

    Seminario STORIA DEI PAESI ISLAMICI Professoressa Maria ...

    in the general public or in a group of persons or particular persons, intimidate a population or compel a government or an international organization to do or to abstain from doing any act, which constitute offences within the scope of...
  • Osvrt na Prvi dijabetoloski kongres FBiH

    Osvrt na Prvi dijabetoloski kongres FBiH

    glycated haemoglobin (HbA. 1c; the long-term average glucose level)Central to discussion is the emphasis on monitoring and treating PPG elevations, given that fasting and HbA. 1c. measures have traditionally been principal targets for treatment. .
  • Paragraphing

    Paragraphing

    Paragraph Unity - the entire paragraph should revolve around a single idea or focal point. If it begins with one idea, it should not end with another. ... Problem to Solution Pattern - Introduce a problem, explain it briefly, and...
  • Kent Chemical Debriefing - Weebly

    Kent Chemical Debriefing - Weebly

    Kent Chemical Products management were not involved in decisions being made by the international subsidiaries. Changes were being made at facilities without regard of how it may affect other facilities around the world. The new financial operating system.
  • Medical Language Second Edition CHAPTER 3 Gastroenterology Medical
  • Ocular Status, Health Seeking Behaviors and Barriers to

    Ocular Status, Health Seeking Behaviors and Barriers to

    Among 36.6% abnormality about 73% of them never went to an eye care specialist,: financial constraints (16.30%), didn't feel necessary (33.70%), infrequently other reasons were lack of escort, lack of time, traditional belief and not aware of hospital doctor was...
  • PARODY, SATIRE, & POETRY - The Last Saskatchewan Pirate ...

    PARODY, SATIRE, & POETRY - The Last Saskatchewan Pirate ...

    A parody is a humorous imitation of a (usually) serious piece of writing. Often people call it a 'takeoff' because the writer borrows the form or the topic of a well-known work and copies it in a funny way. Also...
  • Northwest Northeast Knox Maries Maries Cedar Polk Camden

    Northwest Northeast Knox Maries Maries Cedar Polk Camden

    Clark Harrison Nodaway Gentry Sullivan Adair Knox Grundy Holt Lewis Andrew Daviess De Kalb Linn Macon Shelby Marion Livingston Caldwell Clinton Buchanan Ralls Monroe Randolph Platte Pike Ray Clay Urban Region Montgomery Lafayette Lincoln Jackson St. Louis County Urban Region...