The interaction of OSA, Fatigue, & Sleep:

A Performance Model

November, 2011

Steven R. Hursh, Ph.D

Operational Definition

DOT Human Factors Coordinating Committee, 1998

● Fatigue is more than sleepiness and its effects are more than falling asleep.

● Fatigue is a complex state characterized by a lack of alertness and reduced mental and physical performance, often accompanied by drowsiness.

2

Symptoms versus Root Causes

Symptoms

Operational Consequences

● Measurable Changes in Performance

● Lapses in attention and vigilance

● Delayed reactions

● Impaired logical reasoning and decision-making

● Reduced “situational awareness”

● Low motivation to perform “optional” activities

● Poor assessment of risk or failure to appreciate consequences of action

● Operator inefficiencies

● Fatigue is one

potential root cause.

● No direct measure,

physiological marker,

or “blood test” for

fatigue.

3

Root Cause Analysis

Modeling Provides an Objective

Metric for Fatigue ● 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 fatigue risk.

4

5

ALERTNESS & COGNITIVE PERFORMANCE

CIRCADIAN

RHYTHM CUMULATIVE

SLEEP DEBT

ALERTNESS &

COGNITIVE

PERFORMANCE

Time of Day Sleep History and Time on Duty

Daily Variations in Effectiveness

SAFTE

● The Sleep, Activity, Fatigue, and Task Effectiveness (SAFTE) Model is based on 17 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 warfighter fatigue model.

6

SAFTE Model Components

7

Practical Software for Fatigue

Assessment and Management

● Fatigue Avoidance Scheduling Tool (FAST)

● FAST is a fatigue assessment tool using the SAFTE model

● Developed for the US Air Force and the US Army.

● Department of Transportation has sponsored work leading to

enhancements for transportation and industrial applications.

● DOT validated and calibrated.

8

Now available as a tool for commercial applications.

9

Walter Reed Restricted Sleep Study PVT Speed

Chronic Restriction Adaptation

50

65

80

95

110

0 T1 T2 B E1 E2 E3 E4 E5 E6 E7 R1 R2 R3Day

Mea

n S

pe

ed

(as a

% o

f B

ase

line

)

9 Hr

7 Hr

5 Hr

3 Hr

PVT Speed

Chronic Restriction Adaptation

50

65

80

95

110

0 T1 T2 B E1 E2 E3 E4 E5 E6 E7 R1 R2 R3Day

Me

an

Sp

ee

d

(as a

% o

f B

ase

line

)

9 Hr

7 Hr

5 Hr

3 Hr

SAFTE/FAST R2 = 0.94

Restriction Recovery Baseline

Belenky, G, Wesensten, NJ, Thorne, DR, Thomas, ML, Sing, HC, Redmond, DP, Russo, MB, and Balkin, TJ (2003). Patterns of performance

degradation and restoration during sleep restriction and subsequent recovery: a sleep dose-response study. Journal of Sleep Research, 12, 1-12.

Two-Step Sleep/Performance Model

Sleep Input

Actigraph or Logbook

Physiological Model Sleep

Estimator

Work Schedule 10

OSA

Questions

1. Can actigraphy detect sleep disturbances

produced by sleep apnea?

2. Can actigraphy detect differences in

severity of sleep apnea?

3. Do these differences translate into

meaningful differences in estimated sleep

and predicted performance?

4. Will exposure to feedback based on

actigraphy and predicted performance

reinforce greater CPAP use and adherence

to treatment protocols?

Single Day of Activity, Example

Noon, March 24 to noon, March 25

Major Sleep Period – Time in Bed

Complete 24 hr Records for the Entire Month

Before CPAP

After CPAP

Focus on Sleep Periods Only

Before CPAP

After CPAP

Pre- and Post-CPAP Activity Patterns

Post-CPAP: Noon, April 12 to noon, April 14

Major Sleep Period – Time in Bed

Pre-CPAP: Noon, March 24 to noon, March 25

Major Sleep Period – Time in Bed

Sleep Quantity per Day

Sleep Hours Per Day (Two Days Means)

0

2

4

6

8

10

12

3/18/2006 3/23/2006 3/28/2006 4/2/2006 4/7/2006 4/12/2006 4/17/2006

Days

Ho

urs

Untreated

Treated

CPAP TreatmentUntreated Sleep Apnea

Sleep Fragmentation Before and After

Major Sleep Episodes per Day (Separated by > 2.5 min Awake)

0

2

4

6

8

10

12

14

16

3/18/2006 3/23/2006 3/28/2006 4/2/2006 4/7/2006 4/12/2006 4/17/2006

Days

Ep

iso

de

s

Untreated

Treated

CPAP TreatmentUntreated Sleep Apnea

How does this translate to better

brain functioning and Performance?

● The principle effect of sleep loss is reduced

cognitive ability and performance lapses.

● Difficult to measure directly.

● Fatigue model can predict how the “average”

person’s performance would improve with

increased of sleep.

● Performance feedback could be provided

daily to reinforce adherence.

● Does this additional information matter?

0

200

400

600

800

1000

1200

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

mse

c

PVT trial

Lapses (RT > 500ms) Mean = 6.07 SD = 6.61

(Note: the ‘down’ periods (green) were calculated with the automatic option – to illustrate how this calculation fails for this data set)

Hypothetical “Normal” Sleeper Performance while awake (black line) is always above 90%

Before CPAP After CPAP

12 Hr Sleep

Performance Estimate Based on Sleep Recording Performance gradually improves to near ideal after the CPAP

Performance Effectiveness

Lapses

Daily Performance Before and After CPAP

0

10

20

30

40

50

60

70

80

90

100

110

3/2

0/2

00

6

3/2

1/2

00

6

3/2

2/2

00

6

3/2

3/2

00

6

3/2

4/2

00

6

3/2

5/2

00

6

3/2

6/2

00

6

3/2

7/2

00

6

3/2

8/2

00

6

3/2

9/2

00

6

3/3

0/2

00

6

3/3

1/2

00

6

4/1

/20

06

4/2

/20

06

4/3

/20

06

4/4

/20

06

4/5

/20

06

4/6

/20

06

4/7

/20

06

4/8

/20

06

4/9

/20

06

4/1

0/2

00

6

4/1

1/2

00

6

4/1

2/2

00

6

4/1

3/2

00

6

4/1

4/2

00

6

4/1

5/2

00

6

4/1

6/2

00

6

Effectiveness

Percent Below 77Pre-CPAP Post-CPAP

Fully Rested Range

.08 BAC

.05 BAC

Percent of Awake Time Below 77 (0.05 BAC)

Average Effectiveness

0

2

4

6

Pre-CPAP Post-CPAP

Fully Rested Range

Lapse Likelihood

Daily Tendency to Have Lapses

Proposed Studies

● Study 1: Descriptive analysis of actigraph

recordings of patients with sleep apnea

before and after CPAP treatment.

● Study 2: Controlled study of the benefits of

sleep and performance feedback.

● Protocols have been developed and study is

planned for the near future.

Questions?

Steven R. Hursh

shursh@ibrinc.org

Experiment 1: Actigraphy Measures of

Sleep Apnea and CPAP Therapy Technological Proof-of-Concept/Validation

● Enroll equal numbers of patients diagnosed as Mild, Moderate, &

Severe OSA. Case-control so that all groups are equivalent in as many

parameters as possible prior to study. Ideally, we'd have a similarly

case-controlled group of non-sleep-disordered subjects, but not

necessary.

● Outfit each subject with Actigraph for continuous monitoring and

provide a laptop or netbook computer for data upload:

1-2 weeks of no CPAP baseline followed by 2-4 weeks of CPAP usage (realistically,

one week of baseline + 2 weeks CPAP)

No feedback, simply record and upload the data

Analyze sleep parameters as a function of OSA severity with covariate of compliance

rate.

● Answer the simple question: Does actigraphy detect sleep benefits of

CPAP in OSA patients? If it does not, then feedback would be arbitrary

and irrelevant. If it does, then proceed to Experiment 2.

0

200

400

600

800

1000

1200

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

mse

c

PVT trial

Lapses (RT > 500ms) Mean = 1.28 SD = 1.44

Experiment 2: Effects of Actigraphy

Feedback on CPAP Compliance

● Enroll equal numbers of Mild, Moderate, & Severe OSA patients. Case-

control and similar to Experiment 1.

● Outfit each subject with Actigraph for continuous monitoring and

provide a laptop or netbook computer for data upload.

● Randomly assign half to Feedback or No Feedback conditions.

The No Feedback condition is identical to Exp 1.

In the Feedback condition, upon daily data upload, each subject receives a computer-

generated report on their laptop with the following individualized data: Line chart of total sleep, sleep efficiency, and wake episodes for each day of the study (accumulates as study period progresses)

SAFTE/FAST chart of predicted performance effectiveness (accumulates as study period progresses)

Performance effectiveness: % change from pre-CPAP baseline and % change from previous day

● If we want to save time and money, and we're not worried about time-of-

year or other cohort effects, we could simply compare the feedback

groups in Exp 2 to the subjects in Exp 1 who received no feedback, so

long as they are equally case-controlled.

● Answer the Question: Does sleep and performance feedback

increase compliance with CPAP therapy.