Implementing Evidence-Based Resident Work Schedules: Work

to Date and Future Directions

Children’s Hospital of PhiladelphiaNovember 2010

Christopher P. Landrigan, MD, MPHResearch Director, Children’s Hospital Boston Inpatient Pediatrics Service

Director, Sleep and Patient Safety Program, Brigham and Women’s HospitalAssistant Professor of Pediatrics and Medicine, Harvard Medical School

Disclosures

1) In the past 12 months, Dr. Landrigan has served as a paid consultant to: AXDev (non-profit organization), to assist in the development of a study of Shift Work Disorder; this work was supported by an unrestricted research grant from Cephalon, Inc. (commercial entity) to AXDev.

2) In addition, Dr. Landrigan has received monetary awards, honoraria, and travel reimbursement from multiple academic and professional organizations for delivering lectures on sleep deprivation and safety.

Patient Safety

• To Err is Human– Institute of Medicine Report, 1999

– estimated 44,000 to 98,000 deaths annually from adverse events

• Report notably silent on issue of sleep deprivation

Courtesy of D. Weaver, Univ Massachusetts Medical School, Worcester, MA

Human Circadian Pacemaker in Suprachiasmatic Nucleus (SCN) of Hypothalamus

SCN

Core Body Temperature

Plasma Melatonin

Eye BlinkRate

Slow EyeMovements

Stage 1 Sleep

Figure 1

Relative Clock Time (h)8 16 24 8 16

°C

36.5

37.0

37.5

pm

ol/L

0

100

200

% o

f 5-m

in e

pochs

0

8

16

# p

er 30-s

epoch

5

10

15

Elapsed Time Awake (h)

0 8 16 24 32

0

1

2

3

Relative Clock Time (h)8 16 24 8 16

Karolinska Sleepiness Scale

3

5

7

Cognitive Throughput

30

45

60

Elapsed Time Awake (h)

0 8 16 24 32

Probe Recall Memory Test

0

2

4

ms

# o

f A

ttem

pts

# o

f corr

ect P

air

s

200

500

1000

6000

Psychomotor Vigilance Performance

more

tir

ed

mean reaction timemedian reaction time

10% slowest reaction time

10% fastest reaction time

Cajochen et al. Am J Physiol 1999 National Transportation Safety Board Study (SS1995/01)

02468

101214161820

1:00AM

3:00AM

5:00AM

7:00AM

9:00AM

11:00AM

1:00PM

3:00PM

5:00PM

7:00PM

9:00PM

11:00PM

Num

ber o

f Acc

iden

ts

Temporal distribution of fatigue-relatedsingle vehicle truck accidents

• Biological Time of Day (circadian rhythms)

• Consecutive Waking Hours

• Night Sleep Duration

• Sleep Inertia

DETERMINANTS OF ALERTNESS AND PERFORMANCE

0 10 20 30 40 50

-30

-20

-10

0

10

20

Hours of Wakefulness

Su

bje

cti

ve A

lert

ness

Co

gn

itiv

e T

hro

ug

hp

ut

Hours of Wakefulness

Devi

atio

n fro

m M

ean (

mm

)

-30

-20

-10

0

10

0 10 20 30 40 50

Devi

atio

n fro

m M

ean (

# a

ttem

pte

d)

Jewett et al., 2000

Hours of Driving

Acute Sleep Deprivation and Performance

“...after [19] hours of sustained wakefulness (at 3 am) cognitive psychomotor performance decreased to a level equivalent to the performance impairment observed at a blood alcohol concentration of 0.05 %. ... After 24 hours of sustained wakefulness (at 8 am) cognitive psychomotor performance decreased to a level equivalent to the performance deficit observed at a blood alcohol concentration of roughly 0.10 %.” - D. Dawson and K. Reid, Nature 388: 235,

1997.

• Biological Time of Day (circadian rhythms)

• Consecutive Waking Hours

• Night Sleep Duration

• Sleep Inertia

DETERMINANTS OF ALERTNESS AND PERFORMANCE

Cumulative impact of daily sleep curtailmenton risk of vigilance lapses & subjective alertness

Van Dongen et al. Sleep 2003

Objective performance (PVT) Subjective awareness (KSS)

Pediatric Resident Chronic Sleep Deprivation and Performance

• “Heavy Call” (q5 schedule, ACGME compliant, average of 3h sleep on all) vs. Blood EtOH 0.04-0.05

012345678

mean lanevariability (feet)

mean speedvariability (mph)

Light callLight call w / alcoholHeavy call w / placebo

p=0.06

p=0.01

Arnedt et al. JAMA 2005

• Biological Time of Day (circadian rhythms)

• Consecutive Waking Hours

• Night Sleep Duration

• Sleep Inertia

DETERMINANTS OF ALERTNESS AND PERFORMANCE

Jewett et al., J. Sleep Res., 1999

Time Course of Deficits from Sleep Inertia

Fighters

All aircraft

1

2

3

4

Hours since 6 am wake time1 2 3 4 5 6 7

Aircraft Accident Data (Israeli Airforce)

Ribak et al., Aviat. Space Environ. Med., 1983

Sleep Inertia

Physiological Consequences of Healthcare Provider Schedules

Misalignment of circadian phase

Acute total sleep deprivation scheduled frequently

Chronic partial sleep deprivation resulting in cumulative sleep debt

Performance often required within minutes of awakening

Biological Time of Day (circadian phase)

Number of Hours Awake

Nightly Sleep Duration

Sleep Inertia

Sleep Deprivation and Errors in Detection of Cardiac Arrhythmias on ECG

9.64 ± 1.41

1.8 h

(0-3.8 h)

Sleep Deprived

p < 0.001

p < 0.001

Errors on ECG

sustained attention task

Sleep in prior 32 h

Medical Interns

5.21 ± 0.93

7.0 h

(5.5-8.5 h)

Rested

Friedman et al., N Engl J Med 285: 201, 1971

Grantcharov TP, Bardram L, Funch-Jensen P, Rosenberg J. BMJ 2001;323:1222-1223

Impaired speed and errors in

performance: laparoscopic

surgical simulator

•17-hour overnight call duty in a surgical department

•median reported sleep time 1.5 h (range 0-3 h)

Resident Performance and FatiguePhilibert I. Sleep 2005; 28: 1392-1402.

• Meta-analysis 60 studies (959 MDs, 1028 non-MDs) – For MDs, 24 hours with

no sleep leads to major performance drops to:

• 15th percentile of rested MD performance level

• 7th percentile on clinical tasks -4 -3 -2 -1 0 1 2 3 4

Standard Deviations

•Effect of Sleep Deprivation on Physicians’ Mean Clinical Performance: Results of 14 Studies

Harvard Work Hours, Health, and Safety Study

• National survey: To objectively quantify the work schedules experienced by house staff, and determine if increased hours are associated with increased risk of house staff injury– Study of a national sample of house staff

• 1,417 person-years monthly survey data collected from 2,737 interns nationwide in 2002-2003

– Monthly surveys– Work hours, crashes, and injuries– Correlation of work hours and motor vehicle crashes

Barger, L. K. et al. N Engl J Med 2005; 352:125-134

0

5

10

15

20

25

30

0 >0-1 >1-2 >2-3 >3-4 >4-5 >5-6 >6-7 >7-8 >8

Reported hours of sleep obtained during extended duration work shifts

Per

cen

tag

e o

f m

on

thly

rep

ort

s

Harvard Work Hours, Health, and Safety Study: Results

0

0.2

0.4

0.6

0.8

1

1.2

Crashes per 1000 commutes home

OR: 2.3 (95% CI, 1.6-3.3)

Barger LK et al. NEJM 2005; 352:125-134

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Injuries per 1000 opportunities

Motor Vehicle Crashes Percutaneous InjuriesOR: 1.6 (95%CI, 1.5-1.8)

Ayas, et al. JAMA 2006; 296:1055-1062

Extended shifts

Non-extended shifts

Dose-Response Relationship

• For each additional extended duration work shift scheduled per month, interns had:– 8.8% increased monthly risk of any

Motor Vehicle Crash; – 16.0% increased monthly risk of a Motor

Vehicle Crash on the commute home from work

• Therefore, q3 schedule (10 overnights per month) = 160% increase over baseline risk

Harvard Work Hours, Health, and Safety Study: Part 4

• Objective: To determine if interns report making more harmful medical errors when working 24-hour shifts– Odds of reporting a harmful fatigue-related error

was 7-fold higher when working five or more 24h shifts in a month (compared with self when working no 24h shifts)

– Odds of a fatal error due to fatigue 4-fold higher

Barger, L.K. et al. PLoS Medicine 2006;3:e487

02468

1 01 21 41 61 8

E r r o r w / A d v e r s eo u t c o m e

F a t a l e r r o r

0 2 4 h s h i f t s

1 - 4 2 4 h s h i f t s

> = 5 2 4 h s h i f t s

OR 7.0 (4.3-11) OR 4.1 (1.4-12)

Errors/ 1000 person-months

Harvard Work Hours, Health, and Safety Study

• 1 of every 5 interns admitted making a fatigue-related mistake that injured a patient (700% when interns worked >24 hours in a row)

• 1 of every 20 interns admitted making a fatigue-related mistake that resulted in a patient’s death (300% in months interns worked five >24 hour shifts)

Randomized Trial with the following EXPERIMENTAL QUESTION:

Would ICU patients fare better when the physicians caring for them consisted of:

1. Current standard TRADITIONAL team of 3 residents working on a Q3 schedule which minimized handoffs by relying on repetitive 30-hour scheduled work shifts; or

2. An INTERVENTION team of 4 residents working on a schedule which increased handoffs in order to limit scheduled work shifts to no greater than 16 hours

Intern Sleep and Patient Safety Study

Intern A

Intern B

Intern C

Clock time (h)

TueMonSunSatFriThuWed

Intern 1

Intern 2

Intern 3

Intern 4

B

A

0 12 0 12 0 12 0 120 12 0 12 0 12 0

0 12 0 12 0 12 0 120 12 0 12 0 12 0

Traditional (A) vs. Intervention (B) Intern Schedule

Landrigan, C. P. et al. N Engl J Med 2004;351:1838-1848

Results: Sleep and Work Duration

Lockley, S. W. et al. N Engl J Med 2004;351:1829-1837

Attentional Failures at Night: 11pm-7am

•0.69 (traditional) vs. 0.33 (intervention) attentional failures per hour, p=0.02

•Non-significant trend toward decreased day / evening attentional failures as well

0

20

40

60

80

100

120

140

160

SeriousMedical Errors

- Total

SeriousMedication

Error

SeriousDiagnostic

Error

Traditional "q3" 24-30hour shifts

Intervention Schedule- <16 hour scheduledshifts

Serious Medical Errors•Interns made 36% more serious errors on traditional schedule, including 5 times as many serious diagnostic errors

Landrigan, C. P. et al. N Engl J Med 2004;351:1838-1848

p<0.001

p<0.001

p=0.03

Landrigan, C.P. et al. N Engl J Med 2004;351:1838-1848

Err

ors

per

10

00 p

t da

ys

Effects of Reducing or Eliminating Shifts >16h on Patient Safety and Quality of Care: a Systematic Review

Levine AC, Adusumilli J, Landrigan CP. Sleep 2010; 33: 1043-1053Source Setting Intervention Outcomes

Afessa et al. Medical ICU Transition from Q3 to a 14 hour work shift

model

No change in mortality or LOS

Bhavsar et al. Cardiology service

Elimination of extended shifts for senior

residents (no change for interns)

Improved guideline adherence, decreased

LOS, and improvement in 6-month mortality

de Virgilio et al.

Trauma service

Decrease from average of Q4.8 call to Q6.4

No change in complication rate or

mortality rateGoldstein et

al.Surgical inpatient service

Transition from Q4 to night float system with

12-14 hour shifts

Improvement in patient and nurse ratings of care

qualityGottlieb et al. VA medicine

serviceTransition from Q4 to

schedule with maximum 16h shifts

Decrease in LOS and medication errors

Horwitz et al. Medicine service

Elimination of extended shifts for residents (no

change for interns)

Decrease in ICU admissions and pharmacy

interventions to prevent med errors

Effects of Reducing or Eliminating Shifts >16h on Patient Safety and Quality of Care: a Systematic Review (continued)

Levine AC, Adusumilli J, Landrigan CP. Sleep 2010; 33: 1043-1053Source Setting Intervention Outcomes

Hutter et al. Surgical service

Reduction from Q3 to Q4 call

No change in complication or mortality

ratesLandrigan et

al. Medical and

Cardiac ICUsComparison of Q3 and 16 hour shift systems

Decrease in serious medical errors

Malangoni et al.

Surgical service

50% reduction in call shifts per month per

resident

Decrease in mortality

Mann et al. Radiology Department

Elimination of extended shifts without sleep for

radiology residents

Decrease in errors reading films requiring patient call back to ER

Sawyer et al. Surgical service

Comparison of interns on Q2, Q3, and Q4

No difference in error rates

n Significant improvement

No change Significant decrement

Summary 11 7 4 0

Reduction vs. Elimination of Shifts >16 hours

• Of programs that reduced the frequency of shifts >16 hours but did not eliminate them (e.g., transition from Q3 to Q4 schedule), 1/4 (25%) saw improvements in safety or quality

• Of programs that eliminated 24h shifts, 6/7 (86%) saw improvements in safety or quality

Did the 2003 ACGME duty hour standards solve the

problem?

• <80 hours per week, averaged over 4 weeks

• <30 hours in a row, including time for hand-offs of care and education

• 1 day off in 7, averaged over 4 weeks

Harvard Work Hours, Health, and Safety Study: Compliance with Duty Hour Standards

Work and Sleep, Before and After

‡ p<0.001‡ p<0.001

Landrigan C.P., et al. JAMA 2006;296:1063-1070

Compliance with ACGME Duty Hour Standards

• 83.6% of interns in violation of standards– 85.4% of programs; 90.8% of hospitals

• 44.0 % of all intern-months in violation – 61.5% of inpatient intern-months in violation

‡ p<0.001

Landrigan C.P., et al. JAMA 2006;296:1063-1070

Pediatric Duty Hours StudyLandrigan et al., Pediatrics 2008;122;250-258

• Specific Aim: To comprehensively measure the effects of the ACGME Duty Hour Standards in 3 major pediatric programs on:– house officer work and sleep hours– patient safety– house officer health and safety– medical education– quality of life

Hours of Work and Sleep

0

10

20

30

40

50

60

70

80

Mean Weekly WorkHours

0123456789

101112

Mean Nightly Hours of Sleep

Pre-

Post-

P=0.49

P=0.94

Safety-sensitive US industries

• Truckers : maximum 11 hours in a row

• Pilots : maximum 8h per 24 (domestic routes)

• Nuclear Power : maximum 12 hours

• Train engineers: maximum 12 hours

What are the Personal and Legal Implications of Residents Working >24

Hours?

Resident Car Crashes

• Heather Brewster: Student permanently brain damaged after being rear-ended by a resident from Rush-Presbyterian in Chicago who fell asleep at the wheel after having been awake for 34 of the preceding 36 hours

• Dr. Valentin Barbulescu: Senior resident from New York who died in a one-car crash after falling asleep at the wheel post-call from the CCU

Potential Legal Implications for House Staff

• In New Jersey, “driving after having been without sleep for a

period in excess of 24 consecutive hours” now explicitly considered reckless

– Similar laws pending in several other states

• Brewster case - Illinois court system – appellate court found hospital not liable for resident driving home from work after 36-hour shift, but appeal to Supreme Court of Illinois may be forthcoming

Potential Legal Implications for Hospitals

• Courts in two states have ruled that an employers may be held responsibility for fatigue-related crashes even after workers have left– similar to concept of restaurants / bars being

potentially liable for alcohol-related crashes if they served alcohol to a driver

• Employers have been sued and settled for as much as $10 million in a similar case

What about the risks of lost continuity of care?

Problems in Care Continuity• Night float admission patients had longer LOS

and more tests ordered Lofgren et al. J Gen Intern Med. 1990

• Work hour limits and presumed resulting discontinuities associated with increased hospital complications and test ordering Laine et al. JAMA 1993

• Cross coverage associated with an increased risk of errors (OR 6.0) Petersen et al., Ann. Int. Med 1994

– Signout errors an be improved substantially with structured sign out Petersen et al., Joint Comm J on QI 1998

Communication During Post-operative Patient Hand Off in the Pediatric Intensive Care Unit

Mistry KP, Landrigan CP, Goldmann DA, Bates DWCritical Care Medicine 2005; 33: A12.

0

2

4

6

8

10

12

14

16

18

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

medianmean

Nu

mb

er

of T

ran

sfe

rs

Number of Miscommunication Events

-Audio recording and analysis of 150 post-op sign-outs-100% of sign-outs contained at least one error

Signout Process : A survey of residents at BWH

Carty M, Smith C, Schnipper JL, Harvard Education Day 2004, unpublished data

• 37% said that signout occurred in a quiet place most of the time

• 52% provided written and verbal signout on every patient

• Only 55% of night-float residents said that when called about a patient, the relevant information could be found in the sign-out

What about the effects of lost experience on learning?

Does sleep loss have an effect on Learning?

Memory Consolidation During Sleep – Motor Sequence Task

Stickgold R. Nature 2005; 437: 1272-8

Visual Discrimination Learning Requires Visual Discrimination Learning Requires Sleep after Training*Sleep after Training*

0

5

10

15

20

25

30

0 1 2 3 4 7

Imp

rovem

ent

(ms)

*Stickgold et al., Nature Neurosci 3:1237, 2000

SLEEP

SLEEP

SLEEP

NO SLEEP

0

5

10

15

20

25

30

0 1 2 3 4 7

Imp

rovem

ent

(ms)

SLEEP

SLEEP

NO SLEEP

Visual Discrimination Learning Requires Visual Discrimination Learning Requires Sleep after Training*Sleep after Training*

*Stickgold et al., Nature Neurosci 3:1237, 2000

Stages of Sleep and Memory Consolidation

• Different Stages of Sleep are Correlated with Consolidation of Different Types of Memory– Visual discrimination task: Slow Wave

Sleep (Stages 3 and 4) and REM– Motor sequence task: Stage 2– Motor adaptation task: Slow Wave Sleep

Localized Slow Wave Sleep increase in Cortical Areas of Activation

Ghilardi, M. et al. Brain Res. 2000; 871: 127–145

Complex Tasks and Medical Education

• Mathematical and Geometrical Puzzle Solving show same type of sleep dependent improvement (Kuriyama et al. 2004; Learn Mem. 11: 705–713; Wagner et al. Nature 2004; 427: 352–355)

– Thus, complex as well as simple cognitive abilities affected

• No basic / laboratory research on sleep deprivation’s effects on medical learning but a number of field studies have emerged– Residents working longer hours report decreased

satisfaction with learning environment and decreased motivation to learn Baldwin et al 1997

Systematic Review: Effect of Reducing or Eliminating Shifts > 16h on Medical Education

Levine AC, Adusumilli J, Landrigan CP. Sleep 2010; 33: 1043-1053Source Population Intervention Outcomes

Afessa et al. Internal Med Residents

Transition from Q3 to a 14 hour work shift

model

No change in performance on post

rotation examBarden et al. Surgical

Residents Reduction of ICU Call

from Q2 to Q3; elimination of most call

on other rotations

No change in case volume; improved

ABSITE exam scores

Carey et al. OB/Gyn Residents

Decrease from Q3 to Q7 call

Improved CREOG exam scores

Cockerham et al.

Surgical Residents

Elimination of call with night float system

Reduction in % time spent on non-patient care,

significance not measuredDe Virgilio et

al.Surgical

Residents25% reduction in call

frequencyNo change in ABSITE scores; improved case

volumeFerguson et

al.Surgical

ResidentsReduction in call

frequency from q3 to q4No change in case

volume

Goldstein et al.

Surgical Residents

Transition from q4 to night float system

No change in case volume

Systematic Review: Effect of Reducing or Eliminating Shifts > 16h on Medical Education (cont)

Levine AC, Adusumilli J, Landrigan CP. Sleep 2010; 33: 1043-1053Source Population Intervention Outcomes

Hutter et al. Surgical Residents

Reduction in call frequency from q3 to q4

No change in ABSITE scores; attendings rated

resident skills lowerJarman et al. Surgical

Residents Transition from q3 and q4 to night float system

No change in percent of resident cases

Kelly et al. OB/Gyn Residents

Transition from q3 and q4 to night float system

No change in volume or CREOG exam scores

Malangoni et al.

Surg PGY 4/5 Residents

50% reduction in calls No change in percent of resident cases

McElearney et al.

Surgical Residents

Elimination of extended shifts with night float

No change in case volume

Sawyer et al. Surgical Interns

Comparison of interns on Q2,Q2,Q4

Improved case volume on Q4

Welling et al. Surgical Residents

Comparison of Q4 to night float system

No change in conference attendance

n Significant improvement

No change Significant decrement

Summary 14 4 9 1

Moving Forward

Resident Duty Hours: Enhancing Sleep, Supervision, and Safety

• Study commissioned by Congress, released by IOM in December 2008

• Concluded that it is unsafe for residents to work over 16 hours in a row without sleep

• Two solutions proposed:–Mandatory 5h nap during a 30h shift

or–16h shift limit

• Also called for improved handoff processes, increased supervision, and ongoing research to test best practices for implementation

Proposed 2011 ACGME Duty Hour Standards

• 16h consecutive work limit for Interns• PGY2s and higher continue to work 28h

shifts• Development of standards for supervision,

workload

European Working Time Directive

- 13 hours in a row maximum

- 48-56 hours per week

New Zealand

- 72 hours per week limit

- 16 hours in a row

- in place for 20 years

How do we move to a 16h Shift Limit for Interns?

• Must design schedules that respect principles of sleep and circadian medicine– Limit nights in a row; sufficient time off after night shift for

recovery; proven rotation patterns• Must not excessively overburden those residents

who remain on duty as others are home• Must improve sign-outs, and infrastructure for sign-

outs• Improve supervision• Education will need substantial redesign

– Changes in timing of didactic education• Creative scheduling• Less “on the fly” teaching; more structure required• ? Increased use of simuation / web

– ? Changes in rounding patterns

Slide courtesy of J Rothschild, MD, MPH

Rates of Complications by Attending Physicians After Performing Nighttime Procedures

Rothschild JM et al., JAMA 2009; 302(14):1565-1572

Remember…..

• Adjusting to a new shift system is not easy• Invariably, the first few steps are difficult and

adjustments to an intervention schedule need to be made– Iterative improvement is the key

• Infrastructure needs to change to reflect new schedules

• Changing lifestyle / sleep hygiene patterns is difficult but can be accomplished

Acknowledgements

• Harvard Work Hours, Health, and Safety Group– Especially Charles Czeisler, Steve Lockley, Laura

Barger, and Najib Ayas

• Center of Excellence in Patient Safety at BWH– Especially Jeffrey Rothschild and David Bates

• Children’s Hospital Boston, Brigham and Women’s Hospital and Harvard Medical School

• AHRQ and NIOSH