Keywords
call - night float - home call - ophthalmology residency - postcall relief - resident
sleep
Background
Since the implementation of Accreditation Council for Graduate Medical Education (ACGME)
duty hour restrictions on July 1, 2003, residency programs have been faced with the
challenge of mitigating resident fatigue and work hours while optimizing surgical
training and clinical experience.[1] This need was further emphasized in the ACGME 2020 Common Program requirements,
which required programs to apply strategies to encourage optimal resident well-being.[2] Multiple studies have shown that sleep deprivation experienced by residents is significantly
associated with not only burnout, but also reduced cognition and adverse patient outcomes.[3]
[4]
[5]
[6] In recent years, multiple residency programs have implemented night float systems
in an attempt to improve residents' experience on call and quality of life.[7]
[8]
[9] Per ACGME, residents assigned to night float are assigned on-site duty during night
shifts and do not have daytime assignments.[10] Although schedules vary among programs, night float system typically consists of
consecutive shifts of 10 to 12 hours over the course of 1 to 2 months.[11]
[12]
[13] A systemic review found that in several programs which had implemented night float
call systems, resident satisfaction, morale, and quality of life increased,[5]
[8]
[14]
[15] and one study in ophthalmology residents showed a similar subjective improvement
in burnout, fatigue, and workhours.[12] However, data for whether night float improves resident sleep quantity have been
mixed,[13]
[16]
[17]
[18]
[19]
[20] and it is unclear whether night float improves amount or quality of sleep, especially
when compared to home call.
Wearable activity trackers including the Fitbit Alta HR, employs photoplethysmogram
sensors which measure volumetric variations of blood circulation to track cardiovascular
system activity, in addition to three-axis acceleration sensors, to improve measurement
precision.[21]
[22] Recent studies have demonstrated that Fitbit devices provide reliable measurements
of total sleep time in healthy populations and have a high level of consistency between
devices.[23]
[24] The objective of this study was to use the Fitbit Alta HR device to measure sleep
patterns of ophthalmology residents on different call schedules, including home call,
night float rotations, and time away from call.
Methods
This study was approved by the University of Washington Institutional Review Board
and written consent was obtained from all participants. This was a crossover observational
study assessing patterns of seven postgraduate year (PGY)-2 ophthalmology residents
at the University of Washington from 2019 to 2021 using wrist actigraphy. Overnight
call was scheduled from 5:00 p.m. to 8:00 a.m. on weekdays, and 8:00 a.m. to 8:00
a.m. on weekends. The residency program implemented a partial night float rotation,
during which two to three nights of consecutive call (typically Sunday–Monday 8:00
a.m.–8:00 a.m., Monday 5:00 p.m.–8:00 a.m., and occasionally Tuesday 5:00 p.m.–8:00
a.m.) were assigned to a resident on a research rotation, without other clinical duties.
The remaining nights of call were assigned to residents as home call with postcall
relief, such that residents returned to regularly scheduled clinic after a call shift
but were relieved of clinical duties at noon following overnight call. Typically,
residents were scheduled for approximately 12 call shifts per month on night float
rotations and 6 call shifts per month on traditional home call rotations, with no
call taken on day consult rotations.
Sleep data of PGY-2 ophthalmology residents were collected using the Fitbit Alta HR
commercial device (Fitbit, San Francisco, CA). Residents were asked to wear their
Fitbits as much as tolerated throughout the entire PGY-2 year, with a minimum of at
least 2 weeks on each call schedule over 365 days. Fitbit sleep and activity were
matched to 8-week resident rotations on either night float, home call, or off call.
If residents had a stretch of 7 days or more not on call, typically during a consult
rotation when residents were not assigned regular call duties, this was also defined
as off call time. Postcall recovery sleep was defined as any stretch of sleep recorded
by the Fitbit, for which the majority of the sleep occurred between 8 a.m. and 8 p.m.
on the day the call shift ends. Raw data collected from the participants on the Fitbit
device were processed using the default proprietary algorithm developed by the Fitbit
developers, and the results were uploaded from the Fitbit application synced with
the device. The Fitbit provided data on “total minutes asleep” for every episode of
sleep experienced, with a minimum of 15 minutes of sleep recorded per episode. These
sleep minutes were tabulated for each call shift, as well as on off call dates during
each call schedule rotation. The data were organized into categories including average
amount of sleep over the entire rotation, as well as sleep minutes while on call and
during postcall recovery for both night float and home call schedules. Mixed model
regression was conducted in R software to compare sleep between the home call, night
float, and off call groups.
Results
Seven residents participated in the study, with an average age of 29 and a male:female
ratio of 5:2. Demographics and data collected from residents are listed in [Table 1]. Sleep data were recorded for a total of 1,015 nights, including 503 nights on home
call rotation, 230 nights on night float rotation, and 282 on a rotation without call
duties. All seven residents recorded sleep during home call rotations, while five
residents recorded sleep during night float rotations, and five residents recorded
sleep during periods away from call. Individual residents recorded (mean ± SD) 43 ± 20 nights of sleep on night float, 73 ± 48 nights of sleep on home call,
and 56 ± 45 nights of sleep during periods away from call.
Table 1
Demographics and data collection of night float, home call, and off call cohorts
|
Night float (n = 5)
|
Home call (n = 7)
|
No call (n = 5)
|
|
|
Demographics
|
|
|
Age, mean
|
29
|
29
|
29
|
|
|
M:F ratio
|
4:1
|
5:2
|
4:1
|
|
|
Data collection and adherence
|
|
|
Nights recorded per call schedule, all residents, n
|
230
|
503
|
282
|
|
|
On call nights recorded, all residents, n
|
73
|
161
|
N/A
|
|
|
Number of nights recorded per resident, mean ± SD (range)
|
43 ± 20 (17–61)
|
73 ± 48 (11–145)
|
56 ± 45 (12–107)
|
|
|
Percentage of nights recorded per resident, mean ± SD (range)
|
68 ± 35 (23–94)
|
39 ± 28 (3–83)
|
49 ± 40 (23–99)
|
|
|
Average recorded sleep in hours, mean ± SD
|
p
-Value
|
|
Sleep per night over the entire rotation
|
6.6 ± 1.9
|
6.2 ± 2.2
|
7.1 ± 1.2
|
0.008[a]
|
|
Total sleep during call nights including overnight sleep and postcall recovery
|
5.1 ± 2.8
|
4.2 ± 2.6
|
N/A
|
0.016
|
|
Overnight sleep during call nights
|
2.5 ± 2.2
|
2.6 ± 2.1
|
N/A
|
0.701
|
|
Postcall recovery sleep
|
2.5 ± 2.5
|
1.3 ± 1.6
|
N/A
|
<0.001
|
Abbreviations: M: F, male:female; SD, standard deviation; N/A, Not applicable
a
p-Value of home call versus night float.
Residents slept more when they were not assigned call compared to when they were on
call (both night float and home call; p < 0.001), with mean overall sleep 7.1 ± 1.2 hours on stretches without call and 6.3 ± 2.1 hours
when residents were on rotations with either home call or night float schedule.
Comparing home call with night float, residents slept more overall during rotations
on night float compared to home call (p = 0.008), with residents sleeping on average 6.6 ± 1.9 hours on night float rotations
compared to 6.2 ± 2.2 hours while on home call.
For comparison of call nights, only data from the five residents who had recorded
both home call and night float data were used. When considering sleep during call
nights including overnight sleep and postcall recovery sleep (from 5 p.m. through
5 p.m. the next day), residents on average slept more while on night float than on
home call (p = 0.016). On average, residents slept 4.2 ± 2.6 hours on each shift of home call,
and 5.1 ± 2.8 hours on each shift of night float, including both overnight and during
postcall recovery sleep.
Postcall recovery sleep was significantly different between night float and home call
schedules, as residents slept 2.5 ± 2.5 hours after night float call compared to 1.3 ± 1.6 hours
after home call (p < 0.001). When examining overnight sleep while on call only, overnight sleep while
on night float was 2.5 ± 2.2 hours whereas overnight sleep while on home call was
2.6 ± 2.1 hours, and this difference was not statistically significant (p = 0.701). Thus, postcall recovery sleep accounts for the increased resident sleep
during night float rotations compared to home call rotations.
Discussion
This was a study designed to provide quantitative data comparing sleep of ophthalmology
residents on different call schedules using the Fitbit Alta HR device. We found that
residents on a night float schedule had significantly higher average sleep over a
10-week rotation than on a home call schedule (p = 0.008). Average sleep on call (including overnight and postcall recovery sleep)
was also higher on night float shifts compared to home call shifts (p = 0.016).
Even though postcall relief was built into both night float and home call schedules
to allow for sleep recovery after call, our results demonstrated that ophthalmology
residents continued to sleep less while on either call schedule compared to periods
away from call. This suggests that the sleep disruption and deprivation introduced
by call is not fully remedied by providing time for postcall relief. However, our
finding that residents slept more on night float blocks suggests that implementing
a night float system could alleviate some of the sleep deprivation caused by call
duties. This could be due to the overall lower number of hours on duty when residents
were on night float compared to home call schedules, as residents are scheduled for
approximately 58 hours of duty on night float rotations, while they are scheduled
for an average of approximately 67 hours on home call rotations. While residents slept
the same amount overnight on both night float and traditional home call shifts, residents
slept more postcall during night float; the inconsistency of sleep schedules in the
traditional home call structure may have prevented residents from using postcall recovery
time to sleep. Additionally, postcall relief is provided only after 12 p.m. during
home call schedules whereas residents are relieved immediately after night float shifts.
This suggests that additional sleep during a home call shift may be possible if postcall
relief was offered immediately after home call rather than at 12 p.m.
Residents experienced less than 7 hours of sleep per night on average while on either
call schedule, consistent with sleep deprivation.[25] This is a finding that is consistent across residents of various specialties.[26] Sleeping less than 7 hours per night regularly is associated with adverse health
outcomes, including diabetes, hypertension, and heart disease.[25] We previously demonstrated that ophthalmology residents with decreased sleep experienced
increased feelings of burnout, depression, and anxiety,[20]
[27] and other studies have also shown an association between self-reported sleep deprivation
and an increased prevalence of burnout.[28]
[29] While the improvement of an average 24 minutes of sleep per night on night float
rotations compared to home call seems modest, the cumulative 2.8 hours of additional
sleep per week on night float may be impactful to reducing sleep deprivation. However,
a transition from home call to night float requires shifting daytime clinical training
hours to afterhours for a continuous period of time, so there are additional logistical
and training implications to consider for each individual program.
Further analysis is required to determine whether the quality, in addition to quantity,
of resident sleep between night float and home call schedules significantly varied.
Night shift work has been associated with circadian rhythm misalignment, which can
adversely affect sleep quality.[23]
[24] Past studies using subjective data have shown that residents reported experiencing
more sleep disturbances while on night float rotations,[18] though a recent study using the Fitbit device found no significant difference in
sleep efficiency between night float and on-call residents.[30] Similarly, wrist actigraphy has been shown to be capable of diagnosing circadian
rhythm sleep disorders, when at least 7 days of actigraphy are performed with a sleep
diary.[31]
[32] However, these diaries require a high level of participant burden to increase validity,
which would be difficult for busy residents to maintain.[31] In addition, commercial fitness trackers such as the Fitbit have shown mixed results
when used to evaluate specific sleep stages and pathologic sleep states.[33]
[34] Further studies analyzing sleep data from wrist actigraphy concurrently with sleep
diary completion in residents on varying call schedules would be useful in determining
differences in sleep quality and potential circadian rhythm misalignment.
Limitations of this study include low sample size and variable adherence with Fitbit
usage among individuals, and the study's nonrandomized design with data obtained from
a single ophthalmology program. Although all seven residents wore the Fitbit devices
for at least 2 weeks while on home call rotations, two residents had not worn the
device while on the night float rotation and two residents had not worn the device
during any time away from call and were thus excluded from those analyses.
Additionally, the study did not formally collect subjective feedback regarding call
schedules, or account for outside factors, including additional personal responsibilities
or obligations, that could influence sleep opportunities. However, the objective measurement
of total sleep time using the Fitbit Alta HR device for a large number of recorded
nights on and off call provides unique insight into differences in resident sleep
between two distinct call schedules. More studies are needed to investigate the effects
of night float implementation on resident sleep patterns, as well as sleep quality.
Conclusion
Our data using objective measurements demonstrate that implementation of a partial
night float system allowed PGY-2 ophthalmology residents to experience significantly
higher sleep over the span of a rotation compared to traditional home call. Further
studies with larger sample sizes incorporating subjective data among various residency
programs are required to investigate how night float rotations affect sleep quality,
in addition to sleep quantity.
