Keywords:
Polysomnography - Actigraphy - Validity of Tests - Technology
INTRODUCTION
The quantification and measurement of sleep amongst various interventional and population
research studies and clinical settings is of increasing importance. Sleep monitoring
has also become a substantial consumer industry, with a rising rate of commercial
companies producing various wearable sleep monitoring devices[1]. Although considered the 'gold standard' method of sleep measurement, polysomnography
(PSG) requires a somewhat intrusive and expensive assessment of sleep indices. Wrist
actigraphy is a non-intrusive, cost-effective tool used to estimate sleep quantity
and quality which has been compared to PSG, showing an accuracy of up to 93% in healthy
adults for total sleep time and sleep efficiency[2]
,
[3] and as such is widely used in the sleep literature[4].
Actigraphy involves the use of a device housed in a wristwatch that contains a small
accelerometer capable of sensing movement along any one of three axes[4]. The accelerometer is sampled multiple times per second and the actigraph is downloaded
and either manually or automatically scored for sleep indices. While actigraphy has
become commonplace in both the research and consumer setting, the optimal placement
of the actigraph itself is relatively unknown.
Traditionally, the majority of research studies recommend that the actigraphy device
should be worn on the non-dominant wrist[4], however some studies have suggested that it may be more suitable to wear the actigraph
on the dominant wrist[5], and others do not specify what wrist it should be worn on[6]. Furthermore, the new emerging technology over the past decade has seen improved
automatic scoring actigraphy devices, reducing the need for sleep technicians.
Given the increasing use of actigraphy for monitoring sleep, the new and emerging
technology for automatic scoring of devices and the disparate recommendations in the
literature, this is an important area that needs further clarification. Therefore,
the aim of the current case study was to determine if differences exist between wearing
automatic-scoring actigraphy devices on the non-dominant and dominant wrists in healthy
adults.
METHODS
Participants
A total of 13 healthy adults (8 male / 5 female, mean ± SD, age: 26±7) volunteered
to take part in the study. All participants were free of any diagnosed sleep disorders.
Ethical approval for the study was obtained through the institutions Human Research
Ethics Committee.
Study Design
Participants were required to wear a wrist actigraphy device (Readiband, Fatigue Science,
Vancouver) on each wrist (dominant and non-dominant) over a period of 5 nights. Participants
were instructed to maintain their usual sleep habits and general daily activity patterns
during the monitoring period, and were instructed to leave the devices on at all times.
The Readiband has been validated against PSG, with accuracy levels of 93% being reported[7] and research from our laboratory has also shown that the Readiband results in acceptable
levels of inter-device reliability (ICC = >0.90)[8].
At the conclusion of each recording period, actigraphy data were wirelessly downloaded
to a computer using a Nordic 2.4 GHz ANT transceiver, which was then analysed using
Fatigue Science software (16Hz sampling rate: ReadibandT, Fatigue Science, Vancouver).
The raw activity scores were translated to sleep-wake scores based on computerized
scoring algorithms. Sleep indices including total sleep time (TST), sleep efficiency
(SE%), total time in bed (TTB), sleep latency (SL), wake after sleep onset (WASO),
sleep onset time (SOT) and wake time (WT) were used for comparison between devices.
Statistical Analysis
Simple group statistics are shown as means ± standard deviations unless stated otherwise.
An independent-samples T-test was used to compare dominant and non-dominant wrist
measures using the Statistical Package for Social Science (V. 22.0, SPSS Inc., Chicago,
IL), with statistical significance set at p<0.05. Inter-device agreements for dominant and non-dominant wrists were examined
using Pearson's correlation coefficients (r) with 95% confidence intervals (95% CI) and interpreted as 0.90-1.00 = very high correlation, 0.70-0.89 = high correlation, 0.50-0.69 = moderate correlation, 0.26-0.49 = low correlation and 0.00-0.25 = little, if any correlation[9]. Between-device typical error of estimates (TEE) was determined using an excel spreadsheet[10] and are presented as a coefficient of variation percentage (CV%) and as absolute
values. Similar to Werner et al.[11], we defined an apriori difference between the 2 devices of = 30 min satisfactory
for TST, with a difference < 5% for SE satisfactory.
RESULTS
There were no significant differences between devices for any of the measured sleep
variables (p>0.05).
Mean differences of 6 mins and 2 min between non-dominant and dominant wrists for
TST and TTB were associated with CV% scores of 4.6 and 3.8%, respectively ([Table 1]).
Table 1. Mean +- SD values for the measured sleep variables between non-dominant and dominant
wrist-actigraphy devices. Comparison between devices are reported using mean bias,
Pearson correlations (r), typical error of estimates (TEE) and coefficient of variation
% with 95% confidence intervals.
TST, TTB, SOT and WT all resulted in very high correlations (>0.90), with SE%, SL and WASO resulting in high correlations between devices (0.89, 0.89 and 0.76, respectively).
DISCUSSION
The main finding in the current study was that there were no significant differences
between non-dominant and dominant wrist actigraphy for monitoring sleep in healthy
adults. The non-significant differences between devices were associated with high to very high correlations for all sleep measures and relatively low (~5%) CV's for the key sleep
variables of total sleep time, total time in bed and sleep efficiency. The typical
error of estimate for total sleep time and total time in bed was ~20 minutes and the
mean bias was ~5 minutes, suggesting that there is little difference in what wrist
the actigraph is worn on.
Figure 1 Correlation plots between non-dominant and dominant wrist actigraphy data for: a)
total sleep time (mins), and b) sleep efficiency (%).
The use of wrist-actigraphy for monitoring sleep is becoming increasingly popular
in numerous research fields, including athletic[12], clinical[13], adolescent[14] and pediatric[15] populations. However, while traditionally it was suggested that actigraphy devices
should be worn on the dominant wrist of participants[16], there is a lack of evidence to show whether or not any differences actually exist.
Furthermore, the ever-evolving technology of sleep monitoring via actigraphy has introduced
automatic-scoring devices[8], further identifying the need to investigate differences in the placement of devices.
As previously suggested[6]
,
[17], care should be taken when interpreting results for WASO and SL when measured via
wrist actigraphy, as the accuracy of these measures when compared to PSG is questionable.
The current study would support this, as these were the most variable sleep indices
between the two devices, with typical error of estimates and CV% of ~25 minutes and
~60% for WASO and ~15 minutes and ~100% for SL, respectively.
Results from the current study would suggest that the placement of the actigraphy
device (dominant vs. non-dominant wrist) is not critical for accurate sleep measurement
of key sleep measures including total sleep time, total time in bed, sleep onset and
wake time. Given this, the authors would recommend that individuals wearing actigraphy
devices, either as general consumers or research participants, should opt to wear
their actigraphy device on whatever wrist feels the most comfortable. Indeed, if the
device feels uncomfortable, it is more likely to influence adherence to wearing the
monitor and overall sleep results.
