Apps and fitness trackers that measure sleep: Are they useful?

More and more consumers are using wearable devices and smartphones to monitor and measure various body functions, including sleep. Many patients now present their providers with sleep data obtained from their phones and other devices. But can these devices provide valid, useful clinical information?

This article describes common sleep tracking devices available to consumers and the mechanisms the devices probably use to distinguish sleep from wakefulness (their algorithms are secret), the studies evaluating the validity of device manufacturers’ claims, and their clinical utility and limitations.

DEVICES ARE COMMON

Close to 1 in 10 adults over age 18 owns an activity tracker, and sales are projected to reach $50 billion by 2018.¹ Even more impressive, close to 69% of Americans own a smartphone,¹ and more than half use it as an alarm clock.²

At the same time that these devices have become so popular, sleep medicine has come of age, and experts have been pushing to improve people’s sleep and increase awareness of sleep disorders.^3,4 While the technology has significantly advanced, adoption of data from these devices for clinical evaluation has been limited. Studies examining the validity of these devices have only recently been conducted, and companies that make the devices have not been forthcoming with details of the specific algorithms they use to tell if the patient is asleep or awake or what stage of sleep the patient is in.

WHAT ARE THESE DEVICES?

Consumer tracking devices that claim to measure sleep are easily available for purchase and include wearable fitness trackers such as Fitbit, Jawbone UP, and Nike+ Fuelband. Other sleep tracking devices are catalogued by Ko et al.⁵ Various smartphone applications (apps) are also available.

Fitness trackers, usually worn as a wrist band, are primarily designed to measure movement and activity, but manufacturers now claim the trackers can also measure sleep. Collected data are available for the user to review the following day. In most cases, these trackers display sleep and wake times; others also claim to record sound sleep, light sleep, and the number and duration of awakenings. Most fitness trackers have complementary apps available for download that graphically display the data on smartphones and interact with social media to allow users to post their sleep and activity data.

More than 500 sleep-related apps are available for download to smartphones in the iTunes app store⁵; the Sleep Cycle alarm clock app was among the top 5 sleep-tracking apps downloaded in 2014.⁶ Because sleep data collection relies on the smartphone being placed on the user’s mattress, movements of bed partners, pets, and bedding may interfere with results. In most cases, the apps display data in a format similar to that of fitness trackers. Some claim to determine the optimal sleep phase for the alarm to wake the user.

HOW DOES THE TECHNOLOGY WORK?

Older activity-tracking devices used single-channel electroencephalographic recordings or multiple physiologic channels such as galvanic skin response, skin temperature, and heat flux to measure activity to determine transitions between periods of sleep and wakefulness.^7,8

None of the currently available consumer sleep tracking devices discloses the exact mechanisms used to measure sleep and wakefulness, but most appear to rely on 3-axis accelerometers,⁹ ie, microelectromechanical devices that measure front-to-back, side-to-side, and up-and-down motion and convert the data into an activity count. Activity counts are acquired over 30- or 60-second intervals and are entered into algorithms that determine if the pattern indicates that the patient is awake or asleep. This is the same method that actigraphy uses to evaluate sleep, but most actigraphs used in medicine disclose their mechanisms and provide clinicians with the option of using various validated algorithms to classify the activity counts into sleep or awake periods.^9–11

 

ARE THE MEASURES VALID?

Only a few studies have examined the validity and accuracy of current fitness trackers and apps for measuring sleep.

The available studies are difficult to compare; most have been small and used different actigraphy devices for comparison. Some tested healthy volunteers, others included people with suspected or confirmed sleep disorders, and some had both types of participants. In many studies, the device was compared with polysomnography for only 1 night, making the “first-night effect” likely to be a confounding factor, as people tend to sleep worse during the first night of testing. Technical failures for the devices were noted in some studies.¹² In addition, some currently used apps may use different platforms than the devices used in these studies, limiting the extrapolation of results.

Two fitness tracking devices (Fitbit^13,14  and Jawbone UP^15–17) were compared with polysomnography and actigraphy in several studies in children and adults (Table 1). The devices tended to overestimate sleeping time, sleep efficiency, and latency to sleep onset and underestimate awake time after falling asleep. Some studies noted that differences were more pronounced for those with the most disturbed sleep.

As with fitness trackers, few studies have been done to examine the validity of smartphone apps.⁵ Findings of 3 studies are summarized in Table 2.^17–19 In addition to tracking the duration and depth of sleep, some apps purport to detect snoring, sleep apnea, and periodic limb movements of sleep. Discussion of these apps is beyond the scope of this review.

In general, sleep tracking devices are fair to good at detecting sleep but poor at determining wakefulness. They are inaccurate for determining absolute sleep parameters (ie, total sleep time, sleep efficiency, wake time after sleep onset, and sleep onset latency) and in distinguishing the different sleep stages compared to polysomnography. Age-related differences have been found between consumer sleep devices compared with polysomnography and actigraphy-derived measures; because adults are likelier to lie still when awake, activity monitors are prone to overestimate sleep time in adults. Comparisons in patients with sleep apnea are conflicting.^12,15 Claims that the “sensitive” mode may be appropriate for users with sleep disorders are thus far unsubstantiated.

ARE THE DEVICES CLINICALLY USEFUL?

Although a thorough history remains the cornerstone of a good evaluation of sleep problems, testing is sometimes essential, and in certain situations, objective data can complement the history and clarify the diagnosis.

Polysomnography remains the gold standard for telling when the patient is asleep vs awake, diagnosing sleep-disordered breathing, detecting periodic limb movements and parasomnias, and aiding in the diagnosis of narcolepsy.

Actigraphy, which uses technology similar to fitness trackers, can help distinguish sleep from wakefulness, reveal erratic sleep schedules, and help diagnose circadian rhythm sleep disorders. In patients with insomnia, actigraphy can help determine daily sleep patterns and response to treatment.²⁰ It can be especially useful for patients who cannot provide a clear history, eg, children and those with developmental disabilities or cognitive dysfunction.

Consumer sleep tracking devices, like actigraphy, are portable and unobtrusive, providing a way to measure sleep duration and demonstrate sleep patterns in a patient’s natural environment. Being more accessible, cheaper, and less time-consuming than clinical tests, the commercially available devices could be clinically useful in some situations, eg, for monitoring overall sleep patterns to look for circadian sleep-wake disorders, commonly seen in shift workers (shift work disorder) or adolescents (delayed sleep-wake phase disorder); or in patients with poor motivation to maintain a sleep diary. Because of their poor performance in clinical trials, they should not be relied upon to distinguish sleep from wakefulness, quantify the amount of sleep, determine sleep stages, and awaken patients exclusively from light sleep.

Discerning poor sleep hygiene from insomnia

Patients with insomnia tend to take longer to go to sleep (have longer sleep latencies), wake up more (have more disturbed sleep with increased awakenings), and have shorter sleep times with reduced sleep efficiencies.²¹ Sleep tracking devices tend to be less accurate for patients with short sleep duration and disturbed sleep, limiting their usefulness in this group. Furthermore, patients with insomnia tend to underestimate their sleep time and overestimate sleep latency; some devices also tend to overestimate the time to fall asleep, reinforcing this common error made by patients.^22,23

On the other hand, data from sleep tracking devices could help distinguish poor sleep hygiene from an insomnia disorder. For example, the data may indicate that a patient has poor sleep habits, such as taking long daytime naps or having significantly variable time in and out of bed from day to day. The total times asleep and awake in the middle of the night may also be substantially different on each night, which would also possibly indicate poor sleep hygiene.

 

Detecting circadian rhythms

A device may show that a patient has a clear circadian preference that is not in line with his or her daily routines, suggesting an underlying circadian rhythm sleep-wake disorder. This may be evident by bedtimes and wake times that are consistent but substantially out of sync with one’s social or occupational needs.

Measuring overall sleep duration

In people with normal sleep, fitness trackers perform reasonably well for measuring overall sleep duration. This information could be used to assess a patient with daytime sleepiness and fatigue to evaluate insufficient sleep as an etiologic factor.

Table 3 summarizes how to evaluate the data from sleep apps and fitness tracking devices for clinical use. While these features of consumer sleep tracking devices could conceivably help in the above clinical scenarios, further validation of devices in clinical populations is necessary before their use can be recommended without reservation.

ADVISING PATIENTS

Patients sometimes present to clinicians with concerns about the duration of sleep time and time spent in various sleep stages as delineated by their sleep tracking devices. Currently, these devices do not appear to be able to adequately distinguish various sleep stages, and in many users, they can substantially underestimate or overestimate sleep parameters such as time taken to fall asleep or duration of awakenings in the middle of the night. Patients can be reassured about this lack of evidence and should be advised to not place too much weight on such data alone.

Sleep “goals” set by many devices have not been scientifically validated. People without sleep problems should be discouraged from making substantial changes to their routines to accommodate sleep targets set by the devices. Patients should be counseled about the pitfalls of the data and can be reassured that little evidence suggests that time spent in various sleep stages correlates with adverse daytime consequences or with poor health outcomes.

Some of the apps used as alarm clocks claim to be able to tell what stage of sleep people are in and wait to awaken them until they are in a light stage, which is less jarring than being awakened from a deep stage, but the evidence for this is unclear. In the one study that tested this claim, the app did not awaken participants from light sleep more often than is likely to occur by chance.¹⁷ The utility of these apps as personalized alarm clocks is still extremely limited, and patients should be counseled to obtain an adequate amount of sleep rather than rely on devices to awaken them during specific sleep stages.

The rates for discontinuing the use of these devices are high, which could limit their utility. Some surveys have shown that close to 50% of users stop using fitness trackers; 33% stop using them within 6 months of obtaining the device.²⁴ Also, there is little evidence that close monitoring of sleep results in behavior changes or improved sleep duration. Conversely, the potential harms of excessive monitoring of one’s sleep are currently unknown.

More and more consumers are using wearable devices and smartphones to monitor and measure various body functions, including sleep. Many patients now present their providers with sleep data obtained from their phones and other devices. But can these devices provide valid, useful clinical information?

This article describes common sleep tracking devices available to consumers and the mechanisms the devices probably use to distinguish sleep from wakefulness (their algorithms are secret), the studies evaluating the validity of device manufacturers’ claims, and their clinical utility and limitations.

DEVICES ARE COMMON

Close to 1 in 10 adults over age 18 owns an activity tracker, and sales are projected to reach $50 billion by 2018.¹ Even more impressive, close to 69% of Americans own a smartphone,¹ and more than half use it as an alarm clock.²

At the same time that these devices have become so popular, sleep medicine has come of age, and experts have been pushing to improve people’s sleep and increase awareness of sleep disorders.^3,4 While the technology has significantly advanced, adoption of data from these devices for clinical evaluation has been limited. Studies examining the validity of these devices have only recently been conducted, and companies that make the devices have not been forthcoming with details of the specific algorithms they use to tell if the patient is asleep or awake or what stage of sleep the patient is in.

WHAT ARE THESE DEVICES?

Consumer tracking devices that claim to measure sleep are easily available for purchase and include wearable fitness trackers such as Fitbit, Jawbone UP, and Nike+ Fuelband. Other sleep tracking devices are catalogued by Ko et al.⁵ Various smartphone applications (apps) are also available.

Fitness trackers, usually worn as a wrist band, are primarily designed to measure movement and activity, but manufacturers now claim the trackers can also measure sleep. Collected data are available for the user to review the following day. In most cases, these trackers display sleep and wake times; others also claim to record sound sleep, light sleep, and the number and duration of awakenings. Most fitness trackers have complementary apps available for download that graphically display the data on smartphones and interact with social media to allow users to post their sleep and activity data.

More than 500 sleep-related apps are available for download to smartphones in the iTunes app store⁵; the Sleep Cycle alarm clock app was among the top 5 sleep-tracking apps downloaded in 2014.⁶ Because sleep data collection relies on the smartphone being placed on the user’s mattress, movements of bed partners, pets, and bedding may interfere with results. In most cases, the apps display data in a format similar to that of fitness trackers. Some claim to determine the optimal sleep phase for the alarm to wake the user.

HOW DOES THE TECHNOLOGY WORK?

Older activity-tracking devices used single-channel electroencephalographic recordings or multiple physiologic channels such as galvanic skin response, skin temperature, and heat flux to measure activity to determine transitions between periods of sleep and wakefulness.^7,8

None of the currently available consumer sleep tracking devices discloses the exact mechanisms used to measure sleep and wakefulness, but most appear to rely on 3-axis accelerometers,⁹ ie, microelectromechanical devices that measure front-to-back, side-to-side, and up-and-down motion and convert the data into an activity count. Activity counts are acquired over 30- or 60-second intervals and are entered into algorithms that determine if the pattern indicates that the patient is awake or asleep. This is the same method that actigraphy uses to evaluate sleep, but most actigraphs used in medicine disclose their mechanisms and provide clinicians with the option of using various validated algorithms to classify the activity counts into sleep or awake periods.^9–11

 

ARE THE MEASURES VALID?

Only a few studies have examined the validity and accuracy of current fitness trackers and apps for measuring sleep.

The available studies are difficult to compare; most have been small and used different actigraphy devices for comparison. Some tested healthy volunteers, others included people with suspected or confirmed sleep disorders, and some had both types of participants. In many studies, the device was compared with polysomnography for only 1 night, making the “first-night effect” likely to be a confounding factor, as people tend to sleep worse during the first night of testing. Technical failures for the devices were noted in some studies.¹² In addition, some currently used apps may use different platforms than the devices used in these studies, limiting the extrapolation of results.

Two fitness tracking devices (Fitbit^13,14  and Jawbone UP^15–17) were compared with polysomnography and actigraphy in several studies in children and adults (Table 1). The devices tended to overestimate sleeping time, sleep efficiency, and latency to sleep onset and underestimate awake time after falling asleep. Some studies noted that differences were more pronounced for those with the most disturbed sleep.

As with fitness trackers, few studies have been done to examine the validity of smartphone apps.⁵ Findings of 3 studies are summarized in Table 2.^17–19 In addition to tracking the duration and depth of sleep, some apps purport to detect snoring, sleep apnea, and periodic limb movements of sleep. Discussion of these apps is beyond the scope of this review.

In general, sleep tracking devices are fair to good at detecting sleep but poor at determining wakefulness. They are inaccurate for determining absolute sleep parameters (ie, total sleep time, sleep efficiency, wake time after sleep onset, and sleep onset latency) and in distinguishing the different sleep stages compared to polysomnography. Age-related differences have been found between consumer sleep devices compared with polysomnography and actigraphy-derived measures; because adults are likelier to lie still when awake, activity monitors are prone to overestimate sleep time in adults. Comparisons in patients with sleep apnea are conflicting.^12,15 Claims that the “sensitive” mode may be appropriate for users with sleep disorders are thus far unsubstantiated.

ARE THE DEVICES CLINICALLY USEFUL?

Although a thorough history remains the cornerstone of a good evaluation of sleep problems, testing is sometimes essential, and in certain situations, objective data can complement the history and clarify the diagnosis.

Polysomnography remains the gold standard for telling when the patient is asleep vs awake, diagnosing sleep-disordered breathing, detecting periodic limb movements and parasomnias, and aiding in the diagnosis of narcolepsy.

Actigraphy, which uses technology similar to fitness trackers, can help distinguish sleep from wakefulness, reveal erratic sleep schedules, and help diagnose circadian rhythm sleep disorders. In patients with insomnia, actigraphy can help determine daily sleep patterns and response to treatment.²⁰ It can be especially useful for patients who cannot provide a clear history, eg, children and those with developmental disabilities or cognitive dysfunction.

Consumer sleep tracking devices, like actigraphy, are portable and unobtrusive, providing a way to measure sleep duration and demonstrate sleep patterns in a patient’s natural environment. Being more accessible, cheaper, and less time-consuming than clinical tests, the commercially available devices could be clinically useful in some situations, eg, for monitoring overall sleep patterns to look for circadian sleep-wake disorders, commonly seen in shift workers (shift work disorder) or adolescents (delayed sleep-wake phase disorder); or in patients with poor motivation to maintain a sleep diary. Because of their poor performance in clinical trials, they should not be relied upon to distinguish sleep from wakefulness, quantify the amount of sleep, determine sleep stages, and awaken patients exclusively from light sleep.

Discerning poor sleep hygiene from insomnia

Patients with insomnia tend to take longer to go to sleep (have longer sleep latencies), wake up more (have more disturbed sleep with increased awakenings), and have shorter sleep times with reduced sleep efficiencies.²¹ Sleep tracking devices tend to be less accurate for patients with short sleep duration and disturbed sleep, limiting their usefulness in this group. Furthermore, patients with insomnia tend to underestimate their sleep time and overestimate sleep latency; some devices also tend to overestimate the time to fall asleep, reinforcing this common error made by patients.^22,23

On the other hand, data from sleep tracking devices could help distinguish poor sleep hygiene from an insomnia disorder. For example, the data may indicate that a patient has poor sleep habits, such as taking long daytime naps or having significantly variable time in and out of bed from day to day. The total times asleep and awake in the middle of the night may also be substantially different on each night, which would also possibly indicate poor sleep hygiene.

 

Detecting circadian rhythms

A device may show that a patient has a clear circadian preference that is not in line with his or her daily routines, suggesting an underlying circadian rhythm sleep-wake disorder. This may be evident by bedtimes and wake times that are consistent but substantially out of sync with one’s social or occupational needs.

Measuring overall sleep duration

In people with normal sleep, fitness trackers perform reasonably well for measuring overall sleep duration. This information could be used to assess a patient with daytime sleepiness and fatigue to evaluate insufficient sleep as an etiologic factor.

Table 3 summarizes how to evaluate the data from sleep apps and fitness tracking devices for clinical use. While these features of consumer sleep tracking devices could conceivably help in the above clinical scenarios, further validation of devices in clinical populations is necessary before their use can be recommended without reservation.

ADVISING PATIENTS

Patients sometimes present to clinicians with concerns about the duration of sleep time and time spent in various sleep stages as delineated by their sleep tracking devices. Currently, these devices do not appear to be able to adequately distinguish various sleep stages, and in many users, they can substantially underestimate or overestimate sleep parameters such as time taken to fall asleep or duration of awakenings in the middle of the night. Patients can be reassured about this lack of evidence and should be advised to not place too much weight on such data alone.

Sleep “goals” set by many devices have not been scientifically validated. People without sleep problems should be discouraged from making substantial changes to their routines to accommodate sleep targets set by the devices. Patients should be counseled about the pitfalls of the data and can be reassured that little evidence suggests that time spent in various sleep stages correlates with adverse daytime consequences or with poor health outcomes.

Some of the apps used as alarm clocks claim to be able to tell what stage of sleep people are in and wait to awaken them until they are in a light stage, which is less jarring than being awakened from a deep stage, but the evidence for this is unclear. In the one study that tested this claim, the app did not awaken participants from light sleep more often than is likely to occur by chance.¹⁷ The utility of these apps as personalized alarm clocks is still extremely limited, and patients should be counseled to obtain an adequate amount of sleep rather than rely on devices to awaken them during specific sleep stages.

The rates for discontinuing the use of these devices are high, which could limit their utility. Some surveys have shown that close to 50% of users stop using fitness trackers; 33% stop using them within 6 months of obtaining the device.²⁴ Also, there is little evidence that close monitoring of sleep results in behavior changes or improved sleep duration. Conversely, the potential harms of excessive monitoring of one’s sleep are currently unknown.

More and more consumers are using wearable devices and smartphones to monitor and measure various body functions, including sleep. Many patients now present their providers with sleep data obtained from their phones and other devices. But can these devices provide valid, useful clinical information?

This article describes common sleep tracking devices available to consumers and the mechanisms the devices probably use to distinguish sleep from wakefulness (their algorithms are secret), the studies evaluating the validity of device manufacturers’ claims, and their clinical utility and limitations.

DEVICES ARE COMMON

Close to 1 in 10 adults over age 18 owns an activity tracker, and sales are projected to reach $50 billion by 2018.¹ Even more impressive, close to 69% of Americans own a smartphone,¹ and more than half use it as an alarm clock.²

At the same time that these devices have become so popular, sleep medicine has come of age, and experts have been pushing to improve people’s sleep and increase awareness of sleep disorders.^3,4 While the technology has significantly advanced, adoption of data from these devices for clinical evaluation has been limited. Studies examining the validity of these devices have only recently been conducted, and companies that make the devices have not been forthcoming with details of the specific algorithms they use to tell if the patient is asleep or awake or what stage of sleep the patient is in.

WHAT ARE THESE DEVICES?

Consumer tracking devices that claim to measure sleep are easily available for purchase and include wearable fitness trackers such as Fitbit, Jawbone UP, and Nike+ Fuelband. Other sleep tracking devices are catalogued by Ko et al.⁵ Various smartphone applications (apps) are also available.

Fitness trackers, usually worn as a wrist band, are primarily designed to measure movement and activity, but manufacturers now claim the trackers can also measure sleep. Collected data are available for the user to review the following day. In most cases, these trackers display sleep and wake times; others also claim to record sound sleep, light sleep, and the number and duration of awakenings. Most fitness trackers have complementary apps available for download that graphically display the data on smartphones and interact with social media to allow users to post their sleep and activity data.

More than 500 sleep-related apps are available for download to smartphones in the iTunes app store⁵; the Sleep Cycle alarm clock app was among the top 5 sleep-tracking apps downloaded in 2014.⁶ Because sleep data collection relies on the smartphone being placed on the user’s mattress, movements of bed partners, pets, and bedding may interfere with results. In most cases, the apps display data in a format similar to that of fitness trackers. Some claim to determine the optimal sleep phase for the alarm to wake the user.

HOW DOES THE TECHNOLOGY WORK?

Older activity-tracking devices used single-channel electroencephalographic recordings or multiple physiologic channels such as galvanic skin response, skin temperature, and heat flux to measure activity to determine transitions between periods of sleep and wakefulness.^7,8

None of the currently available consumer sleep tracking devices discloses the exact mechanisms used to measure sleep and wakefulness, but most appear to rely on 3-axis accelerometers,⁹ ie, microelectromechanical devices that measure front-to-back, side-to-side, and up-and-down motion and convert the data into an activity count. Activity counts are acquired over 30- or 60-second intervals and are entered into algorithms that determine if the pattern indicates that the patient is awake or asleep. This is the same method that actigraphy uses to evaluate sleep, but most actigraphs used in medicine disclose their mechanisms and provide clinicians with the option of using various validated algorithms to classify the activity counts into sleep or awake periods.^9–11

 

ARE THE MEASURES VALID?

Only a few studies have examined the validity and accuracy of current fitness trackers and apps for measuring sleep.

The available studies are difficult to compare; most have been small and used different actigraphy devices for comparison. Some tested healthy volunteers, others included people with suspected or confirmed sleep disorders, and some had both types of participants. In many studies, the device was compared with polysomnography for only 1 night, making the “first-night effect” likely to be a confounding factor, as people tend to sleep worse during the first night of testing. Technical failures for the devices were noted in some studies.¹² In addition, some currently used apps may use different platforms than the devices used in these studies, limiting the extrapolation of results.

Two fitness tracking devices (Fitbit^13,14  and Jawbone UP^15–17) were compared with polysomnography and actigraphy in several studies in children and adults (Table 1). The devices tended to overestimate sleeping time, sleep efficiency, and latency to sleep onset and underestimate awake time after falling asleep. Some studies noted that differences were more pronounced for those with the most disturbed sleep.

As with fitness trackers, few studies have been done to examine the validity of smartphone apps.⁵ Findings of 3 studies are summarized in Table 2.^17–19 In addition to tracking the duration and depth of sleep, some apps purport to detect snoring, sleep apnea, and periodic limb movements of sleep. Discussion of these apps is beyond the scope of this review.

In general, sleep tracking devices are fair to good at detecting sleep but poor at determining wakefulness. They are inaccurate for determining absolute sleep parameters (ie, total sleep time, sleep efficiency, wake time after sleep onset, and sleep onset latency) and in distinguishing the different sleep stages compared to polysomnography. Age-related differences have been found between consumer sleep devices compared with polysomnography and actigraphy-derived measures; because adults are likelier to lie still when awake, activity monitors are prone to overestimate sleep time in adults. Comparisons in patients with sleep apnea are conflicting.^12,15 Claims that the “sensitive” mode may be appropriate for users with sleep disorders are thus far unsubstantiated.

ARE THE DEVICES CLINICALLY USEFUL?

Although a thorough history remains the cornerstone of a good evaluation of sleep problems, testing is sometimes essential, and in certain situations, objective data can complement the history and clarify the diagnosis.

Polysomnography remains the gold standard for telling when the patient is asleep vs awake, diagnosing sleep-disordered breathing, detecting periodic limb movements and parasomnias, and aiding in the diagnosis of narcolepsy.

Actigraphy, which uses technology similar to fitness trackers, can help distinguish sleep from wakefulness, reveal erratic sleep schedules, and help diagnose circadian rhythm sleep disorders. In patients with insomnia, actigraphy can help determine daily sleep patterns and response to treatment.²⁰ It can be especially useful for patients who cannot provide a clear history, eg, children and those with developmental disabilities or cognitive dysfunction.

Consumer sleep tracking devices, like actigraphy, are portable and unobtrusive, providing a way to measure sleep duration and demonstrate sleep patterns in a patient’s natural environment. Being more accessible, cheaper, and less time-consuming than clinical tests, the commercially available devices could be clinically useful in some situations, eg, for monitoring overall sleep patterns to look for circadian sleep-wake disorders, commonly seen in shift workers (shift work disorder) or adolescents (delayed sleep-wake phase disorder); or in patients with poor motivation to maintain a sleep diary. Because of their poor performance in clinical trials, they should not be relied upon to distinguish sleep from wakefulness, quantify the amount of sleep, determine sleep stages, and awaken patients exclusively from light sleep.

Discerning poor sleep hygiene from insomnia

Patients with insomnia tend to take longer to go to sleep (have longer sleep latencies), wake up more (have more disturbed sleep with increased awakenings), and have shorter sleep times with reduced sleep efficiencies.²¹ Sleep tracking devices tend to be less accurate for patients with short sleep duration and disturbed sleep, limiting their usefulness in this group. Furthermore, patients with insomnia tend to underestimate their sleep time and overestimate sleep latency; some devices also tend to overestimate the time to fall asleep, reinforcing this common error made by patients.^22,23

On the other hand, data from sleep tracking devices could help distinguish poor sleep hygiene from an insomnia disorder. For example, the data may indicate that a patient has poor sleep habits, such as taking long daytime naps or having significantly variable time in and out of bed from day to day. The total times asleep and awake in the middle of the night may also be substantially different on each night, which would also possibly indicate poor sleep hygiene.

 

Detecting circadian rhythms

A device may show that a patient has a clear circadian preference that is not in line with his or her daily routines, suggesting an underlying circadian rhythm sleep-wake disorder. This may be evident by bedtimes and wake times that are consistent but substantially out of sync with one’s social or occupational needs.

Measuring overall sleep duration

In people with normal sleep, fitness trackers perform reasonably well for measuring overall sleep duration. This information could be used to assess a patient with daytime sleepiness and fatigue to evaluate insufficient sleep as an etiologic factor.

Table 3 summarizes how to evaluate the data from sleep apps and fitness tracking devices for clinical use. While these features of consumer sleep tracking devices could conceivably help in the above clinical scenarios, further validation of devices in clinical populations is necessary before their use can be recommended without reservation.

ADVISING PATIENTS

Patients sometimes present to clinicians with concerns about the duration of sleep time and time spent in various sleep stages as delineated by their sleep tracking devices. Currently, these devices do not appear to be able to adequately distinguish various sleep stages, and in many users, they can substantially underestimate or overestimate sleep parameters such as time taken to fall asleep or duration of awakenings in the middle of the night. Patients can be reassured about this lack of evidence and should be advised to not place too much weight on such data alone.

Sleep “goals” set by many devices have not been scientifically validated. People without sleep problems should be discouraged from making substantial changes to their routines to accommodate sleep targets set by the devices. Patients should be counseled about the pitfalls of the data and can be reassured that little evidence suggests that time spent in various sleep stages correlates with adverse daytime consequences or with poor health outcomes.

Some of the apps used as alarm clocks claim to be able to tell what stage of sleep people are in and wait to awaken them until they are in a light stage, which is less jarring than being awakened from a deep stage, but the evidence for this is unclear. In the one study that tested this claim, the app did not awaken participants from light sleep more often than is likely to occur by chance.¹⁷ The utility of these apps as personalized alarm clocks is still extremely limited, and patients should be counseled to obtain an adequate amount of sleep rather than rely on devices to awaken them during specific sleep stages.

The rates for discontinuing the use of these devices are high, which could limit their utility. Some surveys have shown that close to 50% of users stop using fitness trackers; 33% stop using them within 6 months of obtaining the device.²⁴ Also, there is little evidence that close monitoring of sleep results in behavior changes or improved sleep duration. Conversely, the potential harms of excessive monitoring of one’s sleep are currently unknown.