This does not include the use of actigraphy as a component of portable sleep monitoring (see Policy Guidelines section).
This policy does not address the use of actigraphy when used as a component of portable sleep monitoring under CPT codes 95800 or 95806 (see Related Policies). When used as a component of portable sleep monitoring, actigraphy should not be separately reported.
Actigraphy is the assessment of activity patterns based on recorded body movement that is interpreted by computer algorithms as sleep/wake cycles. These cycles may be altered in patients with sleep disorders including insomnia and circadian rhythm sleep disorders. In addition, actigraphy could potentially be used to assess sleep/wake disturbances associated with numerous other diseases or disorders. Actigraphy might also be used to measure the level of physical activity.
Actigraphy refers to the assessment of activity patterns by devices typically placed on the wrist or ankle that record body movement, which is interpreted by computer algorithms as periods of sleep (absence of activity) and wake (activity). Actigraphic devices are typically placed on the non-dominant wrist with a wristband and are worn continuously for at least 24 hours. Activity is usually recorded for a period of 3 days to 2 weeks but can be collected continuously over extended time periods with regular downloading of data onto a computer. The activity monitors may also be placed on the ankle for the assessment of restless legs syndrome, or on the trunk to record movement in infants.
The algorithms for detection of movement are variable among devices and may include “time above threshold,” the “zero crossing method” (the number of times per epoch that activity level crosses zero), or “digital integration” method, resulting in different sensitivities. Sensitivity settings (e.g., low, medium, high, automatic) can also be adjusted during data analysis. The digital integration method reflects both acceleration and amplitude of movement; this form of data analysis may be most commonly used today.
Data on patient bed times (lights out) and rise times (lights on) are usually entered into the computer record from daily patient sleep logs or by patient-activated event markers. Proprietary software is then used to calculate periods of sleep based on the absence of detectable movement, along with movement-related level of activity and periods of wake. In addition to providing graphic depiction of the activity pattern, device-specific software may analyze and report a variety of sleep parameters including sleep onset, sleep offset, sleep latency, total sleep duration, and wake after sleep onset. Actigraphy might also be used to measure the level of physical activity.
Actigraphy has been used for more than 2 decades as an outcome measure in sleep disorders research. For clinical applications, actigraphy is being evaluated as a measure of sleep/wake cycles in sleep disorders including insomnia and circadian rhythm sleep disorders. In addition, actigraphy is being investigated as a measure of sleep/wake disturbances associated with numerous other diseases and disorders.
Numerous actigraphy devices have received FDA clearance for marketing through the 510(k) process. Some actigraphy devices are designed and marketed to measure sleep/wake states while others are designed and marketed to measure levels of physical activity. FDA product code: OLV.
Medical policies are systematically developed guidelines that serve as a resource for Company staff when determining coverage for specific medical procedures, drugs or devices. Coverage for medical services is subject to the limits and conditions of the member benefit plan. Members and their providers should consult the member benefit booklet or contact a customer service representative to determine whether there are any benefit limitations applicable to this service or supply. This medical policy does not apply to Medicare Advantage.
Assessment of a diagnostic technology typically focuses on 3 categories of evidence:
This policy was initially based primarily on 2003 practice parameters issued by the American Academy of Sleep Medicine (AASM). (1) Because all the specific clinical indications for actigraphy were classified as guidelines or options, the AASM practice parameters indicated that all indications for actigraphy would be considered investigational. In a review paper that served as the basis for the 2003 practice parameters, (2) AASM pointed out the challenges in evaluating the diagnostic accuracy of actigraphy:
The 2005 update for the AASM Practice Parameters (3) continued to list actigraphy as an option and suggested areas such as restless legs syndrome and characterized circadian rhythm patterns for further evaluation. No controlled studies had been conducted to compare the results of actigraphy with other methods to determine if actigraphy would provide incremental information that would result in improved health outcomes.
In 2007, AASM published updated practice parameters on the use of actigraphy in the assessment of sleep and sleep disorders. (4) Whereas the 2005 practice parameters focused on the comparison of actigraphy with polysomnographically recorded sleep, the 2007 update included 108 additional studies comparing actigraphy with a number of standard clinical assessment tools that included sleep logs, subjective questionnaires, caregiver reports, and circadian phase markers. Actigraphy was recommended as a “standard” only as a method to estimate total sleep time in patients with obstructive sleep apnea syndrome when PSG is not available. Other indications changed from “option” to “guideline” but failed to reach a recommendation of “standard” due primarily to the absence of high-quality trials. Few of the studies reviewed had provided technical details related to the administration and scoring of actigraphy. In addition, most of the studies lacked a description of blinding, and there was “an inadequate description of whether visual inspection of data is performed, how missing data is handled, and other important decisions made in the analysis of actigraphy data.” The AASM Standards of Practice Committee indicated the need for additional research in the following areas:
In AASM’s 2007 Practice Parameter on evaluation and treatment of circadian rhythm sleep disorders (CRSDs), the use of actigraphy was considered as either an option or guideline, depending on the suspected disorder. (5) Specifically, use of actigraphy was recommended as an option for diagnosis of irregular sleep-wake disorder and free-running disorder and as a guideline for diagnosis of advanced sleep phase disorder, delayed sleep phase disorder, and shift work disorder. The evidence reviewed indicated good agreement between actigraphy and results of other diagnostic tools including polysomnography, sleep logs, and markers of circadian phase. It should be noted, however, that there is a relative lack of evidence for any procedure in the diagnosis or evaluation of treatment of CRSDs. For example, use of sleep logs received a guideline recommendation, based primarily on consensus and inclusion in the second edition of the International Classification of Sleep Disorders (ICSD-2). Insufficient evidence was found to recommend use of circadian phase markers for any CRSDs other than free-running disorder. Polysomnography is not routinely indicated for the diagnosis of CRSDs.(5)
Actigraphy is frequently used as an intermediate outcome in research studies. However, literature review updates have not identified any studies that evaluated whether the use of actigraphy would result in improved health outcomes for patients with sleep disorders. A number of studies have assessed sensitivity and specificity in either healthy or clinical populations (diagnostic accuracy). Following is a summary of key studies to date.
Actigraphy Compared With PSG
Paquet et al. (2007) compared actigraphic assessment of sleep and wake with PSG under varying conditions of sleep disturbance (night time sleep, daytime sleep, daytime sleep with caffeine) in 23 healthy subjects. (6) Data were analyzed from a study that evaluated the effects of caffeine on daytime recovery sleep. The experimental protocol involved 2 visits to the sleep laboratory, each including one night of nocturnal sleep, one night of sleep deprivation, and the next day of recovery sleep (once with placebo and once with 200 mg caffeine). The Actiwatch® and PSG equipment were synchronized before recording, and assessment of sleep and wake were compared for each one-minute interval to evaluate sensitivity, specificity, and accuracy of actigraphy in comparison with manually staged sleep from PSG recordings. Sensitivity was defined as the proportion of all epochs scored as sleep by PSG that were also scored as sleep by actigraphy. Specificity was the proportion of all epochs scored as wake by PSG that were also scored as wake by actigraphy. Accuracy was the proportion of all epochs correctly identified by actigraphy. Four different sensitivity settings/scoring algorithms were compared. In general, as the threshold to detect movement was raised, sensitivity to detect sleep increased, but the ability to detect wake (specificity) decreased. With the medium threshold algorithm, the sensitivity to detect sleep was 95% to 96%. However specificity, or the ability to detect wake, was 54% for night time sleep, 45% for daytime recovery sleep, and 37% for daytime recovery sleep with caffeine. A main finding of the study was that the more disturbed the sleep, the less the actigraph was able to differentiate between true sleep and quiet wakefulness, with an accuracy of 72% for the most disrupted sleep condition. Through experimental manipulation of the level of sleep disturbance, this study provides information about the limitations of this technology for clinical populations with sleep disruption.
Marino et al. (2013) assessed clinical validity of wrist actigraphy to measure nighttime sleep using the Cole-Kripke algorithm in 54 young and older adults, either healthy or with insomnia, and in 23 night-workers during daytime sleep. (7) Epoch by epoch comparison with PSG showed sensitivity (ability to detect sleep, 97%) and accuracy (86%) during the usual sleep/lights-out period to be high, but specificity (ability to detect wake, 33%) was low. As the amount of wake after sleep onset increased, the more that actigraphy underestimated this parameter. Several other studies assessed clinical validity in patients with primary or secondary sleep disorders. A 2006 study assessed the sensitivity and specificity of actigraphy in comparison with PSG in older adults treated for chronic primary insomnia. (8) Visual scoring of the PSG data was blinded, and actigraphic records were scored by proprietary software. The study found that actigraphy agreed with PSG scoring of sleep for 95% of the 30-second epochs (sensitivity), but agreed with PSG scoring of wake only 36% of the time (specificity). The study concluded that, “the clinical utility of actigraphy is still suboptimal in older adults treated for chronic primary insomnia.” Kaplan et al. (2012) compared outcomes from actigraphy, PSG, and sleep diary in 27 patients with bipolar disorder, who were between mood episodes, and in 27 age- and sex-matched controls. (9) Blinded evaluation found no significant difference in sleep parameters between patients with bipolar disorder and controls. Sleep parameter estimates from actigraphy and polysomnography were highly correlated. Taibi et al. (2013) found a sensitivity of 96.1% and specificity of 36.4% in a study of 16 older adults with insomnia who underwent 8 nights of concurrent actigraphy and PSG. (10) Sleep efficiency (actual sleep as a percentage of total recording time) was overestimated by actigraphy (84.4%) compared with PSG (66.9%) and the accuracy of actigraphy declined as sleep efficiency declined. Actigraphic and PSG measures of total sleep time were highly correlated, but correlations were marginal for sleep onset latency and wake after sleep onset. Sensitivity and specificity were not assessed. Louter et al. reported a study of actigraphy as a diagnostic aid for rapid eye movement (REM) sleep behavior disorder (RBD) in 45 consecutive patients with Parkinson disease. (11) The study population included patients referred for a variety of reasons, including insomnia, restless legs syndrome, and sleep apnea. Following video PSG, 23 patients were diagnosed with RBD. There was no significant difference between the 2 groups for the presence of other sleep disorders. Using a cutoff of 95 wake bouts per night, actigraphy had a sensitivity of 26.1% and specificity of 95.5%, with a positive predictive value of 85.7%.
Beecroft et al. (2008) reported an observational study of sleep monitoring in the intensive care unit, comparing nurse assessment, actigraphy, and PSG in 12 stable, critically ill, mechanically ventilated patients. (12) PSG showed severely disrupted sleep, with decreased total sleep time and sleep efficiency, high frequency of arousals and awakenings (fragmentation), and abnormal sleep architecture (decreased slow wave and REM sleep). Both the nurse’s and the actigraphic assessment of sleep were found to be inaccurate. Actigraphy overestimated the total sleep time, with a median that was 2 to 3 hours greater than PSG. Median sleep efficiency was estimated at 61% to 95% by actigraphy, depending on the sensitivity setting, which was substantially higher than the 42% median sleep efficiency shown by PSG with sleep staging. Similar findings were reported by van der Kooi et al (2013) in a study of 7 short-term intensive-care unit patients; median specificity was less than 19% when compared with PSG-recorded sleep. (13) In a 2010 study, Actigraphy with a SOMNOwatch™ in patients (n=28) with sleep -disordered breathing showed a sensitivity of 90%, a specificity of 95%, and overall accuracy of 86% in comparison with PSG. (14) Correlations were high for total sleep time (0.89), sleep period time (0.91), and sleep latency (0.89), and moderate for sleep efficiency (0.71) and sustained sleep efficiency (0.65).
Studies continue to assess different modes of data collection and analysis, including varying the sensitivity settings for existing algorithms and developing new scoring algorithms. A 2011 publication compared 3 collection modes (proportional integration, time above threshold, zero crossings) with PSG in 889 older community-dwelling men who participated in the Outcomes of Sleep Disorders in Men (MrOS) study. (15) The proportional integration mode was found to correspond best to PSG, with moderate interclass correlation coefficients of 0.32 to 0.57. Actigraphy in this mode overestimated total sleep time by an average of 13.2 minutes, with an absolute difference (positive or negative direction) of 52.9 minutes. There was a systematic bias for overestimating total sleep time, which increased with decreasing sleep duration.
A systematic review of leg actigraphy to quantify periodic limb movements of sleep (PLMS) found significant heterogeneity for the sensitivity and specificity of different devices.16 Factors contributing to the heterogeneity were variability in devices tested, placement of the devices (e.g., foot or ankle), thresholds to define clinically significant PLMS (e.g., 5, 10, or 15/hour), and algorithms used to calculate the periodic limb movements. The inability to combine actigraphy data from both legs also presents a limitation for clinical use at this time.
Actigraphy Compared With Sleep Diaries
Levenson et al. (2013) evaluated the utility of sleep diaries and actigraphy to differentiate older adults with insomnia (n=79) from good sleeper controls (n=40). (17) Sensitivity and specificity were determined for sleep onset latency, wake after sleep onset, sleep efficiency, and total sleep time; patients with insomnia completed PSG studies, but controls did not. Using receiver operating characteristic curve analysis, sleep diary measurements produced areas under the curve in the high range (0.84-0.97), whereas actigraphy performed less well at discriminating between older adults with insomnia and controls (areas under the curve, 0.58-0.61).
Children and Adolescents
Actigraphy Compared With PSG
In 2011, O’Driscoll et al reported a comparison of actigraphy with PSG in 130 children who had been referred for assessment of sleep-disordered breathing. (18) The arousal index and apnea-hypopnea index (AHI) scored from PSG were compared with the number of wake bouts/hour and actigraphic fragmentation index. Using a PSG-determined AHI of greater than 1 event/hour, the actigraphic measure of wake bouts/hour had a sensitivity and specificity of 14.9% and 98.8%, respectively, and the fragmentation index had a sensitivity and specificity of 12.8% and 97.6%, respectively. Using a PSG-determined arousal index greater than 10 events per hour as the reference standard, the actigraphic measure of wake bouts/hour had a sensitivity and a specificity of 78.1% and 52.6% and the fragmentation index had a sensitivity and specificity of 82.2% and 50.9%, both respectively. Based on receiver operating characteristic curves, the ability of actigraphic measures to correctly classify a child as having an AHI of greater than 1 event/hour was considered poor.
A 2007 study examined the validity of actigraphy for determining sleep and wake in children with sleep disordered breathing with data analyzed over 4 separate activity threshold settings (low, medium, high, auto). (19) The low and auto activity thresholds were found to adequately determine sleep (relative to PSG) but significantly underestimated wake, with sensitivity of 97% and specificity of 39%. The medium- and high-activity thresholds significantly underestimated sleep time (sensitivity, 94% and 90%, respectively) but were not found to be significantly different from the total PSG estimates of wake time (specificity, 59% and 69%). Overall agreement rates between actigraphy and PSG (for both sleep and wake) were 85% to 89%. Belanger et al. (2013) assessed Sensitivity and specificity of different scoring algorithms in healthy preschoolers. (20) An algorithm designed specifically for children showed the highest accuracy (95.6%) in epoch-by-epoch comparison with PSG.
Insana et al. (2010) compared ankle actigraphic recording and PSG in 22 healthy infants (13-15 months of age). (21) Actigraphy was found to underestimate total sleep time by 72 minutes and overestimate wake after sleep onset by 14 minutes. In 55% of the infants, total sleep time was underestimated by 60 minutes or more. Sensitivity was calculated for total sleep time (92%), stages 1 and 2 combined (91%), slow wave sleep (96%), and REM sleep (89%). Specificity for identifying wake was 59%, and accuracy was 90%. Overall, actigraphy identified sleep relatively well but was unable to discriminate wake from sleep. A 2011 study compared wrist actigraphy with PSG in 149 healthy school-aged children. (22) Although the sleep period time was not significantly different, actigraphy was found to underestimate total sleep time by 32 minutes (correlation coefficient, 0.47) and overestimate wake after sleep onset by 26 minutes (correlation coefficient, 0.09). The authors concluded that actigraphy is relatively inaccurate for the determination of sleep quality in this population.
Actigraphy Compared With Sleep Diaries
Werner et al. (2008) assessed agreement between actigraphy and parent diary or questionnaire for sleep patterns in 50 children, aged 4 to 7 years, recruited from kindergarten schools in Switzerland. Sixty-eight (10%) of 660 invited families agreed to participate. (23) Each child was home-monitored with an actigraph for 6 to 8 consecutive nights, and parents were requested to complete a detailed sleep diary (15-minute intervals) during the monitoring days to indicate bedtime, estimated sleep start, wake periods during the night, and estimated sleep end. Parents’ assessment of habitual wake time, get up time, bedtime, time of lights off, sleep latency, and nap duration were obtained through questionnaire. Satisfactory agreement, defined a priori as differences smaller than 30 minutes, was achieved between actigraphy and diary for sleep start, sleep end, and assumed sleep. Actual sleep time and nocturnal wake time differed by an average of 72 minutes and 55 minutes, respectively. Satisfactory agreement was not reached between actigraphy and questionnaire for any of the parameters. The study concluded that the diary is a cost-effective and valid source of information about children’s sleep-schedule time, while actigraphy may provide additional information about nocturnal wake time or may be used if parents are unable to report in detail. Compliance and accuracy in the diaries is likely to be affected by the motivation of the parents, who were self-selected in this study.
Discrepancy between actigraphic and sleep diary measures of sleep in adolescents was reported by Short et al in 2012. (24) A total of 290 adolescents (13-18 years) completed 8 days of sleep diaries and actigraphy. Actigraphic estimates of total sleep time (median, 6 hours 57 minutes) were significantly less than total sleep time recorded in adolescent’s sleep diaries (median, 8 hours 17 minutes) or parent reports (median, 8 hours 51 minutes). Wake after sleep onset averaged 7 minutes in sleep diaries and 74 minutes by actigraphy. Actigraphy estimated wake after sleep onset of up to 3 hours per night in the absence of any wakening from sleep diaries, suggesting an overestimation of wake in this population. The discrepancy between actigraphy and sleep diary estimates of sleep was greater for boys than for girls, consistent with PSG studies showing increased nocturnal motor behavior in boys.
Actigraphy Compared With Behavioral Observations
A validation study of actigraphy for determining sleep and wake was conducted in 10 preterm infants using videotaped behavioral observations. (25) The study was conducted for a 24-hour period each week while the infants were in the nursery, resulting in a total of 38 studies. Wakefulness was scored as quiet wake with eyes open and “bright,” active wake with eyes open and gross body movements, or crying. Sleep included quiet sleep with regular breathing and eyes closed, active sleep with irregular breathing and REMs, and indeterminate sleep, during which characteristics of both active and quiet sleep were observed. Behavioral sleep-wake scoring was carried out blinded to the knowledge of the actigraphy data. The actigraph, which was synchronized to the video recording, was placed in a custom-designed sleeve bandage and positioned on the infant’s leg midway between the knee and ankle. The agreement rate between actigraphic determination of sleep and wake, and behavioral scoring ranged from 66% for the high sensitivity setting at the youngest gestational age (30-33 weeks) to 89% at the low sensitivity setting for infants of 37 to 40 weeks of gestational age. For the youngest infants, sensitivity and specificity at the low threshold were 88% and 34%, respectively. For infants of 37 to 40 weeks of gestational age, sensitivity and specificity were 97% and 32%, respectively. Similar results (97% sensitivity and 24% specificity) were obtained in a 2008 study with an epoch-by-epoch comparison of actigraphy and videosomnography in 22 autistic, 11 developmentally delayed, and 25 normally developing preschool children.(26)
Actigraphy Compared With Video-Electroencephalography
A prospective validation study of actigraphy for determining sleep-wake patterns in children with epilepsy was reported in 2014. (27) In this study, 27 children with medically refractory epilepsy wore activity monitors while being evaluated with at least 24-hour video-electroencephalography (vEEG; mean, 70.5 hours) in an inpatient epilepsy monitoring unit. vEEG and actigraphy data were evaluated by 2 independent and blinded reviewers. Although sensitivity and specificity were not reported, correlation coefficients between the 2 measures were very high (r range, 0.93-0.99) for night sleep period, night sleep time, duration of night wake time, and percent time of sleep during the day. Consistent with lower specificity to detect awakenings during sleep, the correlation for the number of awakenings after sleep onset was less robust.
Summary of Evidence
The clinical validity of actigraphy, the assessment of activity patterns by devices typically placed on the wrist or ankle that record body movement, depends, to a large extent, on the modality with which it is being compared.
Overall, progress has been made since the 2007 American Academy of Sleep Medicine (AASM) research recommendations in assessing the reliability and validity of different algorithms in comparison with the reference standard. Although actigraphy appears to provide reliable measures of sleep onset and wake time in some patient populations, the clinical utility of actigraphy over the less expensive sleep diary has not been demonstrated. Moreover, evidence indicates that actigraphy does not provide a reliable measure of sleep efficiency in clinical populations. Evidence to date does not indicate that this technology is as beneficial as the established alternatives. Therefore, actigraphy is considered investigational.
Practice Guidelines and Position Statements
American Academy of Sleep Medicine
The recommendations of AASM are categorized as standards, guidelines, or options. Standards describe a generally accepted patient care strategy, which reflects a high degree of clinical certainty. Guidelines reflect a moderate degree of clinical certainty, while options imply either inconclusive or conflicting evidence or conflicting expert opinion. As noted here, there is only one recommendation considered a standard, and this addresses the technical performance of actigraphic devices (first bullet next paragraph). There is also only 1 recommended guideline (second bullet next paragraph), and this addresses the small subset of patients with insomnia and restless legs syndrome with specific indications. All of the other recommendations are considered options.
Recommendations of AASM from 2003 (1):
A 2005 Update for the AASM practice parameters (3) continued to list actigraphy as an option and also suggested areas, such as restless legs syndrome and characterizing circadian rhythm patterns, for further evaluation.
Updated practice parameters in 2007 on the use of actigraphy in the assessment of sleep and sleep disorders (including a separate practice parameter on circadian rhythm sleep disorders) recommended actigraphy as a “standard” only as a method to estimate total sleep time in patients with obstructive sleep apnea syndrome when PSG is not available. (4, 5) Other indications changed from option to guideline but failed to reach a recommendation of standard due primarily to the absence of high-quality trials.
AASM practice parameters from 2008 on clinical management of chronic insomnia in adults reference the 2007 practice parameters on actigraphy, stating that actigraphy is indicated as a method (Option) to characterize circadian rhythm patterns or sleep disturbances in individuals with insomnia, including insomnia associated with depression.(28)
U.S. Preventive Services Task Force Recommendations
Medicare National Coverage
There is no national coverage determination (NCD). In the absence of an NCD, coverage decisions are left to the discretion of local Medicare carriers.
Actigraphy testing, recording, analysis, interpretation, and report (minimum of 72 hours to 14 consecutive days of recording)
Add policy to Medicine Section - New Policy
Update Scope and Disclaimer - No other changes.
Replace Policy - Policy reviewed with literature search; references added; no change in policy statement.
Replace Policy - Policy reviewed with literature search; references added; no change in policy statement.
Code Update - Code 95803 added, effective 1/1/09.
Replace Policy - Policy reviewed with literature search; references added; no change in policy statement.
Replace Policy - Policy updated with literature review through December 2010; references added and reordered; policy statement unchanged.
Replace policy. Policy updated with literature review through November 2011; references 10 and 12 added and references reordered; some references removed. Policy statement unchanged.
Updated Related Policy title, 2.01.503.
Replace policy. Rationale section updated based on a literature review through January 4, 2013. References 8 and 15 added; others renumbered or removed. Policy statement unchanged.
Update Related Policies. Change title to policy 2.01.503.
Annual Review. Policy updated with literature review through January 6, 2014; references 7, 10, 12, 15, 18, and 25 added; policy statement unchanged. ICD-9 and ICD-10 diagnosis and procedure codes removed; they are not utilized in adjudication.
Update Related Policies. Change title to 2.01.503.
Annual Review. Policy updated with literature review through January 6, 2015. References 11, 16, and 27 added; others renumbered. Policy statement unchanged.
Disclaimer: This medical policy is a guide in evaluating the medical necessity of a particular service or treatment. The Company adopts policies after careful review of published peer-reviewed scientific literature, national guidelines and local standards of practice. Since medical technology is constantly changing, the Company reserves the right to review and update policies as appropriate. Member contracts differ in their benefits. Always consult the member benefit booklet or contact a member service representative to determine coverage for a specific medical service or supply. CPT codes, descriptions and materials are copyrighted by the American Medical Association (AMA).