Approximately 10 years ago, Fitbit (Fitbit, Inc) introduced its first wearable model for use by health-conscious consumers. Nonetheless, these devices, which rely entirely on movement-based algorithms, lack sleep-stage assessment capability additionally, they tend to overestimate sleep duration. Wrist actigraphy, which senses accelerated motion, was introduced some 35 years ago by Ambulatory Monitoring Inc and is now used in conjunction with proprietary interpretative algorithms to conduct outpatient sleep screenings through estimation of key sleep parameters. EEG wearables enable at-home evaluation of sleep architecture and staging but they are expensive and somewhat technologically complicated.
Sleep diary methods are simple and economical ways of tracking and appraising sleep by consumers but because they entail subjective self-ratings, they are often inaccurate and incomplete furthermore, they do not assess sleep architecture and stages. Thus, it is not surprising less than 50% of sleep studies nowadays are conducted in formal sleep facilities.
Additionally, PSG requires a special facility plus oversight by skilled technicians, making it expensive and precluding, under most circumstances, investigation of between-night variation of sleep quality. However, the environment and instrumentation of conventional PSG can be uncomfortable, anxiety producing, and even sleep disturbing. PSG is regarded as the gold standard for diagnosis of sleep disorders and conduct of sleep research. Polysomnography (PSG) consists of simultaneous electroencephalographic (EEG), electromyographic, electrooculographic, electrocardiographic, and other assessments.
Sleep-staging Fitbit models showed higher sensitivity (0.95-0.96) and specificity (0.58-0.69) values in detecting sleep epochs than nonsleep-staging models and those reported in the literature for regular wrist actigraphy. Sleep-staging Fitbit models, in comparison to PSG, showed no significant difference in measured values of WASO ( P=.25 heterogenicity: I 2=0%, P=.92), TST ( P=.29 heterogenicity: I 2=0%, P=.98), and SE ( P=.19) but they underestimated SOL ( P=.03 heterogenicity: I 2=0%, P=.66). Recent-generation Fitbit models that collectively utilize heart rate variability and body movement to assess sleep stages performed better than early-generation nonsleep-staging ones that utilize only body movement. In reference to PSG, nonsleep-staging Fitbit models correctly identified sleep epochs with accuracy values between 0.81 and 0.91, sensitivity values between 0.87 and 0.99, and specificity values between 0.10 and 0.52. In reference to polysomnography (PSG), nonsleep-staging Fitbit models tended to overestimate total sleep time (TST range from approximately 7 to 67 mins effect size=-0.51, P<.001 heterogenicity: I 2=8.8%, P=.36) and sleep efficiency (SE range from approximately 2% to 15% effect size=-0.74, P<.001 heterogenicity: I 2=24.0%, P=.25), and underestimate wake after sleep onset (WASO range from approximately 6 to 44 mins effect size=0.60, P<.001 heterogenicity: I 2=0%, P=.92) and there was no significant difference in sleep onset latency (SOL P=.37 heterogenicity: I 2=0%, P=.92). After eliminating duplicates and in compliance with inclusion and exclusion criteria, 22 articles qualified for systematic review, with 8 providing quantitative data for meta-analysis. The search yielded 3085 candidate articles.