Issues of Validity in Actigraphic Sleep Assessment

Objectives The Standards of Practice Committee of the American Sleep Disorders Association has supported the use of actigraphy in the assessment of sleep disorders. Pollak et al disagree. The objective of this paper is to identify and critically evaluate several theoretic and methodological issues that are central to these divergent views regarding the valid use of actigraphy for sleep assessment.

Design Critical review, analysis, and comment.

Setting N/A

Patients/Participants N/A

Interventions N/A

Measurements and Results N/A

Conclusions (1) Coefficients of the validity of actigraphy exceed those associated with common medical tests and the best psychological tests. (2) Reasons why actigraphy should be held to a substantially higher empirical standard than common medical tests and the best psychological tests have yet to be advanced. (3) Differences between actigraphy and polysomnography are not random and can be reduced. (3a) Sleep onset is a gradual rather than a discrete process. Actigraphy keys on an earlier phase of the sleep-onset process than does polysomnography, resulting in systematic rather than random differences. (3b) A sleep switch device can be used to substantially increase the accuracy of sleep-onset times. (3c) The residual unreliability of polysomnographic data explains a portion of the differences between actigraphy and polysomnography. Actigraphy cannot be expected to agree more completely with polysomnography than polysomnography does with itself. (4) Complete agreement between actigraphy and polysomnography has been presumed, but achieving such a limit is theoretically impossible. Some lower maximum agreement limit should be identified. (5) Conclusions that actigraphy is not an accurate sleep-wake indicator and that it is inappropriate to infer sleep from actigraphy data are not supported by the findings. Conclusions reached, including caveats, by the Standards of Practice Committee remain supported.

Link to Abstract: http://www.ncbi.nlm.nih.gov/pubmed/14998254