Actigraphy reliability with normal sleepers

Actigraphy has gained popularity as an objective method for measuring sleep in a home setting. We evaluated whether missing data affects the utility of actigraphy for the measurement of sleep parameters in normal sleepers.

We evaluated actigraphy data from 60 normal sleepers who participated in a study of the effects of changes in setting on sleep. Participants were asked to wear a Mini Mitter Actiwatch actigraph for 35 days and to use event markers to record bedtime and arising time. Counts of nights on which participants failed to supply usable data were computed using the following criteria: missing nights, missing bedtime markers, missing arising time markers, and multiple markers supplied at bedtime or arising time. A night on which any of these problems occurred was counted as unscorable.

We evaluated a total of 2,100 nights, of which 559 (27%) nights were deemed unscorable due to missing data. Missing markers at bedtime (206) and arising time (172) accounted for the majority of missing data. Trends over the five weeks indicated that incidence of missing data increased over the assessment period.

We found that missing data was a significant problem for long-term assessment of sleep using actigraphy. We suggest that researchers consider compensatory strategies, such as extending the assessment period and using adjunctive measures, in order to obtain sufficient data for analysis. We also recommend that future improvements in actigraphy instruments should aim to address the sources of missing data.