Night-time activity forecast by season and weather in a longitudinal design - natural light effects on three years' rest-activity cycles in nursing home residents with dementia.

Abstract

Motor activity of 20 elderly nursing home residents living with dementia was analyzed using recordings of continuously worn wrist activity monitors over a three-year period. The average recording duration was 479 ± 206 days per participant (mean ± SD). Regional cloud amount and day length data from the local weather station (latitude: 52°56'N) were included in the analysis to investigate their effects on several activity variables. Nocturnal rest, here defined as the five consecutive hours with the least motor activity during 24 hours (L5), was the most predictable activity variable per participant. There was a significant interaction of night-time activity with day length and cloud amount (F 1,1174 = 4.39; p = 0.036). Night-time activity was higher on cloudy short days than on clear short days (p = 0.007), and it was also higher on cloudy short days than on cloudy long days (p = 0.032). The need for sufficient zeitgeber (time cue) strength during winter time, especially when days are short and skies are cloudy, is crucial for elderly people living with dementia. Activity forecast by season and weather might be a valuable approach to anticipate adequately complementary use of electrical light and thereby foster lower night-time activity.