Characterizing the time of day and year of falls in people with probable Parkinson's disease.

Temporal aspects of falls are not well characterized in people with Parkinson's disease (PD), particularly those later in life. The purpose of the current study is to characterize the timing (across day and year) of falls in people with probable PD compared to non-PD peers. Swedish registries were used to identify 441 people (mean age, 83.7 years, range = [59.0-100.0]) with one or more prospectively registered falls over the course of approximately 2.5 years. Of these participants, 40 were prescribed drugs of the class N04 "Anti-Parkinson Drugs". Of these 40, 30 were prescribed dopamine or dopamine derivatives, suggesting presence of PD, and 10 were prescribed "other PD-related medications". Given the unclear nature of these 10 participants, they were excluded from all analyses. Chi-squared tests and Generalized Estimating Equations were used to assess the time of day and year of falls in people with probable PD (n = 30) and non-PD (n = 401) accounting for covariates (age, gender, number of prescription medications). There was a statistically-significant effect of PD-group on fall time of day (accounting for gender, age, & number of prescription medications), such that Swedish residents with probable PD fell less often in the evening than the day (p = 0.015, OR = 0.61). Across the whole sample, female participants fell less frequently at night than the day (p = 0.032, OR = 0.73). People with probable PD fell more frequently in the spring months than other seasons (×2 = 32.1, p < 0.001). The finding that people with probable PD exhibit more falls during the day than night is consistent with previous work and extends knowledge to show a similar result in older, more frail people with PD. Characterizing the temporal nature of falls in people with PD can provide additional context to other fall circumstances, thus improving our ability to treat and predict falls in this group.