A Poisson-like model of sub-clinical signs from the examination of healthy aging subjects
- First Online:
- 21 Downloads
Background and aims: Our studies of the standard neurological examination on 66 middleaged (50–64 yrs) and elderly subjects (65–84 yrs) demonstrate that healthy elders have neurological deficits (or “signs”) that are not associated with specific known neurological disease. The purpose of the current study is to describe this loss of neurological function in healthy aging subjects as seen through accumulated subclinical neurological signs present. Methods: Logistic regression is applied to the data on each of six signs. Parameters determined are used to describe the distribution of first occurrence times for each sign. The results are then used to construct a Poisson-like model that describes the accumulation in the number of signs present over time on average. This model is also used to simulate a longitudinal population to explore the variability in the number of signs present over time in an aging population. Results: As the rate of arrival of the signs is heterogeneous, as determined through logistic regression, and the number of signs detected is finite, the resulting distributions of the number of signs over time have a different nature than Poisson. Our results suggest that we can expect to see on average one neurological deficit in healthy people by the age of 62, and that the expected number of deficits increases linearly at the rate of 1 additional sign every 12 years over a wide age range (age 70–90). The distribution of the number of deficits over time is also described. Conclusions: The linearity in the average rate at which signs appear in this population is somewhat of a surprise, in that an increasing (accelerating) rate might be anticipated. In addition to characterizing the neurological exam results in this group, we demonstrate a methodology that allows the comparison of groups and defines a rate of neurological aging.
Key wordsAging index geriatrics logistic regression neurological examintion Poisson model
Unable to display preview. Download preview PDF.