Random change point models: investigating cognitive decline in the presence of missing data

Abstract

With the aim of identifying the age of onset of change in the rate of cognitive decline while accounting for the missing observations, we considered a selection modelling framework. A random change point model was fitted to data from a population-based longitudinal study of ageing (the Cambridge City over 75 Cohort Study) to model the longitudinal process. A missing at random mechanism was modelled using logistic regression. Random effects such as initial cognitive status, rate of decline before and after the change point, and the age of onset of change in rate of decline were estimated after adjustment for risk factors for cognitive decline. Among other possible predictors, the last observed cognitive score was used to adjust the probability of death and dropout. Individuals who experienced less variability in their cognitive scores experienced a change in their rate of decline at older ages than individuals whose cognitive scores varied more.

Breaking the sticks: a hierarchical change‐point model for estimating ontogenetic shifts with stable isotope data

“…2000; Beckage et al. 2007; Ghosh & Vaida 2007; Muniz‐Terrera, den Hout & Matthews 2011), and how to fit them using the BUGS language (Lunn et al. 2000).…”

“…Such models have been used in epidemiology to infer the onset of cognitive decline (Hall et al. 2000; Muniz‐Terrera, den Hout & Matthews 2011), of prostate cancer (Bellera et al. 2008) or of HIV immunologic response decline (Ghosh & Vaida 2007).…”

“…This paper thus deals with the problem of estimating individual ontogenic shifts with longitudinal isotopic data, that is repeated measurements of d 13 C and d 15 N on the same organism throughout its life. We present a Bayesian change-point model to estimate jointly individual ontogenic shifts in d 13 C and d 15 N. Our aim is to bring forward to a larger audience the vast literature on change-point models (Hall et al 2000;Beckage et al 2007;Ghosh & Vaida 2007;Muniz-Terrera, den Hout & Matthews 2011), and how to fit them using the BUGS language (Lunn et al 2000).…”

Derivative estimation for longitudinal data analysis: Examining features of blood pressure measured repeatedly during pregnancy

“…The use of spline mixed models can lead to an incorrect population mean estimate and incorrect pointwise standard errors, due to the effect of covariance structure of the individual effects on the population mean estimate. Heckman pointed out that this could be remedied by ensuring that the function space for the individual effects is a subspace of the function space for the population mean curve. We have used a simple approach to apply this condition which is to select the knots for the individual curves ( as a subset of the knots of the population curves ( κ k ).…”

Can we spot deleterious ageing in two waves of data? The Lothian Birth Cohort 1936 from ages 70 to 73

“…Even samples screened for clinical manifestations of impairment would inevitably include some who had entered this period but remained undiagnosed, and their numbers could be expected to increase with age, which could create the declines in mean function with age that studies consistently show. As longitudinal data samples have proliferated and statistical analytical techniques have improved, research efforts have been directed toward describing individual trajectories of decline (e.g., Finkel et al, 2005;McGue & Christensen, 2002) and/or the specific interval before death at which some period of 'terminal decline' begins (e.g., Rabbitt et al, 2011;Sliwinski et al, 2006;Terrera et al, 2011;Wilson et al, 2003). These studies have produced widely varying results about the extent of change, its rate of acceleration if considered, and the length of any terminal decline interval.…”