The Chatham Blood Pressure Study. An Application of Bayesian Growth Curve Models to a Longitudinal Study of Blood Pressure in Children

Abstract

Recent developments in statistics have produced powerful methods that facilitate the analysis of longitudinal studies. These methods are illustrated by an analysis of a longitudinal study of blood pressure in children. The results of the study show a clear tendency for blood pressure to increase with age, and Asian children tend to have lower blood pressures than their Caucasian counterparts of the same age. There is evidence to support the hypothesis that blood pressures track.

Mathematical and Analytical Aspects of Tracking

“…In many studies, the concept of tracking is related to modeling the longitudinal development of a certain variable Y(t) (12,44,45). All of these studies have to do with growth in general and, although they are very interesting, they do not estimate some sort of predictive value.…”

Longitudinal analysis of incomplete adolescent data

“…Alternative approaches which have been applied to the analysis of longitudinal blood pressure data include that of Lauer, Mahoney and Clarke (1986), which applies the linear growth curve model to individuals' serial percentile values rather than observed measurements, and Darby and Fearn's (1979) Bayesian analysis of preadolescent blood pressure which also assumes a linear growth curve model and permits unbalanced and incomplete data. These approaches are similar to the random-effects model we used to model the adult blood pressures.…”

Longitudinal principal components and non-linear regression models of early childhood growth

“…The authors claim it is reasonable to assume uncorrelated errors since our observed ages are widely separated in time. Many authors (Day 1966, El Lozy 1978, Preece and Baines 1978, Darby and Fearn 1979, make this assumption, although methods which allow correlated errors are becoming more accessible (Glasbey 1979, Sandland and McGilchrist 1979, Bock and Thissen 1980. The reader should refer to Berkey (1982) for empirical evidence supporting our assumptions.…”