Premium
A quality control measure for longitudinal studies with continuous outcomes
Author(s) -
Ambrosius Walter T.,
Hui Siu L.
Publication year - 2000
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/(sici)1097-0258(20000530)19:10<1339::aid-sim480>3.0.co;2-#
Subject(s) - measure (data warehouse) , quality (philosophy) , computer science , control (management) , statistics , econometrics , mathematics , data mining , artificial intelligence , philosophy , epistemology
At the time of data acquisition in a longitudinal study a decision needs to be made whether or not the latest measurement of the primary outcome is a potential outlier. If the data point does not fit with the subject's prior data, the patient can be immediately remeasured before he/she leaves the office. From the third visit onwards, a least squares approach can be used to generate prediction intervals for the value of the response at that visit. We propose a Bayesian method for calculating a prediction interval that can incorporate external information about the process that can be used beginning at the first visit. Both the least squares and Bayesian approaches will be used to prospectively clean longitudinal data. An example using longitudinally measured bone density measurements in the elderly will be discussed. In addition, simulation studies will be described which show that both cleaning methods are better than doing nothing and that the Bayesian approach outperforms the least squares method. Copyright © 2000 John Wiley & Sons, Ltd.