Curve fitting for repeated measurements made at irregular time-points

Area of Science:

• Biostatistics
• Epidemiology
• Human Biology

Background:

• Longitudinal studies often face challenges with non-simultaneous measurements across subjects.
• Classical curve fitting methods are unsuitable for irregularly timed repeated measures.
• Existing methods may not adequately address data collected at varied time points.

Purpose of the Study:

• To develop a statistical method for analyzing longitudinal data with non-simultaneous measurements.
• To adapt curve fitting techniques for irregularly spaced repeated measures.
• To provide a robust method for analyzing human population studies with staggered data collection.

Main Methods:

• An estimation procedure using iteratively reweighted least squares was developed.
• An intraclass correlation matrix was assumed for measurements within subjects.
• The model was generalized to incorporate covariables with minimal procedural modification.

Main Results:

• The proposed method effectively handles longitudinal data with unevenly spaced time points.
• The estimation procedure demonstrated robustness in analyzing repeated measurements.
• The generalized model successfully incorporated covariables.

Conclusions:

• The developed statistical method offers a viable solution for analyzing longitudinal studies with asynchronous data collection.
• This approach is applicable to various fields, including human population studies and biomedical research.
• The method provides a flexible framework for analyzing bone mass changes in postmenopausal women.