Mixed longitudinal studies: their efficiency for the estimation of trends over time

Area of Science:

• Biostatistics
• Epidemiology
• Longitudinal Data Analysis

Background:

• Mixed longitudinal studies provide valuable data for trend estimation, offering advantages over purely cross-sectional or longitudinal designs.
• Analyzing trends using mixed longitudinal data requires specific statistical approaches distinct from standard longitudinal analysis.

Purpose of the Study:

• To evaluate the efficiency of mixed longitudinal studies for measuring trends in anthropometric measurements (e.g., height, weight).
• To compare the efficiency of mixed longitudinal designs against traditional series of cross-sectional studies.
• To determine how design parameters, follow-up rates, and within-subject correlations impact efficiency.

Main Methods:

• Calculation of relative efficiency for various mixed longitudinal study designs.
• Assessment of efficiency based on differing follow-up percentages and correlation coefficients between repeated measurements.
• Analysis considering the number of surveys and the age range of participants.

Main Results:

• Mixed longitudinal studies demonstrated a relative efficiency generally exceeding 0.8 compared to cross-sectional studies.
• Efficiency sometimes surpassed 1.0, indicating superior performance over cross-sectional designs.
• Efficiency tended to decrease with an increased number of surveys but increase with a wider age group difference.

Conclusions:

• Mixed longitudinal studies are an efficient method for estimating trends in measurements over time.
• The design of mixed longitudinal studies, including participant follow-up and age stratification, significantly influences their efficiency.
• These findings support the use of mixed longitudinal designs for robust trend analysis in various scientific fields.