Comparing correlations of continuous observations from two independent populations using a sequential approach

Area of Science:

• Biostatistics
• Genetic Epidemiology
• Quantitative Genetics

Background:

• Sequential study designs offer greater efficiency than fixed sample designs, particularly in genetic and epidemiological research.
• Existing sequential methods primarily focus on dichotomous traits in matched case-control studies.
• There is a need for sequential analysis methods applicable to continuous paired observations, especially in twin studies.

Purpose of the Study:

• To extend sequential analysis to compare associations within two independent groups of paired continuous observations.
• To develop a sequential twin method based on intraclass correlation for analyzing phenotypic correlations.
• To demonstrate the efficiency gains and flexibility of the proposed sequential approach in familial studies.

Main Methods:

• Development of a sequential statistical method for paired continuous data, utilizing intraclass correlation.
• Application of the method to a heritability study of dysplasia, incorporating body mass index as a covariate.
• Comparison of monozygotic twins with pairs of singleton sisters using the sequential approach.

Main Results:

• The sequential twin method significantly reduces the number of required observations compared to fixed sample designs.
• The method maintains study power and error rates while increasing efficiency.
• The approach allows for the straightforward inclusion of explanatory factors, such as body mass index.

Conclusions:

• Sequential analysis is a valuable tool for enhancing the efficiency of twin and familial studies involving continuous traits.
• The developed sequential twin method provides a robust and flexible framework for analyzing phenotypic correlations.
• This methodology can lead to substantial savings in resources (time, cost) without compromising statistical rigor.