Between-subject and within-subject statistical information in dental research

Area of Science:

• Dental research
• Biostatistics
• Epidemiology

Background:

• Dental research often involves correlated data from multiple sites within the same patient.
• Standard statistical methods like generalized estimating equations and generalized linear mixed models are used for risk factor analysis.
• These methods often implicitly assume consistency between between-subject and within-subject comparisons.

Purpose of the Study:

• To highlight the potential for biased estimates and interpretation issues when between- and within-subject comparisons yield conflicting results.
• To illustrate the impact of differing statistical approaches on risk factor significance in dental studies.
• To emphasize the need for assessing the consistency of conclusions from different data sources in dental research.

Main Methods:

• Review of statistical assumptions in analyzing correlated dental data.
• Presentation of case examples from periodontal disease studies.
• Comparison of results from different statistical methods applied to the same datasets.

Main Results:

• Conflicting between- and within-subject comparisons can lead to biased statistical estimates.
• Different statistical methods can produce varying estimates and significance levels for the same risk factor.
• Examples from periodontal disease research demonstrate these discrepancies.

Conclusions:

• Statistical analyses in dental research must account for potential conflicts between between- and within-subject comparisons.
• It is crucial to verify if different sources of information lead to similar conclusions regarding risk factors and treatments.
• Ensuring consistency across analyses is vital for reliable dental research findings.