Statistical testing against baseline in orthodontic research: a meta-epidemiologic study.

Area of Science:

• Orthodontics
• Biostatistics
• Scientific Publishing

Background:

• Within-group comparisons from baseline to follow-up are frequently used in orthodontic research.
• This practice can lead to flawed interpretations of study findings.
• The prevalence and predictors of this statistical issue require investigation.

Purpose of the Study:

• To assess the prevalence of within-group comparisons in published orthodontic articles.
• To identify associations between this statistical problem and study characteristics.
• To promote accurate statistical reporting for improved clinical decision-making.

Main Methods:

• A systematic search of four leading orthodontic journals (AJODO, EJO, ANGLE, OCR) was conducted up to December 2017.
• 339 articles were analyzed for the use and interpretation of within-group comparisons.
• Univariable and multivariable modified Poisson regression were used to identify predictors.

Main Results:

• 18% of included studies (60 out of 339) presented findings based solely on within-group comparisons.
• Statistical significance of the primary outcome was a strong predictor (RR: 2.33) of this flawed approach.
• The majority of articles were published in ANGLE (46%).

Conclusions:

• The use of within-group statistical testing and interpretation is prevalent in orthodontic research.
• Accurate statistical conduct and reporting are essential for optimal clinical decision-making.
• Further research should address the effect of time on these statistical practices.