Simulation of Random Differential Periodontitis Outcome Misclassification with Perfect Specificity

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Clinical periodontitis measurement can involve classification errors, especially with partial-mouth protocols.
• While specificity is often perfect, sensitivity can be nondifferential, leading to an expectation of no bias in etiological studies.
• However, random errors can introduce differential misclassification, the impact of which is not well understood.

Purpose of the Study:

• To quantify the probability, magnitude, and expected impacts of differential periodontitis outcome misclassification due to random error.
• To provide guidance for interpreting research findings where nondifferential misclassification mechanisms are known to be present.

Main Methods:

• Simulated datasets with varying sample sizes, exposure effects, prevalence, incidence, and outcome sensitivity.
• Introduced random misclassification using a Bernoulli trial, repeating each scenario 10,000 times.
• Analyzed the impact on risk ratios and odds ratios.

Main Results:

• Differential misclassification occurred at least 37% of the time across simulations.
• Risk ratios were biased toward the null when sensitivity was higher in unexposed groups and away from the null when higher in exposed groups.
• Bias extent for absolute sensitivity differences ≥0.04 ranged from 0.05 to 0.19; odds ratio bias depended on incidence, sensitivity, and effect size.

Conclusions:

• Simulation results offer quantitative insights into the effects of random classification errors in periodontitis research.
• Understanding the probability and impact of these errors is crucial for accurate research scrutiny.
• Findings aid in interpreting studies with known nondifferential outcome misclassification and perfect specificity.