A Latent Growth Curve Model for Predicting Periodontal Disease Progression over 12 Months

Background: The progression of periodontal diseases is heterogeneous among individuals and related to behavioral, systemic, and clinical parameters.However, classical modeling methods usually do not account for interindividual diversity, hampering the precise prediction of risk.Aim: To understand the utility of latent growth curve modeling (LGCM) in describing changes in clinical attachment level (CAL) and predicting both baseline disease severity and follow-up trajectory over 12 months.Materials and methods: Simulated longitudinal data of 200 patients with baseline, 3, 6, and 12 months were used for modeling CAL.Latent growth curve modeling (LGCM) (structural equation modeling) estimated latent intercepts and slopes and included predictors, consisting of age, smoking status, diabetes status, oral health literacy, and plaque index.Goodness of fit was measured by the Chi-squared (χ²), comparative fit index (CFI), Tucker-Lewis index (TLI), root mean squared error approximation (RMSEA), and the standardized root mean residual (SRMR).Slope tertiles were used to define trajectory clusters, and group differences were examined using ANOVA and Chi-squared tests. Results:The mean CAL increased from 3.45 to 3.92 mm after 12 months (p < 0.001).Latent growth curve modeling indicated that there was a large inconsistency regarding both intercepts and slopes across classes (p < 0.01).Current smoking, diabetes, and higher plaque index were associated with more baseline CAL and accelerated progression, while better oral health literacy was protective.In trajectory analysis, stability and slow and fast progressors were observed, indicating variable disease courses.Fit indices demonstrated that the models fit very well (CFI = 0.985, TLI = 0.977, RMSEA = 0.042, SRMR = 0.028). Conclusion:The latent growth curve model successfully described the pattern of CALs in each subject and the contributing factors to disease progression.Trajectory-based stratification could potentially be useful in detecting high-risk patients and subsequent prevention interventions. Clinical implications:The LGCM concept could improve precision periodontology by providing risk-adapted treatments and appropriate selection of resources with better long-term outcomes using personalized care.