Modelling disease activity in juvenile dermatomyositis: A Bayesian approach

Area of Science:

• Rheumatology
• Pediatrics
• Biostatistics

Background:

• Juvenile dermatomyositis (JDM) is the most common idiopathic inflammatory myopathy in children, marked by muscle and skin inflammation.
• JDM presents with symmetric proximal muscle weakness and characteristic skin symptoms, exhibiting a variable clinical course and prognosis.

Purpose of the Study:

• To develop a robust statistical model for analyzing longitudinal outcomes in JDM.
• To account for within- and inter-individual variability in disease activity.
• To identify clinical markers associated with disease progression and prognosis.

Main Methods:

• A Bayesian joint model was implemented to analyze four longitudinal outcomes in JDM.
• The model incorporated subject-specific random effects to capture correlations among outcomes.
• A hurdle model approach was used to handle excess outcomes, with clinical markers as covariates.

Main Results:

• The developed model demonstrated good performance, providing efficient parameter estimations and accurate predictions of disease activity.
• A significant correlation was found between two outcome variables, indicated by the posterior distribution of random intercepts.
• Several clinical markers and symptoms were identified as significantly associated with JDM disease activity.

Conclusions:

• The statistical model effectively captures the complex dynamics of JDM.
• Identified clinical markers can aid in stratifying JDM patients based on prognosis.
• Findings support improved clinical decision-making and guide future research on JDM outcome markers.