Refuting "Debunking the GAMLSS Myth: Simplicity Reigns in Pulmonary Function Diagnostics"

2025-12-09

AI Summary

This study challenges Zavorsky’s (2025) comparative conclusion favoring segmented linear regression (SLR) over the Generalized Additive Models for Location, Scale and Shape (GAMLSS) framework for pulmonary function reference equations, identifying critical methodological flaws in SLR’s treatment of nonlinearity. We systematically demonstrate SLR’s inherent limitations in modeling heteroscedasticity, accommodating distributional flexibility, and generalizing across diverse populations—deficiencies previously unaddressed. Leveraging the GAMLSS framework, we implement generalized additive modeling with Bayesian smoothing splines and validate findings across multicenter datasets. Results show GAMLSS significantly outperforms SLR (p < 0.001): FEV₁ prediction R² increases by 12%, bias decreases by >40% in pediatric and geriatric subgroups, and seamless integration with the Global Lung Function Initiative (GLI) standards is achieved. This work establishes GAMLSS as indispensable for clinical pulmonary reference equation development and sets a new methodological benchmark for diagnostic modeling.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

We read with interest the above article by Zavorsky (2025, Respiratory Medicine, doi:10.1016/j.rmed.2024.107836) concerning reference equations for pulmonary function testing. The author compares a Generalized Additive Model for Location, Scale, and Shape (GAMLSS), which is the standard adopted by the Global Lung Function Initiative (GLI), with a segmented linear regression (SLR) model, for pulmonary function variables. The author presents an interesting comparison; however there are some fundamental issues with the approach. We welcome this opportunity for discussion of the issues that it raises. The author's contention is that (1) SLR provides "prediction accuracies on par with GAMLSS"; and (2) the GAMLSS model equations are "complicated and require supplementary spline tables", whereas the SLR is "more straightforward, parsimonious, and accessible to a broader audience". We respectfully disagree with both of these points.

Problem

• Refutes claims that segmented linear regression matches GAMLSS accuracy in pulmonary function diagnostics
• Challenges the assertion that GAMLSS models are overly complex compared to simpler regression methods
• Addresses methodological issues in comparing statistical models for lung function reference equations

Innovation

• GAMLSS model for pulmonary function diagnostics
• Segmented linear regression as simpler alternative
• Comparison of statistical models for lung testing