DLNM Meta-Analysis

What problem does it solve?

This Skill streamlines the process of combining results from multiple Distributed Lag Non-Linear Models (DLNMs) fitted to different geographical locations or datasets, enabling robust meta-analysis and the identification of overall trends and location-specific variations.

Core Features & Use Cases

• Two-Stage Meta-Analysis: Facilitates fitting city-specific DLNM models and then pooling their reduced coefficients using packages like mixmeta or mvmeta.
• Meta-Regression: Allows for the investigation of heterogeneity by including location-level covariates in the meta-analysis model.
• BLUP Estimation: Computes Best Linear Unbiased Predictions to obtain refined, location-specific estimates that borrow strength across studies.
• Use Case: When analyzing air pollution's impact on health across multiple cities, this Skill helps pool the city-specific exposure-response curves to derive a global estimate and understand how factors like local climate or socioeconomic status modify the effect.

Quick Start

Use the DLNM Meta-Analysis skill to pool the crossreduced coefficients and their variance-covariance matrices from multiple city-specific DLNM models using the 'reml' method in mixmeta.