DLNM Model Specification

What problem does it solve?

This Skill simplifies the process of specifying and fitting complex Distributed Lag Non-Linear Models (DLNMs) for time series data, particularly in environmental epidemiology.

Core Features & Use Cases

• Canonical Model Structure: Provides a template for fitting GLMs with cross-basis terms, time trends, and confounders.
• Outcome & Family Guidance: Recommends appropriate families (quasi-Poisson, Poisson, negative binomial) based on outcome type and over-dispersion.
• Confounding Control: Details strategies for adjusting for time trends, seasonality, meteorological factors, and other temporal confounders.
• Use Case: Fit a quasi-Poisson GLM to daily mortality data, including a DLNM for temperature, natural splines for long-term trends and day-of-week, and a separate spline for daily temperature.

Quick Start

Fit a DLNM model for daily deaths using cross-basis for temperature, natural splines for time and temperature, and day of week.