DLNM Prediction & Interpretation

Extract and interpret DLNM predictions using crosspred() and crossreduce() in R.

Updated Mar 7, 2026

Search Tags:#prediction #interpretation #epidemiology #dlnm #r language #crosspred #crossreduce

Ready to run

npx skills add https://github.com/ntluong95/agent-skills-statistics --skill dlnm-prediction-interpretation

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: DLNM Prediction & Interpretation
Source: https://github.com/ntluong95/agent-skills-statistics/tree/main/skills/dlnm/prediction-interpretation
Command: npx skills add https://github.com/ntluong95/agent-skills-statistics --skill dlnm-prediction-interpretation

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps statisticians and epidemiologists understand and extract meaningful insights from complex Distributed Lag Non-Linear Models (DLNMs) by interpreting predictions and model outputs.

Core Features & Use Cases

Generate Predictions: Create predictions across the exposure-lag-response surface using crosspred().
Extract Key Metrics: Obtain overall RRs, RR matrices, and cumulative RRs from model outputs.
Interpret Centering: Understand and adjust the reference value (cen) for meaningful risk ratio calculations.
Summarize Effects: Reduce complex models to interpretable one-dimensional curves using crossreduce().
Use Case: After fitting a DLNM to air pollution and mortality data, use this Skill to generate and visualize the exposure-response curve and the lag-response curve at a specific pollution level, with clear interpretation of the centering value used.

Quick Start

Use the DLNM Prediction & Interpretation skill to generate predictions from the fitted model 'model' using the crossbasis object 'cb', centering at the median exposure.

Frequently Asked Questions about DLNM Prediction & Interpretation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I interpret a distributed lag non-linear model prediction in R?▼

To interpret a distributed lag non-linear model prediction in R, extract overall risk ratios and cumulative effects across the exposure-lag-response surface using crosspred, adjusting the centering value for meaningful baseline comparisons.

How do I use crossreduce to summarize DLNM exposure-response curves?▼

Use crossreduce to simplify complex DLNM models into interpretable one-dimensional exposure-response or lag-response curves, extracting key metrics and risk ratios relative to your specified centering value.

What does the centering value (cen) do in a DLNM risk ratio calculation?▼

The centering value in DLNM risk ratio calculations sets the reference exposure level for comparisons, ensuring meaningful relative risk interpretations across the exposure-lag-response surface.

How do I generate predictions from a fitted DLNM using crosspred?▼

Generate DLNM predictions by passing your fitted model and crossbasis object to crosspred, which computes overall risk ratios, cumulative effects, and the full exposure-lag-response surface matrix.

Are there common pitfalls when extracting cumulative effects from DLNMs?▼

Common DLNM pitfalls include misinterpreting the centering value, incorrectly extracting cumulative risk ratios, and failing to properly reduce the exposure-lag-response surface for one-dimensional interpretation.

Can I visualize the lag-response curve at a specific exposure level?▼

Yes, you can visualize the lag-response curve at a specific exposure level by generating predictions with crosspred and reducing the complex surface to an interpretable one-dimensional summary using crossreduce.