Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Using generalized additive models to reduce residual confounding.

Andrea Benedetti1, Michal Abrahamowicz

  • 1McGill University, Department of Epidemiology and Biostatistics, Canada.

Statistics in Medicine
|December 8, 2004
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

PROTOCOL: Comparative Effectiveness of Tools for Communicating Health Research Results to Study Participants and Others With Relevant Lived Experience: A Living Systematic Review of Randomized Controlled Trials.

Campbell systematic reviews·2026
Same author

Comparison of the accuracy of latent factor and sum scoring of the Hospital Anxiety and Depression Scale to screen for major depression: An individual participant data meta-analysis.

Journal of affective disorders·2026
Same author

The impact of opioid prescribing guidelines on opioid dispensation: A systematic review and meta-analysis.

Canadian journal of pain = Revue canadienne de la douleur·2026
Same author

Reply to Li and Ni "Positive Airway Pressure Therapy for Cognitive Enhancement in Parkinson's Disease: From Statistical Signals to Clinical Practice".

Sleep·2026
Same author

Meta-Analysis of Median Survival Times With Inverse-Variance Weighting.

Statistics in medicine·2026
Same author

Human Prion Disease: Pathogenesis, Diagnosis and Public Health.

Viruses·2026
Same journal

Optimal Weighted Tests for Replication Studies and the 'Two-Trials Rule' With Multiple Hypotheses.

Statistics in medicine·2026
Same journal

Identifiable Copula-Double-Cox Models: A Fully Parametric Framework for Dependent Right-Censored Survival Data.

Statistics in medicine·2026
Same journal

Moving From Individualized Risk-Based Prevention to Benefit-Based Prevention: Estimating Individualized Life-Years Gained From Prevention Services as a Basis for Eligibility.

Statistics in medicine·2026
Same journal

A Mixture of Distributed Lag Non-Linear Models to Account for Spatially Heterogeneous Exposure-Lag-Response Associations.

Statistics in medicine·2026
Same journal

Practical Considerations for Gaussian Process Modeling for Causal Inference in Quasi-Experimental Studies With Panel Data.

Statistics in medicine·2026
Same journal

Covariate Adjustment for Wilcoxon Two Sample Statistic and Test.

Statistics in medicine·2026
See all related articles

Generalized additive models (GAM) reduce residual confounding from continuous variables better than traditional regression. GAMs improve accuracy and avoid errors when confounder-outcome relationships are non-linear.

Area of Science:

  • Epidemiology
  • Biostatistics
  • Statistical Modeling

Background:

  • Confounding by continuous variables is typically managed using linear or categorical terms in regression models.
  • Residual confounding arises from mis-modeling the confounder's effect on the outcome.
  • Parametric models require pre-specified functional forms, which are often unknown and can lead to bias.

Purpose of the Study:

  • To compare the efficacy of parametric multiple logistic regression versus generalized additive models (GAM) in controlling for continuous confounders.
  • To investigate implementation issues of GAMs, including degrees of freedom selection and confounder inclusion/representation criteria.
  • To assess the impact of confounder-outcome association shape, sample size, and exposure-confounder correlation.

Main Methods:

Related Experiment Videos

  • Simulations were conducted to compare parametric logistic regression and non-parametric GAM extensions.
  • The study evaluated scenarios with linear and non-linear confounder-outcome relationships.
  • Key implementation aspects of GAMs were examined, such as variable selection and smoothing parameter choices.

Main Results:

  • When confounder-outcome associations were non-linear, GAMs reduced mean squared error and Type I error inflation for the adjusted exposure effect compared to parametric models.
  • GAMs performed comparably to parametric logistic regression when the confounder-outcome relationship was truly linear.
  • Joint non-linear modeling of exposure and confounder is necessary to avoid spurious non-linearity findings.

Conclusions:

  • Generalized additive models (GAM) offer significant improvements over conventional parametric modeling for reducing residual confounding by continuous variables.
  • GAMs provide a more flexible and robust approach, particularly when the functional form of confounder-outcome relationships is unknown or non-linear.
  • The findings support the use of GAMs for more accurate and reliable estimation of exposure effects in epidemiological and biostatistical analyses.