Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Review
. 2020 Apr 1:209:107944.
doi: 10.1016/j.drugalcdep.2020.107944. Epub 2020 Feb 27.

Tutorial in Biostatistics: The use of generalized additive models to evaluate alcohol consumption as an exposure variable

Laura F White  1 , Wenqing Jiang  2 , Yicheng Ma  3 , Kaku So-Armah  4 , Jeffrey H Samet  4 , Debbie M Cheng  2
Affiliations
Review

Tutorial in Biostatistics: The use of generalized additive models to evaluate alcohol consumption as an exposure variable

Laura F White et al. Drug Alcohol Depend. .

Abstract

Alcohol consumption is a commonly studied risk factor for many poor health outcomes. Various instruments exist to measure alcohol consumption, including the AUDIT-C, Single Alcohol Screening Questionnaire (SASQ) and Timeline Followback. The information gathered by these instruments is often simplified and analyzed as a dichotomous measure, risking the loss of information of potentially prognostic value. We discuss generalized additive models (GAM) as a useful tool to understand the association between alcohol consumption and a health outcome. We demonstrate how this analytic strategy can guide the development of a regression model that retains maximal information about alcohol consumption. We illustrate these approaches using data from the Russia ARCH (Alcohol Research Collaboration on HIV/AIDS) study to analyze the association between alcohol consumption and biomarker of systemic inflammation, interleukin-6 (IL-6). We provide SAS and R code to implement these methods. GAMs have the potential to increase statistical power and allow for better elucidation of more nuanced and non-linear associations between alcohol consumption and important health outcomes.

Keywords: AUDIT-C; Alcohol consumption; Generalized additive models.

PubMed Disclaimer

Conflict of interest statement

Conflict of Interest

No conflict declared.

Figures

Figure 1.
Figure 1.
a) Example of difference between GAM, linear model and piecewise linear model; b) residual plot from linear regression model fit.
Figure 1.
Figure 1.
a) Example of difference between GAM, linear model and piecewise linear model; b) residual plot from linear regression model fit.
Figure 2.
Figure 2.
Illustrations of GAM and a (a) normally distributed outcome, (b) count outcome, and (c) dichotomous outcome. Solid lines show a generalized linear model fit, assuming the relationship between the outcome (Y, log(y) or logit(y)) and the covariate is linear. The dotted line shows the fit from a GAM, allowing a flexible spline fit to the data.
Figure 2.
Figure 2.
Illustrations of GAM and a (a) normally distributed outcome, (b) count outcome, and (c) dichotomous outcome. Solid lines show a generalized linear model fit, assuming the relationship between the outcome (Y, log(y) or logit(y)) and the covariate is linear. The dotted line shows the fit from a GAM, allowing a flexible spline fit to the data.
Figure 2.
Figure 2.
Illustrations of GAM and a (a) normally distributed outcome, (b) count outcome, and (c) dichotomous outcome. Solid lines show a generalized linear model fit, assuming the relationship between the outcome (Y, log(y) or logit(y)) and the covariate is linear. The dotted line shows the fit from a GAM, allowing a flexible spline fit to the data.
Figure 3.
Figure 3.
Results from fitting a generalized additive model to the data. a) Nonlinear portion of the model fit; this is the standard plot produced by most software packages; b) estimated values of log of IL-6 for varying levels of drinking from the GAM model with other covariates set to mean values.
Figure 3.
Figure 3.
Results from fitting a generalized additive model to the data. a) Nonlinear portion of the model fit; this is the standard plot produced by most software packages; b) estimated values of log of IL-6 for varying levels of drinking from the GAM model with other covariates set to mean values.
Figure 4.
Figure 4.
a) Example of a piecewise linear model, with GAM and linear regression model fit. b) Piecewise linear model fit of relationship between drinking and log of IL-6 show with GAM and linear model fit. Covariates in the model are all set to their mean values.
Figure 4.
Figure 4.
a) Example of a piecewise linear model, with GAM and linear regression model fit. b) Piecewise linear model fit of relationship between drinking and log of IL-6 show with GAM and linear model fit. Covariates in the model are all set to their mean values.
See this image and copyright information in PMC

References

    1. Altman DG, Royston P, 2006. Statistics Notes The cost of dichotomising continuous variables. BMJ 332, 1080. - PMC - PubMed
    1. Assessing Alcohol Problems-Second Edition [WWW Document], n.d. URL https://pubs.niaaa.nih.gov/publications/assessingalcohol/index.htm (accessed 11.20.19).
    1. Bell S, Daskalopoulou M, Rapsomaniki E, George J, Britton A, Bobak M, Casas JP, Dale CE, Denaxas S, Shah AD, Hemingway H, 2017. Association between clinically recorded alcohol consumption and initial presentation of 12 cardiovascular diseases: population based cohort study using linked health records. BMJ 356, j909 10.1136/bmj.j909 - DOI - PMC - PubMed
    1. Berhane K, Tibshirani RJ, 1998. Generalized additive models for longitudinal data. Can. J. Stat 26, 517–535. 10.2307/3315715 - DOI
    1. Burns E, Gray R, Smith LA, 2010. Brief screening questionnaires to identify problem drinking during pregnancy: A systematic review. Addiction 105, 601–614. 10.1111/j.1360-0443.2009.02842.x - DOI - PubMed

Publication types