Gretchen Bandoli1, Kristin Palmsten2, Christina D Chambers1,3, Laura L Jelliffe-Pawlowski4,5, Rebecca J Baer1,4, Caroline A Thompson6,7. 1. Department of Pediatrics, University of California, San Diego, La Jolla, CA, USA. 2. HealthPartners Institute, Minneapolis, MN, USA. 3. Department of Family Medicine and Public Health, University of California, San Diego, La Jolla, CA, USA. 4. Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, CA, USA. 5. California Preterm Birth Initiative, University of California San Francisco School of Medicine, San Francisco, CA, USA. 6. Graduate School of Public Health, San Diego State University, San Diego, CA, USA. 7. Palo Alto Medical Foundation Research Institute, Sutter Health, Palo Alto, CA, USA.
Abstract
BACKGROUND: A "Table Fallacy," as coined by Westreich and Greenland, reports multiple adjusted effect estimates from a single model. This practice, which remains common in published literature, can be problematic when different types of effect estimates are presented together in a single table. The purpose of this paper is to quantitatively illustrate this potential for misinterpretation with an example estimating the effects of preeclampsia on preterm birth. METHODS: We analysed a retrospective population-based cohort of 2 963 888 singleton births in California between 2007 and 2012. We performed a modified Poisson regression to calculate the total effect of preeclampsia on the risk of PTB, adjusting for previous preterm birth. pregnancy alcohol abuse, maternal education, and maternal socio-demographic factors (Model 1). In subsequent models, we report the total effects of previous preterm birth, alcohol abuse, and education on the risk of PTB, comparing and contrasting the controlled direct effects, total effects, and confounded effect estimates, resulting from Model 1. RESULTS: The effect estimate for previous preterm birth (a controlled direct effect in Model 1) increased 10% when estimated as a total effect. The risk ratio for alcohol abuse, biased due to an uncontrolled confounder in Model 1, was reduced by 23% when adjusted for drug abuse. The risk ratio for maternal education, solely a predictor of the outcome, was essentially unchanged. CONCLUSIONS: Reporting multiple effect estimates from a single model may lead to misinterpretation and lack of reproducibility. This example highlights the need for careful consideration of the types of effects estimated in statistical models.
BACKGROUND: A "Table Fallacy," as coined by Westreich and Greenland, reports multiple adjusted effect estimates from a single model. This practice, which remains common in published literature, can be problematic when different types of effect estimates are presented together in a single table. The purpose of this paper is to quantitatively illustrate this potential for misinterpretation with an example estimating the effects of preeclampsia on preterm birth. METHODS: We analysed a retrospective population-based cohort of 2 963 888 singleton births in California between 2007 and 2012. We performed a modified Poisson regression to calculate the total effect of preeclampsia on the risk of PTB, adjusting for previous preterm birth. pregnancy alcohol abuse, maternal education, and maternal socio-demographic factors (Model 1). In subsequent models, we report the total effects of previous preterm birth, alcohol abuse, and education on the risk of PTB, comparing and contrasting the controlled direct effects, total effects, and confounded effect estimates, resulting from Model 1. RESULTS: The effect estimate for previous preterm birth (a controlled direct effect in Model 1) increased 10% when estimated as a total effect. The risk ratio for alcohol abuse, biased due to an uncontrolled confounder in Model 1, was reduced by 23% when adjusted for drug abuse. The risk ratio for maternal education, solely a predictor of the outcome, was essentially unchanged. CONCLUSIONS: Reporting multiple effect estimates from a single model may lead to misinterpretation and lack of reproducibility. This example highlights the need for careful consideration of the types of effects estimated in statistical models.
Authors: Hannah H Chang; Jim Larson; Hannah Blencowe; Catherine Y Spong; Christopher P Howson; Sarah Cairns-Smith; Eve M Lackritz; Shoo K Lee; Elizabeth Mason; Andrew C Serazin; Salimah Walani; Joe Leigh Simpson; Joy E Lawn Journal: Lancet Date: 2012-11-16 Impact factor: 79.321
Authors: Chander P Arora; Marian Kacerovsky; Balazs Zinner; Tibor Ertl; Iuliana Ceausu; Igor Rusnak; Serhiy Shurpyak; Meenu Sandhu; Calvin J Hobel; Daniel A Dumesic; Sandor G Vari Journal: Croat Med J Date: 2015-04 Impact factor: 1.351
Authors: Jennifer P Jain; Lila A Sheira; Edward A Frongillo; Torsten B Neilands; Mardge H Cohen; Tracey E Wilson; Aruna Chandran; Adaora A Adimora; Seble G Kassaye; Anandi N Sheth; Margaret A Fischl; Adebola A Adedimeji; Janet M Turan; Phyllis C Tien; Sheri D Weiser; Amy A Conroy Journal: AIDS Date: 2022-08-10 Impact factor: 4.632
Authors: Heather A Pines; Shirley J Semple; Steffanie A Strathdee; Craig W Hendrix; Alicia Harvey-Vera; Pamina M Gorbach; Carlos Magis-Rodríguez; Gustavo Martinez; Thomas L Patterson Journal: BMC Public Health Date: 2018-08-14 Impact factor: 3.295