Skip to main content
OSTI.GOVU.S. Department of Energy Office of Scientific and Technical Information
U.S. Department of Energy
Office of Scientific and Technical Information
 

Effects of Omitting Non-confounding Predictors From General Relative-Risk Models for Binary Outcomes

Journal Article · · Journal of Epidemiology
 [1];  [2];  [2];  [3]
  1. Radiation Effects Research Foundation, Hiroshima (Japan); DOE/OSTI
  2. Radiation Effects Research Foundation, Hiroshima (Japan)
  3. Shiga Univ. of Medical Science (Japan)
+ Show Author Affiliations
The effects, in terms of bias and precision, of omitting non-confounding predictive covariates from generalized linear models have been well studied, and it is known that such omission results in attenuation bias but increased precision with logistic regression. However, many epidemiologic risk analyses utilize alternative models that are not based on a linear predictor, and the effect of omitting non-confounding predictive covariates from such models has not been characterized. We employed simulation to study the effects on risk estimation of omitting non-confounding predictive covariates from an excess relative risk (ERR) model and a general additive-multiplicative relative-risk mixture model for binary outcome data in a case-control setting. We also compared the results to the effects with ordinary logistic regression. For these commonly employed alternative relative-risk models, the bias was similar to that with logistic regression when the risk was small. More generally, the bias and standard error of the risk-parameter estimates demonstrated patterns that are similar to those with logistic regression, but with greater magnitude depending on the true value of the risk. The magnitude of bias and standard error had little relation to study size or underlying disease prevalence. Prior conclusions regarding omitted covariates in logistic regression models can be qualitatively applied to the ERR and the general additive-multiplicative relative-risk mixture model without substantial change. Quantitatively, however, these alternative models may have slightly greater omitted-covariate bias, depending on the magnitude of the true risk being estimated.
Research Organization:
National Academy of Sciences, Washington, DC (United States)
Sponsoring Organization:
USDOE Office of Environment, Health, Safety and Security (AU)
Grant/Contract Number:
HS0000031
OSTI ID:
1613772
Alternate ID(s):
OSTI ID: 22866121
Journal Information:
Journal of Epidemiology, Journal Name: Journal of Epidemiology Journal Issue: 3 Vol. 29; ISSN 0917-5040
Publisher:
Japan Epidemiological Association (JEA)Copyright Statement
Country of Publication:
United States
Language:
English

References (22)

Link Functions in Multi‐Locus Genetic Models: Implications for Testing, Prediction, and Interpretation journal April 2012
Relationship between anthropometric factors, radiation exposure, and colon cancer incidence in the Life Span Study cohort of atomic bomb survivors journal October 2012
Use of the Box-Cox transformation with binary response models journal January 1982
Biased estimates of treatment effect in randomized experiments with nonlinear regressions and omitted covariates journal January 1984
A geometric approach to assess bias due to omitted covariates in generalized linear models journal January 1993
General Relative risk Functions for Case-Control Studies journal July 1985
A Note on the Noncollapsibility of Rate Differences and Rate Ratios journal May 2016
Associations of Ionizing Radiation and Breast Cancer-Related Serum Hormone and Growth Factor Levels in Cancer-Free Female A-Bomb Survivors journal June 2011
Selecting Controls for Assessing Interaction in Nested Case-control Studies journal January 2003
Some Surprising Results about Covariate Adjustment in Logistic Regression Models
  • Robinson, Laurence D.; Jewell, Nicholas P.
  • International Statistical Review / Revue Internationale de Statistique, Vol. 59, Issue 2 https://doi.org/10.2307/1403444
journal August 1991
General Relative-Risk Models for Survival Time and Matched Case-Control Analysis journal December 1981
On the Application of Linear Relative Risk Regression Models journal March 1986
Modelling the Relative and Absolute Risks of Radiation-Induced Cancers journal January 1987
Effects of omitting a covariate in poisson models when the data are balanced journal June 2000
Describing time and age variations in the risk of radiation-induced solid tumour incidence in the Japanese atomic bomb survivors using generalized relative and absolute risk models journal January 1999
Risk Factors for Primary Breast Cancer in Japan: 8-Year Follow-Up of Atomic Bomb Survivors journal January 1997
Including known covariates can reduce power to detect genetic effects in case-control studies journal July 2012
Improving the efficiency of nested case-control studies of interaction by selecting controls using counter matching on exposure journal May 2004
General Relative risk Regression Models for Epidemiologic Studies1 journal November 1987
A Note on the Noncollapsibility of Rate Differences and Rate Ratios journal May 2016
Risk set sampling in epidemiologic cohort studies journal January 1996
Associations of Ionizing Radiation and Breast Cancer-Related Serum Hormone and Growth Factor Levels in Cancer-Free Female A-Bomb Survivors journal June 2011

Similar Records

Joint Non-Parametric Correction Estimator for Excess Relative Risk Regression in Survival Analysis with Exposure Measurement Error
Journal Article · Sun Feb 26 19:00:00 EST 2017 · Journal of the Royal Statistical Society: Series B (Statistical Methodology) · OSTI ID:1957676
Epidemiologic programs for computers and calculators: use of Poisson regression models in estimating incidence rates and ratios
Journal Article · Thu Jan 31 23:00:00 EST 1985 · Am. J. Epidemiol.; (United States) · OSTI ID:5751958
Use of logistic regression for modelling risk factors: with application to non-melanoma skin cancer
Journal Article · Tue Oct 31 23:00:00 EST 1978 · Am. J. Epidemiol.; (United States) · OSTI ID:6851781

Related Subjects

60 APPLIED LIFE SCIENCES
97 MATHEMATICS AND COMPUTING
Public
Environmental & Occupational Health
bias
binary outcomes
general relative-risk model
generalized nonlinear model
logistic regression
omitted covariate