On quantifying the magnitude of confounding.

*(English)*Zbl 1437.62504Summary: When estimating the association between an exposure and outcome, a simple approach to quantifying the amount of confounding by a factor \(Z\), is to compare estimates of the exposure-outcome association with and without adjustment for \(Z\). This approach is widely believed to be problematic due to the nonlinearity of some exposure-effect measures. When the expected value of the outcome is modeled as a nonlinear function of the exposure, the adjusted and unadjusted exposure effects can differ even in the absence of confounding [S. Greenland et al., Stat. Sci. 14, No. 1, 29–46 (1999; Zbl 1059.62506)]; we call this the nonlinearity effect. In this paper, we propose a corrected measure of confounding that does not include the nonlinearity effect. The performances of the simple and corrected estimates of confounding are assessed in simulations and illustrated using a study of risk factors for low birth-weight infants. We conclude that the simple estimate of confounding is adequate or even preferred in settings where the nonlinearity effect is very small. In settings with a sizable nonlinearity effect, the corrected estimate of confounding has improved performance.

##### MSC:

62P10 | Applications of statistics to biology and medical sciences; meta analysis |

PDF
BibTeX
XML
Cite

\textit{H. Janes} et al., Biostatistics 11, No. 3, 572--582 (2010; Zbl 1437.62504)

Full Text:
DOI

##### References:

[1] | (1980) |

[2] | Smoking and lung cancer: recent evidence and a discussion of some questions, 22, 173-203 (1959) |

[3] | Testing collapsibility in contingency tables, 48, 197-205 (1986) · Zbl 0608.62063 |

[4] | Indirect corrections for confounding under multiplicative and additive models, 13, 119-130 (1988) |

[5] | Sensitivity analysis, Monte-Carlo risk analysis, and Bayesian uncertainty assessment, 21, 579-583 (2001) |

[6] | The impact of prior distributions for uncontrolled confounding and response bias: a case study of the relation of wire codes and magnetic fields to childhood leukemia, 98, 47-54 (2003) · Zbl 1047.62106 |

[7] | Multiple-bias modeling for analysis of observational data (with discussion), 168, 267-308 (2005) · Zbl 1099.62129 |

[8] | Inference on collapsibility in generalized linear models, 36, 771-782 (1994) · Zbl 0850.62559 |

[9] | Closed form and dually consistent methods for inference on strict collapsibility in 2 x 2 x K and 2 x J x K tables, 37, 335-343 (1988) |

[10] | Confounding in health research, 22, 189-212 (2001) |

[11] | Causal diagrams for epidemiologic research, 10, 37-48 (1999) |

[12] | Identifiability, exchangeability, and epidemiological confounding, 15, 413-418 (1986) |

[13] | Confounding and collapsibility in causal inference, 14, 29-46 (1999) · Zbl 1059.62506 |

[14] | Specification tests in econometrics, 46, 1251-1271 (1978) · Zbl 0397.62043 |

[15] | A definition of causal effect for epidemiological research, 58, 265-271 (2004) |

[16] | Estimating causal effects from epidemiological data, 60, 578-586 (2006) |

[17] | A statistical test for the equality of differently adjusted incidence rate ratios, 167, 512-522 (2008) |

[18] | (2000) |

[19] | (1982) |

[20] | A sensitivity analysis to separate bias due to confounding from bias due to predicting misclassification by a variable that does both, 11, 544-549 (2000) |

[21] | Estimating causal effects, 31, 422-429 (2002) |

[22] | Confounding: essence and detection, 114, 593-603 (1981) |

[23] | A comparison of cluster-specific and population-averaged approaches for analyzing correlated binary data, 59, 25-35 (1991) |

[24] | On the application of probability theory to agricultural experiments: essay on principles, section 9, 5, 465-480 (1923) · Zbl 0955.01560 |

[25] | Causal diagrams for empirical research, 82, 669-710 (1995) · Zbl 0860.62045 |

[26] | (2000) |

[27] | Quantifying and reporting uncertainty from systematic errors, 14, 459-466 (2003) |

[28] | The control of confounding by intermediate variables, 8, 679-701 (1989) |

[29] | (1998) |

[30] | Marginal structural models versus structural nested models as tools for causal inference, 95-134 (1999) |

[31] | Identifiability and exchangeability for direct and indirect effects, 3, 143-155 (1992) |

[32] | (2002) |

[33] | Estimating causal effects of treatments in randomized and nonrandomized studies, 66, 688-701 (1974) |

[34] | Bayesian inference for causal effects: the role of randomization, 6, 34-58 (1978) · Zbl 0383.62021 |

[35] | Marginal structural models as a tool for standardization, 14, 680-686 (2003) |

[36] | Confounding in epidemiologic studies: the adequacy of the control group as a measure of confounding, 43, 751-765 (1987) · Zbl 0715.62219 |

[37] | Collapsibility of multidimensional contingency tables, 40, 328-340 (1978) · Zbl 0413.62041 |

This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.