Multiple-bias Sensitivity Analysis Using Bounds

Abstract

Confounding, selection bias, and measurement error are well-known sources of bias in epidemiologic research. Methods for assessing these biases have their own limitations. Many quantitative sensitivity analysis approaches consider each type of bias individually, although more complex approaches are harder to implement or require numerous assumptions. By failing to consider multiple biases at once, researchers can underestimate-or overestimate-their joint impact. We show that it is possible to bound the total composite bias owing to these three sources and to use that bound to assess the sensitivity of a risk ratio to any combination of these biases. We derive bounds for the total composite bias under a variety of scenarios, providing researchers with tools to assess their total potential impact. We apply this technique to a study where unmeasured confounding and selection bias are both concerns and to another study in which possible differential exposure misclassification and confounding are concerns. The approach we describe, though conservative, is easier to implement and makes simpler assumptions than quantitative bias analysis. We provide R functions to aid implementation.

Are E-values too optimistic or too pessimistic? Both and neither!

“…Furthermore, what should be considered realistic will vary greatly across contexts. We suspect that these problems will increase as one includes more bias components in the bounding exercise (as in Smith et al 21 ).…”

Data Quality in Electronic Health Record Research: An Approach for Validation and Quantitative Bias Analysis for Imperfectly Ascertained Health Outcomes Via Diagnostic Codes

“… Gianfrancesco & Goldstein (2021) articulate four central challenges to validity of EHR-based research: issues of selection and representativeness, data availability and interpretation (including measurement error), missing measurements, and missing visits. No EHR study would be complete without reflecting upon all such considerations, and one can conduct a sensitivity analysis of the total bias that includes multiple threats to validity concurrently ( Smith et al, 2021 ).…”

Means to valuable exploration - II. How to explore data to modify existing claims and create new ones (revised)

“…This informs the readership of how probing against a particular bias scenario (Lash et al, 2014;Smith et al, 2021;VanderWeele & Mathur, 2020) relates to meeting a norm. Although seldom used, such Bayesian bias models make assumptions on bias transparent.…”