Accounting for Data Errors Discovered from an Audit in Multiple Linear Regression

Abstract

SummaryA data coordinating team performed on-site audits and discovered discrepancies between the data sent to the coordinating center and that recorded at sites. We present statistical methods for incorporating audit results into analyses. This can be thought of as a measurement error problem, where the distribution of errors is a mixture with a point mass at 0. If the error rate is non-zero, then even if the mean of the discrepancy between the reported and correct values of a predictor is 0, naive estimates of the association between two continuous variables will be biased. We consider scenarios where there are 1) errors in the predictor, 2) errors in the outcome, and 3) possibly correlated errors in the predictor and outcome. We show how to incorporate the error rate and magnitude, estimated from a random subset (the audited records), to compute unbiased estimates of association and proper confidence intervals. We then extend these results to multiple linear regression where multiple covariates may be incorrect in the database and the rate and magnitude of the errors may depend on study site. We study the finite sample properties of our estimators using simulations, discuss some practical considerations, and illustrate our methods with data from 2815 HIV-infected patients in Latin America, of whom 234 had their data audited using a sequential auditing plan.

When exposure is subject to nondifferential misclassification, are validation data helpful in testing for an exposure–disease association?

“…In fact, the former article provides an example of a validation study in the context of a study drawing upon electronic health record databases, with further such examples including Cea Soriano, Soriano‐Gabarró & Garca Rodrguez (). And in a related direction, Shepherd and Yu () and Shepherd, Shaw & Dodd () consider “data auditing” of clinical data, where the choice of how many (and which) records to audit is, in essence, a choice of how much (and which) validation data to collect. In these contexts and many others, careful thinking about the cost of validation data and the benefit toward the stated inferential goals is of evident importance.…”

Evaluating risk-prediction models using data from electronic health records

“…However, errors in outcomes and risk factors could be correlated due to their shared dependence on patient characteristics. Research has focused on correcting for correlated errors in covariates and continuous outcomes [Shepherd, Shaw and Dodd (2012), Shepherd and Yu (2011)]. Further research is needed to correct for correlated errors in covariates and binary outcomes.…”

Influence factor analysis and prediction models for component removal time in manufacturing

“…Due to its good regression and prediction ability, it can be widely used to deal with such a variety of issues as biological, economic, chemical and engineering problems. [30][31][32] In this work, it is used to predict the removal time of a component in an assembly.…”