Latent Propensity Score Approaches for Causal Effect Estimation Allowing Covariate Measurement Error and Non-Directly Observable Outcome
Loading...
Links to Files
Permanent Link
Author/Creator
Author/Creator ORCID
Date
2020-01-20
Type of Work
Department
Mathematics and Statistics
Program
Statistics
Citation of Original Publication
Rights
Distribution Rights granted to UMBC by the author.
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu
Subjects
Abstract
Real-world evidence (RWE) is playing an increasing role in facilitating the development of innovative therapies and assessing the safety of post-marketed medical products. However, it is common that a covariate is measured with error in RWE studies, which violates the unconfoundedness assumption. Ignoring such error and using standard propensity score approaches may lead to biased ATE (average treatment effect) estimate. To help approach selection, the performance of three causal approaches allowing covariate measurement error were compared by simulation studies under various scenarios with respect to of the type of outcome (Gaussian or binary), the type of error-free covariate (continuous or discrete), the size of ATE and the size of measurement error. The results indicate that the bias correction approach is recommended under a Gaussian outcome and small measurement error, while the inverse probability weighting approach is recommended under a binary outcome and small measurement error; and the latent propensity score (LPS) approach using MCMC method in estimation is recommended under large measurement error, no matter outcome is Gaussian or binary. Besides measurement error, another common issue in RWE studies is a non-directly observable outcome measured indirectly through multiple manifest outcomes. By considering the non-directly observable outcome as a latent outcome and extending the existing LPS framework, we innovatively proposed an extended LPS approach allowing both covariate measurement error and non-directly observable outcome. It has two major advantages: First, it directly estimates the ATE on the true outcome. Second, it works when manifest outcomes are of different types. Simulations studies under various scenarios consistently indicate that this extended LPS approach works when the non-directly observable true outcome is Gaussian or binary, and it works the best with smaller ATE and larger covariate measurement error. Compared with the existing LPS approach that estimates the ATE on an observed manifest outcome (Gaussian or binary), the extended LPS approach shows better or at least competent performance in all the scenarios.