Work

Estimating and testing blip effects of treatments in sequence via standardized point effects of treatments

Year: 2025

Type: article

Abstract: In longitudinal studies, treatments are often assigned in the form of a sequence to achieve a certain outcome of interest. The blip effect of treatment in sequence is the net effect of treatment on the outcome. In this article, we introduce a method of estimating and testing the blip effects via the standardized point effects of treatments in sequence. First, we apply available methods to estimate the point effects referring to single-point treatments. Then we standardize the point effects to a small number of strata of relevance to the blip effects of interest. Finally, we use the standardized point effects to estimate and test the blip effects. Our method addresses two issues in complex longitudinal studies: a dimension reduction without strict treatment assignment conditions and a targeted analysis of the blip effects of interest across different times. The simulation study shows that our method achieves unbiased estimates of the blip effect, maintains nominal coverage probability, and demonstrates high power for hypothesis testing. A medical example illustrates the application of our method in observational studies.

Source: Frontiers in Applied Mathematics and Statistics

DOI: 10.3389/fams.2025.1650059

Authors: Yaqin Liao, Yihong Lan, Li Yin, Xiaoqin Wang

Institutions: Xiamen University, Karolinska Institutet, University of Gävle

Language: English

Topic: Advanced Causal Inference Techniques

Subfield: Statistics and Probability

Field: Mathematics

Domain: Physical Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $1,418