Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 83 - Contributed Poster Presentations: Health Policy Statistics Section
Type: Contributed
Date/Time: Monday, August 3, 2020 : 10:00 AM to 2:00 PM
Sponsor: Health Policy Statistics Section
Abstract #313422
Title: Causal Inference with Cross-Temporal Design
Author(s): Yi Cao* and Roee Gutman
Companies: and Brown University
Keywords: causal inference; instrumental variable; temporal correlation; Bayesian analysis

Driven by the task of investigating the impact of the hospice use on the end-of-life expenditures of nursing home residents, we formulate a causal inference framework to address the possibility of confounding between hospice users to non-users. We take advantage of the expansion of hospice use between 2004 and 2009 and construct a quasi-experimental design that assumes residents in 2009 are encouraged to use hospices compared to residents in 2004. The year represents a binary encouragement variable. We incorporate this design within the principal stratification framework and partition the residents into 3 groups: new-users (compliers), traditional-users (always-takers), and non-users (never-takers). The intention-to-treat effect within the new-user group describes the effect of hospice use. Because the year of service may impact the potential outcomes regardless of hospice use, the stochastic exclusion restrictions assumptions are violated. We use a Bayesian approach for inference and apply our method to estimate the effect of hospice use on end-of-life expenditures. We perform a simulation study and sensitivity analysis to assess the validity of the estimates.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program