Activity Number:
|
189
- Nonparametric Methods in Big or Complex Data
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 4, 2020 : 10:00 AM to 2:00 PM
|
Sponsor:
|
Section on Nonparametric Statistics
|
Abstract #310945
|
|
Title:
|
Estimating Treatment Effect in Electronic Health Records (EHRs) with Differential Number and Spacing of the Follow-Ups Visits Between the Treatment Groups: A 3-Step Novel Approach
|
Author(s):
|
Yingjie Weng* and Manisha Desai
|
Companies:
|
Stanford Univ, School of Medicine and Stanford University
|
Keywords:
|
Electronic Health Records (EHRs) ;
differential follow-ups;
propensity score matching;
nearest neighbor matching
|
Abstract:
|
The use of electronic health records (EHRs) for research and program evaluation is promising. Unlike prospective data collection in traditional clinical trials, however, data recorded in EHRs are triggered by the needs that arise in the care of the patient and therefore present challenges when evaluating the impact of treatment on change or trajectory. We propose a novel approach to estimate the treatment effect under the setting of differential follow-ups. The method involves the use of propensity score-based techniques that incorporate patient-level features as well as the timing of and length of time between visits. Assuming differential missing data mechanisms between treatment arms, we evaluate the statistical properties of the method and compare it to standard longitudinal tools including generalized estimating equations and generalized linear mixed effects models. The method will be implemented in a real-world example.
|
Authors who are presenting talks have a * after their name.