|
Activity Number:
|
108
- Innovations in Testing and Inference
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 8, 2022 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
|
Abstract #322297
|
|
|
Title:
|
Use of Previously Published Results for More Powerful Testing of Factors Related to Adverse Birth Outcomes
|
|
Author(s):
|
Sergey Tarima* and Wanlin Juan and Zhanna Zenkova and Mary Homan
|
|
Companies:
|
Division of Biostatistics, Medical College of Wisconsin and Medical College of Wisconsin and Tomsk State University and Medical College of Wisconsin
|
|
Keywords:
|
auxiliary information;
mean squared error;
variance;
external information
|
|
Abstract:
|
Preterm births and the associated complications impact physical and mental health of mothers and children as well the health of the community. The rates of adverse birth outcomes within minority populations are especially troubling perpetuating a significant health disparity that threatens health equity. Analyses of factors related to adverse birth outcomes based on the Wisconsin Pregnancy Resource Assessment Monitoring System (PRAMS) can be improved using previously published results (additional information). It is possible that the published results estimate different quantities creating risk of bias. This manuscript explores a mean squared error (MSE) minimization approach for incorporating previously published results. Our MSE minimization allows preserving consistency of estimation even if additional information is not appropriate for a given dataset. If additional information has no conflict with the data, then a more accurate estimation and a more powerful testing is achieved with the MSE minimization. In addition to the PRAMS example, we present Monte-Carlo simulation studies for type 1 error, power and coverage probability of parametric bootstrap confidence intervals.
|
Authors who are presenting talks have a * after their name.
