Online Program Home
My Program

Abstract Details

Activity Number: 175
Type: Contributed
Date/Time: Monday, August 1, 2016 : 10:30 AM to 12:20 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #319407
Title: Optimal Sampling Designs of Two-Compartment Nonlinear Regression Models
Author(s): Noa Molshatzki* and Sandrah P. Eckel
Companies: University of Southern California and University of Southern California
Keywords: Air pollution ; Biomarkers ; Nonlinear regression ; Fisher information matrix ; Study design

Exhaled nitric oxide (FeNO) is a biomarker of airway inflammation increasingly of interest for clinical and epidemiological applications (e.g., asthma management and air pollution health effects studies). FeNO measured at multiple exhalation flow rates can be partitioned into airway and alveolar sources using nonlinear regression to estimate three "NO parameters" in a two compartment model. Despite the growing number of multiple flow FeNO studies, there is no standardized flow rate sampling design. In this work, we use statistical theory and simulation studies to derive optimal flow rate sampling designs for FeNO. Criteria for optimal designs include unbiased parameter estimation and minimization of standard errors (by inverse Fisher Information). We found that designs optimal for minimizing the SE of the different NO parameters were similar, typically including the minimum and maximum flows. Furthermore, optimal designs for nonlinear models are known to depend on parameters values, but we found similar optimal designs for the range of NO parameter values. Results from this theoretical exploration of optimal designs can inform guidelines for multiple flow FeNO studies.

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

Back to the full JSM 2016 program

Copyright © American Statistical Association