|
Activity Number:
|
285
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #305076 |
|
Title:
|
Estimating Household Income Percentiles from a Public Health Survey Using Log-Linear Bootstrap Interpolation
|
|
Author(s):
|
Robert Feyerharm*+
|
|
Companies:
|
Oklahoma State Department of Health
|
|
Address:
|
1000 NE 10th Street, Oklahoma City, OK, 73117,
|
|
Keywords:
|
public health survey ; log-linear regression ; bootstrap
|
|
Abstract:
|
Oklahoma PRAMS is a population-based surveillance system administered by the Oklahoma State Department of Health which annually surveys 2,000--3,000 women who have recently delivered a live infant in Oklahoma. PRAMS participants are asked to report the number of dependents living in their household and their annual household income. This information is valuable for estimating the percentage of births to women in Oklahoma having household incomes at or below 185 percent of the Federal Poverty Level (FPL), and thus are qualified to receive State Medicaid assistance for prenatal care and delivery costs. The categorical nature of PRAMS income data discourages a direct calculation of an income percentile estimate. The author proposes log-linear regression to interpolate household income percentiles < = 185% of FPL, using bootstrap resampling to estimate confidence intervals.
|
