emy of Sciences to provide information to the panel on DHHS Collection of Race and Ethnicity Data. The panel was convened to examine the adequacy of data on race and ethnicity collected or used by DHHS programs, and will issue guidance to the DHHS regarding the following:
What data on race and ethnicity are private organizations and providers collecting?
What is the availability and quality of the data collected?
How can private organizations benefit from the collection of these data?
How are the data used?
What are the barriers to collecting these data?
AAHP was asked to evaluate whether and how health plans collect racial and ethnic data. The panel believed there would be value in using an interview method that provided for and encouraged interactive questioning. AAHP conducted telephone interviews with a sample of member health plans across the country to identify and highlight the issues surrounding such data collection and summarize these efforts.
A sample of 30 health plans was selected for interviews using a twostep methodology.
Step 1. Sixteen AAHP member health plans were chosen, because these plans were known to have initiated activities related to this project.
Step 2. Fourteen health plans were selected from the pool of health plans responding to the 2002 AAHP Industry Survey (n = 194). A subset was created from these 194 health plans using the following criteria: current AAHP membership and enrollment of at least 100,000 members. A random selection of 14 plans was chosen from the subset.2
The list of responders to the 2002 AAHP Annual Industry Survey was used as the sampling frame for this project because it contained health plans verified as eligible to participate and had current contact information. The 2002 industry survey sampling methodology used the following selection criteria: five large national plans were sampled with certainty, because to exclude them from the sample would distort the national data, as they represent a large population of health plan members. A subsequent sample of additional plans was selected in a randomly stratified manner as to enrollment: very large (≥6 million members), large (>370,000 but <6 million members), medium (130,000 to <370,000 members), and small (<130,000 members).