 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
Innovations in Research Methodology
Background
Numerous reports and research studies have documented major disparities between key health indicators among racial and ethnic minorities and those among white populations. It is believed, however, that detailed information is needed on the health of these minority populations to effectively plan and target intervention programs that will narrow these disparities. To address these data gaps, the Disadvantaged Minority Health Improvement Act of 1990 (P. L. 101-527) authorized the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention (CDC) to establish a grants program to improve minority health statistics.
The Need for Improved Methodologies
In December 1991, the grants program held an agenda-setting workshop with approximately 85 leading experts in minority health and minority health statistics to assist in setting research priorities which address the critical issues affecting current and future efforts to collect and analyze data on racial and ethnic minority populations. These critical issues point to various problems with current research methodologies that fail to collect sufficient data on minority populations. As a result, the grants program supports projects that develop innovative technology and methodologies which better enable NCHS, CDC to accurately collect and analyze minority health data.
Cognitive Methods for Survey Development
Because it has been shown that current national surveys have been developed according to a dominant white, middle class standard, the grants program has prioritized the need to develop culturally sensitive data collection tools and techniques. To develop these tools and techniques, projects with the Northern California Cancer Center (NCCC) and the Survey Research Laboratory utilized methods drawn from cognitive psychology. Since cognitive methods have only recently been considered for use in survey research, these projects play a crucial part in developing these methods for use in minority health statistics. This research lays the foundation for developing national surveys that better represent the health needs of an increasingly heterogeneous population.
Focus Groups. This technique has been shown to be effective and efficient in assessing the perceptions, attitudes, and beliefs of a population. The NCCC used focus groups to develop surveys for Hispanic, Chinese, Vietnamese, and black populations and to identify variables that are conducive to cross-cultural research. Research assistants recruited from these communities conducted the focus groups, collecting qualitative data that gave the project team a better understanding of how the different minority groups perceive the survey questions and how these questions can be better worded. For the development of future surveys, guidelines were developed, including discussions of methodological issues and difficulties.
Think-Alouds. The Survey Research Laboratory developed think-aloud probes to determine how Mexican Americans, Puerto Ricans, blacks, and whites understand and answer national survey questions. These probes were designed to understand the four-step response process a respondent goes through for each survey item:
question
interpretation;
information
retrieval;
judgment
formation; and
response
editing.
This research identified cultural differences in the cognitive structures of each minority group, providing researchers with a better understanding of the cultural biases in survey development.
Sampling Techniques
A major problem with minority health statistics is a lack of data, leaving researchers unable to draw many reliable conclusions from national surveys. Most current methods for estimating the true number of a population affected by a disorder, for example, can be applied to large numbers, but these methods often fail when applied to sparse and/or incomplete data. Though oversampling can fill these data gaps, this approach is resource intensive and is, therefore, also limited in its use as an alternative method.
Sample-Resample Methodology. A project with the Western Consortium for Public Health adapted methods drawn from ecology for use in minority health statistics research. Similar to the capture-recapture method used to count wildlife populations, the sample-resample method allows researchers to work more accurately and efficiently from sparse and/or incomplete data. Estimations of the true number of individuals affected by a particular disorder are made by examining the intersection of different sources that list individuals' diagnoses. These sources include hospital in-patient records and records from outpatient clinics. This method can also determine the individuals missed by a particular study, making it useful in research using either sparse or large data sets.
Data Linkages to Reduce Racial Misclassification
The accurate classification of racial and ethnic minorities in data collection is needed to obtain accurate assessments of their health. Current classification practices create major data problems for planning intervention programs, particularly for Asian and Pacific Islanders (API's) and American Indians and Alaskan Natives (AIAN's). Data linkages, however, provide a method to begin developing protocol to reduce the occurrence of racial/ethnic misclassification by showing how one data set can fill the gaps of another data set.
The NCHS Linked Birth-Death data set and 1990 census data on APIs. By linking the NCHS Linked Birth-Death data set and 1990 Census data, a project with the Asian American Health Forum (AAHF) evaluated current methods of classifying the children of interracial and interethnic couples.
AIAN tribal data. To aid improvements in reporting race and ethnicity on death certificates, a project with the Northwest Portland Indian Health Board (NPAIHB) established a centralized tribal database and linked it with the Portland Area Indian Health Service patient registration file, State and other agency disease registries, the National Death Index, and other databases such as Medicare and Medicaid. From these linkages, researchers identified the geographic areas where misclassification is most prevalent, allowing them to direct intervention strategies effectively. This was also the first study to show the feasibility of reporting data for a population-based sample of AIANs enrolled in Parts A and B of Medicare.
Easier Data Access
Another priority of the grants program is to make existing national data bases more accessible to those interested in minority health statistics research. NCHS, CDC has supported the development of technology to address this need.
The Minority Health Database Catalog. A project with the University of North Carolina at Chapel Hill (UNC-CH) is developing the most expansive inventory of national and subnational data sets available for minority health research. In order to support minority health research worldwide, it is available via the World Wide Web on the UNC-CH Department of Biostatistics Home Page. The catalog currently contains over 550 descriptions.
This page last reviewed January 11, 2007
|
