227055 Systematic measurement error's influence on disparities in national rates of children with special health care needs across Spanish and English speaking households

Tuesday, November 9, 2010

Adam C. Carle, MA, PhD , Health Policy and Clinical Effectiveness, Cincinnati Children's Hospital and Medical Center, Cincinnati, OH
Introduction: Non-English speaking children often experience poorer health than English speakers. Yet, research suggests large differences in the rates of US children with special health care needs (CSHCN) across English (14%) and Spanish speaking families (4.6%), suggesting that children from Spanish speaking homes may require fewer resources than children from English speaking homes. However, culture and translation may affect how parents answer questions, resulting in differences that reflect systematic measurement error (i.e., measurement bias) not true differences. Before concluding that disparities reflect true differences and allocating resources, research must establish that the instrument used to identify CSHCN, the CSHCN Screener, provides equivalent measurement in English and Spanish.

Methods: I used data from the 2005-2006 National Survey (NS) of CSHCN (n = 350,345) and multiple-group-multiple-indicator-multiple cause models (MG-MIMIC) to investigate whether the Screener provides equivalent measurement across English and Spanish.

Results: MG-MIMIC that included cross-group differences in education and income uncovered no significant measurement bias across English and Spanish translations. The model excluding education and income's influences, found significant measurement bias.

Discussion: Results demonstrate that responses provided equivalent measurement across English and Spanish among children with similar socio-demographic backgrounds, increasing confidence in research showing differences in the prevalence of CSHCN across English and Spanish speaking children examined within similar sociodemographic groups. However, comparisons across Spanish and English speaking families disregarding other sociodemographics will lead to biased results, calling into question work comparing CSHCN across Spanish and English families without taking into account systematic differences in the distribution of education and income.

Learning Areas:
Biostatistics, economics
Chronic disease management and prevention
Epidemiology

Learning Objectives:
At the end of the session, participants will be able to: 1. Define measurement bias and its potential influence on health disparities work. 2. Explain the importance of establishing equivalent measurement across individuals with diverse sociodemographic backgrounds. 3. Describe how measurement bias can impedes CSHCN disparities research specifically and disparities research generally.

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: In the past 5 years, I have published over 20 peer-reviewed articles and delivered over 40 national and international research presentations. Many of these have addressed children with special health care needs, psychometric models, disparities, or all three. For the research presented here, I worked individually, conducted the literature searches and summaries of previous related work, undertook the statistical analyses, and wrote the manuscript.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.