• J A Brown, S E Nederend, R D Hays, P F Short, and D O Farley.
• RAND, Santa Monica, CA 90407-2138, USA. JulievBrown@rand.org
• Med Care. 1999 Mar 1;37(3 Suppl):MS79-88.

ObjectivesThe authors describe the process used to develop and test survey items targeted to Medicaid consumers for the Consumer Assessment of Health Plans Study (CAHPS). In addition, the authors highlight the special challenges in locating and surveying Medicaid recipients and provide recommendations for increasing response rates.MethodsThe RAND CAHPS team reviewed the literature and existing questionnaires to identify health care issues and concepts important to Medicaid consumers. Three focus groups and 66 one-on-one cognitive interviews were conducted to test the relevance of our concepts and items and to identify additional concepts important to Medicaid consumers. After the cognitive interviews, the CAHPS Medicaid consumer survey was field tested using a sample of 930 adults and children receiving both Medicaid and Aid to Families with Dependent Children in Los Angeles County and Oklahoma. To determine if one particular mode were preferable for surveying a Medicaid population, our field test sample was divided randomly into a telephone-mode sample, a mixed-mode sample, and a second telephone-mode sample with enhanced locating procedures. Before finalizing the CAHPS 1.0 surveys, the full CAHPS item set was subjected to a formal literacy review.ResultsThe results of the focus groups and cognitive testing informed iterative versions of the list of concepts addressed by the Medicaid-targeted items. Concepts that were not relevant to Medicaid consumers or that consumers were unable to accurately attribute to a health plan were discarded. New concepts addressing important aspects of health care and the health care experience of Medicaid consumers were identified and added. Item wording and format were revised and refined based on the findings from focus groups, cognitive testing, the field test, and the formal literacy review. In the field test, the mixed-mode method achieved the best results with a 56% completion rate.ConclusionsThe testing and formatting efforts described in this article, in combination with a formal literacy review, led to the development of a Medicaid questionnaire that measures the important health care experiences of Medicaid consumers in a format that is "respondent-friendly." Our recommendations for surveying Medicaid recipients can benefit any survey of a Medicaid population.

25 keywords...

hide…