Disparities in Quality of Care for Midlife Adults (Ages 45–64) Versus Older Adults (Ages >65). 3.2 Methods



The analysis used data from MEPS 2004–2006. The MEPS is a health survey developed to analyze health care use, expenditures and insurance coverage for the U.S. civilian noninstitutionalized population. The MEPS Household Component (MEPS HC) provides estimates of respondents' demographic and socioeconomic characteristics, access to care, health insurance coverage and effectiveness of care for an array of priority clinical conditions, including cardiovascular disease. The MEPS also collects information on diabetes care effectiveness separately through a self-survey, the MEPS Diabetes Care Supplement. All noninstitutionalized MEPS respondents ≥45 were included in this analysis. The overall 2004 MEPS HC response rate was 63.1 percent; for 2005, it was 61.3 percent; and for 2006, it was 58.3 percent. The MEPS protocol involves computer assisted personal interviewing (CAPI). Data from the 2004–2006 MEPS were pooled to bolster and ensure reliable estimates for comparative analysis across race/ethnicity, gender

Study Sample

The analysis focused on noninstitutionalized MEPS HC respondents ≥45 with self-identified diabetes (N=5,077) or self-identified cardiovascular disease (N=23,235). The sample of all diabetics was identified from those who responded in the affirmative to a MEPS HC question asking if they ever received a diagnosis of diabetes. Respondents were required to complete the MEPS Diabetes Care Supplement containing additional diabetes measures of care effectiveness.

The sample of all those with cardiovascular disease was identified from people who responded in the affirmative to a MEPS HC question asking if they ever received a diagnosis of high blood pressure (hypertension), high cholesterol, coronary heart disease, angina or angina pectoris, a heart attack or myocardial infarction, a stroke or transient ischemic attack (TIA or "ministroke";)

Data Analysis

The basic analytic approach was to compare relative differences in all health care quality measures between each age group, stratified by racial/ethnic and gender groups. Age groups included midlife adults 45–64 and older adults ≥65 (Medicare population). Race/ethnicity groups included non-Hispanic Whites, non-Hispanic Blacks and Hispanics of any race; gender groups included males and females. A fixed reference group was used to assess group differences in use of services. The reference group for age group differences was Medicare-eligible (>65) adults. The reference group for gender differences was males. In terms of race/ethnicity comparison, the reference group was Whites. All relative differences between groups were expressed as relative rates. Statistical significance was assessed using z tests.

The MEPS dataset collected sociodemographic information that allowed age group-based and race/ ethnicity comparison of non-Hispanic Whites, non-Hispanic Blacks and Hispanics of all races, and gender comparison of males and females. Because of data limitations, reliable estimates (with minimum cell size criteria of 100 observations or relative standard error >0.3) were not possible for all race/ethnicity, gender and age strata, and unreliable estimates were not included in the tables (identified by *** in Tables 4–10).


This analysis used measures from MEPS, reflecting three domains of care quality described by the IOM: effectiveness of care (with different measures for diabetes vs. cardiovascular patients), timeliness/access to care and patient-centeredness in care (Table 3). Diabetes effectiveness of care includes measures of HbA1c, blood pressure and cholesterol testing; eye and foot examinations; and preventive care, including adult vaccinations, dental check-up and receipt of smoking advice (for self-identified smokers). Measures of cardiovascular care effectiveness include blood pressure testing and receipt of lifestyle counseling for diet and exercise. Access to care includes indicators for delays in getting necessary medical or dental care or prescription medications, and an inability to get care when desired. Patient-centeredness in care includes indicators for patient-physician communication, perception of respect and amount of time a physician spends with a patient. Access to care and patient-centered measures are reported separately for the diabetes and cardiovascular disease populations.

View full report


"report.pdf" (pdf, 548.24Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®