Defining a Health Care Quality Measure
Although many measures are available to assess quality of care, this report focuses on clinical performance measures. Previous authors have defined clinical performance measures as
"tools that assess the delivery of clinical services…[They] estimate the extent to which a [health care] provider delivers clinical services that are appropriate for the patients' condition; provides the clinical services safely, competently, and in the appropriate time frame; and achieves the desired outcomes in terms of those aspects of patient health and satisfaction that can be affected by clinical services"; (Beal et al 2004;
Clinical performance measures can address processes, outcomes, access and patient experiences. There is no standardized set of categories for evaluating quality of care, but the categories developed by Avedis Donabedian for evaluating quality of care includes an assessment of structure, process and outcomes, and are widely used today (AHRQ 2008). Structural measures are not as strongly recommended by the IOM because there is less evidence of an association between structure and outcome (Beal et al 2004). Thus, this literature review focuses on process and outcomes of care measures for the priority condition of diabetes and, to a lesser extent, for cardiovascular conditions and depression. Process and outcomes measures can be used to measure any IOM quality domain of equity (the overarching domain), safety, effectiveness, timeliness/accessibility, patient-centeredness and efficiency of care (AHRQ 2008).
AHRQ defines process of care measures as those assessing whether "appropriate physician and other provider activities are carried out to deliver care"; (e.g., percentage of females of specified age receiving mammography; percentage of patients with asthma for whom appropriate medications are ordered [AHRQ 2008]). Outcomes of care measures assess "results of physician and other provider activities"; (e.g., experience or level of satisfaction with care; test results within a range indicating
The remainder of this chapter describes important examples of disparities based on race/ethnicity, gender or other individual factors, in diabetes measures that reflect IOM quality domains. Because the literature is dominated by studies of racial disparities and effectiveness of care measures, this review is focused on these aspects.
Numerous studies of disparities in diabetes effectiveness of care are based on race and ethnicity. Evidence shows that racial/ethnic minorities receive suboptimal quality of care across an array of diabetes quality measures, compared to Whites (Harris 1999; IOM 2002). Few studies examine quality of care disparities based on other factors, such as gender or socioeconomic status. Moreover, there are limited disparities studies based on age, particularly studies on age-based differences for midlife vs. older adults.
Effective care of diabetes includes receipt of appropriate process of care services, such as lipid and hemoglobin A1c (HbA1c) testing or control; preventive care (e.g., adult vaccinations); and eye and foot examinations. Reviews of effectiveness of care measures in diabetes have found numerous differences among non-White racial/ethnic minorities compared to Whites (AHRQ 2007; Harris 1999; IOM 2002; Peek et al 2007). There is some evidence that these disparities persist even after accounting for health care access and insurance. For example, Puerto Rican adults with diabetes in New York City are less likely than Whites to receive annual HbA1c testing, cholesterol testing, hypertensive medications and pneumococcal vaccinations, despite having equal access to health care, as measured by insurance, medical home and physician visits (Hosler and Melnik 2005; Peek et al 2007). Similarly, in the 2008 Healthcare Disparities Report, AHRQ found that from 2002–2004, Hispanics consistently lagged behind Whites in receipt of recommended diabetes services, including HbA1c testing, eye examination and foot examination. Findings from a study of Medicare managed care plans also demonstrated that plans with larger numbers of racial minority patients have lower effectiveness of care across several diabetes quality measures relative to plans with fewer minority patients (Schneider et al 2002; Trivedi et al 2005, 2006). Two meta-analyses examining data on HbA1c control in Black, Hispanic and White persons with diabetes found that Blacks and Hispanics exhibited consistently higher levels of HbA1c than Whites (Kirk
Disparities in diabetes effectiveness of care have been largely reported according to race/ethnicity, with relatively fewer reports examining disparities based on gender and socioeconomic status. In addition, disparities based on gender are not consistent in the literature. For example, a 2007 issue of the peer-reviewed journal, Women's Health Issues, dedicated to gender disparities on clinical performance measures in effectiveness of care for diabetes and cardiovascular disease, found that women sometimes received lower quality of diabetes care than men, but sometimes received about the same or better quality of diabetes care than men. One study from that issue found that older women with diabetes in Medicare had higher rates of eye examinations but lower rates of lipid screening and nephropathy monitoring than men (Bird et al 2007). Another study found that women <65 with diabetes had rates of cholesterol screening that were about the same as men, with a reported difference of only 1 percentage point (Chou et al 2007a). A consistent finding across all studies was that fewer women achieved recommended cholesterol control than men, both among women enrolled in commercial plans (primarily age <65) and among Medicare beneficiaries (primarily >65). Women who were of lower socioeconomic status or Black had added risks because of the combined effects of gender disparities due to race and socioeconomic status (NCQA 2007). As Table 2 shows, only 37.7 percent of women with diabetes in commercial managed care plans achieved recommended cholesterol control, compared to 43.3 percent of men (Chou et al, 2007b), with a similar disparity for women with diabetes in Medicare (38.5 percent) and men (45.7 percent).
Disparities in diabetes effectiveness of care based on socioeconomic status have also been demonstrated. Using 2001–2004 data, AHRQ reported that those with lower income and lower education were less likely to receive three recommended diabetes services (HbA1c testing, eye examinations and foot examinations) than individuals with higher income and more education (AHRQ 2007). Another study found that avoidable hospitalizations for diabetes decreased as income increased, although the study cautioned that other factors (e.g., quality of primary care, age, relationship with providers, patient self-management skills) could also influence rates (Correa-de-Araujo et al 2006).
Effective diabetes management includes screening for other conditions for which diabetics are at increased risk (e.g., cardiovascular disease). Evidence from the Framingham Heart Study indicates that the presence of diabetes is significantly associated with an increased risk of developing cardiovascular disease (Franco et al 2007). Research has found a lower likelihood of some screenings for these additional risks, including testing and treatment for hypertension and dyslipidemia among Blacks and Hispanics compared to Whites (Peek et al 2007). Review of self-monitoring of blood glucose studies found that while self-monitoring rates were low, they were generally lower among racial and ethnic minorities. English fluency had some influence on self-monitoring rates in some studies, but data were limited for Native Americans and Asian Americans (Kirk et al 2007).
Some studies have found evidence that racial disparities in diabetes effectiveness of care may be lessening among older adults. One study on Medicare managed care members using 1999–2003 data found that Black–White racial disparities in diabetes care were attenuated over time, including disparities in eye examinations, low-density lipoprotein (LDL) testing and control and HbA1c testing (Trivedi et al 2005). A study of Veterans Administration (VA) beneficiaries found that hospital care for mostly older male VA patients with diagnosed diabetes did not differ for Black, Hispanic or White patients (though differences were found among patients with diagnosed congestive heart failure and chronic obstructive pulmonary disease). Examples of process of hospital care include admission history that documented the patient's typical level of blood glucose control; glycosylated hemoglobin measured during the stay; and patients being ready for discharge when acceptable blood glucose control was established (Gordon et al 2003). Results from these studies should be interpreted with caution because male veterans and Medicare managed care beneficiaries are not representative of all midlife or older
Analogous to diabetes complications, data for disparities in diabetes effectiveness of care were rarely stratified by the midlife and older adults. A Medline review of human subject, English-language diabetes studies published in the last five years, using the broadest search terms possible (i.e., diabetes and disparities) and limited to studies that include the midlife (45–64 years), yielded 195 studies, of which only 6 could be determined conclusively to stratify results by midlife vs. older adults. Thus, even when applying the most productive search criteria, there are still considerable gaps in the literature with respect to studies that examine and clarify disparities in diabetes care by both midlife and older adults.
Among the few studies that did stratify results by these two age groups, there are conflicting results. One study of Hispanics with diabetes found that older adults (>65) were less likely to have an HbA1c test in the past year than midlife adults (40–64) (Mainous et al 2007). The midlife age range in this study started at the slightly younger age of 40 years, instead of 45 years. The studies reported in Table 2 do not show a consistent advantage for either age group, and the finding on gender disparity in cholesterol control is consistent across age groups.
Another study stratifying results by the two age groups found that higher socioeconomic status, as measured by higher education, had a protective effect against smoking (as measured by lower probability of smoking) among midlife adults (45–64), but not among older adults (>65) (Karter et al 2007). Yet another study found that given the presence of diabetes, lower extremity amputations were much lower among ages 50–64 than among ages 65–74 or >75 (Sambamoorthi et al 2006). Also in this study, the midlife age range did not exactly encompass 45–64, but the slightly older range of 50–64. Moreover, the older range was stratified into two age groups (65–74 and >75) rather than just one group of >65. The results of this study are not surprising, given that older people probably have lived with diabetes for longer, probably have other comorbidities and may be of frailer health overall, which puts them at higher risk for adverse outcomes, such as amputations.
Safe, Patient-Centered, Timely/Accessible, Efficient Care
Despite recommendations by the IOM, most diabetes quality of care measures and studies do not address the domains of safe, patient-centered, timely/accessible and efficient care. For example, the National Committee for Quality Assurance/American Diabetes Association (NCQA/ADA) Provider Recognition Program indices do not include patient-centered items related to self-management or psychosocial support (Glasgow et al 2008), and measures of care efficiency in general are lacking. Virtually no efficiency measures reviewed in a recent, comprehensive RAND study included a quality dimension in assessing output, and as such, most efficiency measures could more appropriately be termed as merely measures of cost rather than of efficiency (RAND 2008). NCQA has been among the few organizations that have developed and tested nationally-based efficiency measures assessing the relationship between quality of care outcomes relative to resource input required for that care. Its efficiency measures examine care for persons with diabetes, including annual HbA1c testing, LDL screening and eye examinations; and the receipt of medical care for nephropathy (Roski et al 2008). Resource use for those with diabetes was calculated using medical and pharmacy claims in 31 commercial health plans. Early testing results indicate that pharmacy resource use was significantly and positively associated with higher quality of diabetes care (i.e., plans that spent more on pharmacy services for members with diabetes had more favorable diabetes care results), and hold promise for the future development and use of efficiency measures for diabetes care.