In this US prospective study of persons with hypertension or diabetes and eGFR >
60 mL/min/1.73 m2
at baseline, Blacks and Whites at risk of CKD incidence experienced differences in both the presence and quality of their health care, with Blacks more likely to lack both health insurance and a usual source of care and Blacks utilizing different usual sources of health care compared with Whites. Blacks had nearly 60% excess risk of CKD incidence compared with Whites after adjustment for age, sex, study centre and baseline eGFR. Adjustment for socioeconomic, lifestyle and clinical factors explained 64% of the race disparity in CKD incidence, but the disparity was not completely attenuated until models were additionally adjusted for race differences in access to health care. These findings suggest that whilst socioeconomic, lifestyle and clinical factors explain the greatest proportion of the disparity, access to health care may also play a role in ethnic/race disparities in CKD incidence among persons at high risk for CKD incidence and progression.
Our findings are among the first to provide prospective evidence of the role of access to health care in explaining racial disparities in incident CKD above and beyond socioeconomic, lifestyle and clinical factors and have implications for future efforts to narrow disparities in patients with CKD risk factors. Whilst several studies have explored possible explanations for Black–White disparities in the development of ESRD [2–5,38,39
], few have explored potential explanations for race disparities in the incidence of earlier stages of CKD. A previous study found that access to health care explained some of the disparity in less advanced CKD, but this study limited the definition of access to health care as only the presence or absence of health insurance [6
]. Our study incorporated a broader definition of access to care, which is supported by research demonstrating that the presence of a usual source of care is associated with improved continuity, comprehensiveness and timely receipt of health care [17,40
Greater access to health care has been associated with improved treatment and control of hypertension and diabetes, a potential mechanism through which better access to health care could prevent the incidence of CKD. Previous research has demonstrated that differences in types of insurance (Medicaid vs
private insurance) and in the source of health care (e.g. private physicians’ office or emergency rooms) may lead to differences in clinical services rendered [41–46
] as well as health outcomes [47–49
]. Moy et al
. showed that, among persons with hypertension, those without a usual source of health care or health insurance were less likely to receive ‘screening, follow-up care or pharmacologic treatment for hypertension’ [10
]. Studies in individuals with diabetes have demonstrated similar relationships between access to health care and improved glycemic monitoring and control [11,13–15, 50–52
]. Our own analyses, demonstrating a strong association between our measure of access to health care with treatment and control of hypertension and diabetes at baseline, support this hypothesis.
Both race differences in the presence of access to health care as well as race differences in the quality of access may have contributed to our findings of race disparities in CKD incidence. We found Blacks were less likely than Whites to have health insurance or a usual source of health care and that the distribution of types of usual care differed between Blacks and Whites. Whilst the total proportion of Blacks and Whites with no usual source of health care was small, differences in the distribution of the type of usual source of health care reported by Blacks and Whites were significant. Therefore, it is possible that differences in quality of care were more influential than the absence of usual care itself on disparities in CKD incidence. For example, Blacks were less likely than Whites to report seeing a private physician and more likely to use HMOs, even among those with a usual source of health care. These results are consistent with studies of access to health care in other clinical areas [16,18–28
]. We also found that Blacks were less likely than Whites to have health insurance, although data on types of health insurance were not collected in the ARIC Study, precluding a more in-depth review of the effects of differences in insurance type (e.g. public vs
private). Further studies are needed to determine what types of differences in health insurance (e.g. prescription medication coverage, co-pays) or usual source of care (e.g. continuous relationships with a single primary care provider vs
more episodic care, characteristic of health clinics) most strongly impact the clinical outcomes of patients at risk of CKD incidence.
Limitations of this study deserve mention. First, it is possible that ARIC participants at high risk of CKD incidence were not representative of persons at high risk of CKD in the US general population. Access to health care among ARIC participants differed from recent population estimates (7 and 3% of Black and White ARIC participants with no usual source of health care vs
14 and 13% of Blacks and Whites in the US general population in 2007; 23 and 4% of Blacks and Whites with no health insurance in ARIC vs
17 and 11% of Blacks and Whites in the US general population in 2007) [53,54
]. These differences may reflect temporal changes in health care occurring after ARIC baseline data was collected and may limit the generalizability of our findings. Further research will be needed to assess the impact of recent changes in US health insurance policy on race disparities in CKD incidence [55
Second, Blacks were recruited exclusively from Jackson, MS, a rural, southern district where access to health care may be less adequate than other more urban regions. Although we attempted to adjust for differences in participants enrolled at different ARIC Study sites using standard statistical methods, it is possible that we could not completely eliminate confounding effects of geographic differences in access to care, quality of care or other social determinants of health. Such unmeasured and uncontrolled confounding has the potential to overestimate the race disparities in CKD and could explain the higher rates of poor access to care among Blacks in this study compared with national averages. It is also possible that participant's level of access to health care may have changed during follow-up, possibly introducing misclassification bias. Third, we considered participants reporting using pharmacists, regular health clinics and family members in the health care field as having a usual source of health care, but the content of this care is unclear. Fourth, the ARIC Study did not collect data on urine protein at baseline; some participants with kidney damage and normal eGFR could have been misclassified as not having CKD. Furthermore, incident CKD defined by eGFR was detectable only at Visit 4, resulting in stepped data and possibly an underestimate of CKD incidence if cases were more likely to drop out of the study or die before their follow-up visit. Also, CKD hospitalizations and deaths may represent more advanced cases of CKD compared with cases defined by eGFR ≤ 60 mL/min/1.73 m2
and variations in physicians' subjective judgments regarding coding practice could introduce misclassification bias. However, this composite outcome has been used in several prior studies [39,56–58
] and results of an analysis incorporating more limited definitions of CKD incidence were consistent with our main findings. Notwithstanding these limitations, our study provides important evidence regarding influence of access to health care in explaining Blacks’ excess CKD incidence.