National Trends in the Association of Race and Ethnicity With Predialysis Nephrology Care in the United States From 2005 to 2015

This national registry study examines national trends in racial/ethnic disparities in the receipt of predialysis nephrology care at least 1 year before dialysis initiation among adults with end-stage kidney disease in the United States from 2005 to 2015.


Introduction
Receipt of nephrology care before the initiation of treatment for end-stage kidney disease (ESKD) is associated with improved patient survival and other key outcomes, including reduced hospitalizations and complications, increased quality of life, better preparation for dialysis, and greater likelihood of receiving a kidney transplant. [1][2][3] Primary care clinicians are generally the first to detect chronic kidney disease (CKD) and play key roles in deciding when to refer patients to nephrology care. 4 Current guidelines recommend referral to a nephrologist for patients with an estimated glomerular filtration rate of less than 30 mL/min/1.73 m 2 , severely increased albuminuria, rapid decline of the estimated glomerular filtration rate, hematuria, and/or uncontrolled complications of CKD, such as hypertension requiring 4 or more antihypertensive agents, anemia, or electrolyte abnormalities. 5 However, late referral to nephrology care remains common, with recent reports 6 suggesting that approximately one-third of patients with CKD receive nephrology care at least 12 months before the start of ESKD therapy.
Clinicians are also advised to refer their patients with CKD to a nephrologist if they are part of a demographic group known to experience more rapid kidney disease progression, such as young adults and racial/ethnic minorities. 4 However, prior studies [7][8][9][10][11] suggest that racial/ethnic disparities exist in timely receipt of nephrology care. Prakash et al 7 reported that Black patients with CKD are less likely to receive a timely referral to a nephrologist than White patients, and patients living in areas with larger populations of Black residents are less likely than those living in other areas. Non-White patients are also more likely to be referred to a nephrologist during later stages of CKD than their White counterparts. 7 Postulated reasons for these disparities include differences in insurance, socioeconomic status, educational attainment, provider biases, and geographic barriers. [7][8][9][10][11] The Healthy People 2020 (HP2020) initiative, coordinated by the US Department of Health and Human Services, provides a vision and strategy for improving the health of all US residents by setting priorities, identifying baseline data and 10-year targets for specific objectives, monitoring outcomes, and evaluating progress. 12 The HP2020 CKD-specific objectives were designed to monitor and assess efforts to reduce the long-term burden of kidney disease, increase lifespan, improve quality of life, and eliminate related health care disparities. 12 A key HP2020 CKD objective is to "increase the proportion of chronic kidney disease patients receiving care from a nephrologist at least 12 months before the start of renal replacement therapy." 12 In addition, 1 of the 4 overarching goals of HP2020 is to eliminate health care disparities. 12 As such, the primary goal of the present study was to examine national trends and assess whether racial/ethnic disparities in receipt of at least 12 months of predialysis nephrology care narrowed during the last decade in the United States.

Data Sources
This cross-sectional study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. 13 The study used data from the US Renal Data System (USRDS), a national data system that collects, analyzes, and distributes information about ESKD in the

Main Outcomes and Measures
The study population included adults (aged Ն18 years) who initiated maintenance dialysis in the United States from January 1, 2005, to December 31, 2015, as captured in the USRDS. The primary study outcome, receipt of at least 12 months of predialysis nephrology care, was defined as response of yes and a marked checkbox for greater than 12 months by clinician documentation on the ESRD Medical Evidence Report Form CMS 2728 in response to the question, "Prior to ESRD therapy, was patient under care of a nephrologist? If yes, answer: <6 months, 6-12 months, or >12 months." We limited the study population to adults initiating dialysis on or after January 1, 2005, because the USRDS did not ascertain receipt of predialysis nephrology care before 2005. Patients were excluded if they were younger than 18 years, if their ESRD Medical Evidence Report Form CMS 2728 was missing, or if data for race or ethnicity were missing. Patients designated as 1 of the fixed USRDS racial/ethnic categories of non-Hispanic White, non-Hispanic Black, non-Hispanic Asian, or Hispanic (any race) were included in the study. This study was not sufficiently powered to examine outcomes among patients identified as American Indian/Alaska Native, Pacific Islander, Middle Eastern/ Arabian, Indian (subcontinent), or other/unknown owing to the small population sizes in some study years. The date of the analysis being reported was April 17, 2020.

Descriptive Data Analysis
Clinical and demographic characteristics were stratified by patient race/ethnicity and year of dialysis initiation. Wilcoxon rank sum (for continuous variables) and χ 2 (for categorical variables) tests were performed to compare distributions and assess statistical significance.

Regression Models
Associations between race/ethnicity and receipt of at least 12 months of predialysis nephrology care were analyzed using both univariable (estimating crude odds ratios [ORs] with 95% CIs) and

Mediation Analyses
We performed subsequent multivariable logistic regression models to examine the extent to which racial/ethnic differences in the following potential mediators attenuated disparities in predialysis nephrology care: (1) comorbid medical conditions (ie, cancer, atherosclerotic heart disease, congestive heart failure, diabetes, hypertension, chronic obstructive pulmonary disease, or peripheral vascular disease), (2) health insurance type, and (3) type of dialysis modality and type of vascular access. To estimate the influence of these potential mediators on primary study outcomes, these factors were incrementally incorporated into the main multivariable logistic regression models (adjusting for age, sex, body mass index, and ESKD etiology).

Sensitivity Analyses
To test the robustness of our study findings, we repeated primary multivariable logistic regression models to assess racial/ethnic differences in receipt of predialysis nephrology care at any time before ESKD therapy. Finally, we repeated primary analyses using modified Poisson regression models (estimating adjusted risk ratios) to further test the robustness of our primary study findings.

Model Testing and Statistical Significance
Missing variable levels were modeled separately from known variable levels in regression models.
The robustness of estimates was tested by comparing results from an alternate modeling approach to handle missing data (multiple imputation), and inferences remained the same for the study outcomes of interest. Two-tailed P < .05 was considered statistically significant. All analyses were conducted using Stata, version 16/MP for Linux (StataCorp LLC). from crude models were similar to inferences from adjusted models ( Table 3). The full model is provided as eTable 1 in the Supplement. Inferences from additional multivariable logistic regression models examining racial/ethnic differences in receipt of predialysis care at any time before ESKD therapy were also similar to inferences from primary models (eTable 2 in the Supplement). The adjusted risk ratios estimated from modified Poisson regression models performed in sensitivity analyses are provided in eTable 3 in the Supplement.

Factors Associated With Racial/Ethnic Disparities in Receipt of Predialysis Nephrology Care
Exploratory mediation analysis findings from incremental multivariable logistic regression models (

Discussion
In this national registry study of more than 1 million adults who initiated maintenance dialysis treatment in the United States from 2005 to 2015, 31.1% of patients received at least 12 months of predialysis nephrology care. In addition, racial/ethnic disparities in receipt of at least 12 months of predialysis nephrology care did not substantially improve during the study period. Secondary study findings from exploratory mediation analyses suggest that racial/ethnic differences in health insurance type may be more strongly associated with slight attenuation of racial/ethnic disparities in predialysis nephrology care than differences in comorbid medical conditions, dialysis type, or vascular access type among Black and Hispanic patients.   Findings from our study suggest that national strategies designed to target racial/ethnic disparities are needed to achieve equity in access to predialysis nephrology care. Potential strategies may include national efforts to enhance collaborations between primary care providers and nephrologists, particularly for members of racial/ethnic minority groups. 4,14,15 For instance, the United Kingdom has successfully implemented a primary care-based CKD management program that has resulted in improved CKD care and education, dissemination of new and existing educational tools, and early education of primary care trainees. 15 Similar integrated care models could be applied to clinical settings in the United States; in particular, national strategies are needed to improve primary care clinicians' capacities to deliver optimal CKD care and comanagement of patients with nephrologists (eg, team-based decision support). Integrated care coordination will be vital to achieving the HP2020 12 goals of eliminating health care disparities and substantially increasing the proportion of patients with CKD receiving at least 12 months of predialysis care from a nephrologist.
We also found that Black and Hispanic patients are less likely to receive at least 12 months of predialysis nephrology care than White patients, independent of differences in clinical and demographic factors. Although prior studies reported that women are less likely to receive pre-ESKD nephrology care owing to their lower prevalence of CKD, slower progression, and better clinical outcomes of CKD, 9 we found that racial/ethnic disparities are associated with receipt of predialysis nephrology care by both male and female patients. We also found that racial/ethnic differences in health insurance type did not fully explain disparities in predialysis nephrology care. These novel findings build on and expand prior work examining the association of health insurance status with likelihood of receiving predialysis specialty care. [7][8][9][10][11] In a prior study, 10 authors concluded that patients in the US Department of Veterans Affairs system are more likely to receive pre-ESKD nephrology care than the general population outside the Veterans Affairs health care system because of their greater access to subspecialty care, use of electronic health records, case management, and integrated clinical guidelines for early recognition and management.

Strengths and Limitations
Our study has several strengths. The first is the ability to comprehensively analyze a decade of national data from the well-characterized population of adults who initiated treatment for ESKD in the United States. A second strength of the study is the ability to account for differences in many important characteristics that might confound or potentially mediate observed associations.
This study also had limitations, one of which is that the primary study outcome is subject to the accuracy of physician-provided retrospective data documented on the ESRD Medical Evidence Report Form CMS 2728, as well as the inability to further subcategorize the broad racial/ethnic categories available in the national USRDS registry (ie, Hispanic/Latino, Asian, Black/African American, and White). Another limitation is that the study was unable to account for individual-or household-level patient income. The future availability of individual patient or household income in national registries would allow us to better delineate the extent to which trends in socioeconomic measures may contribute to changes in predialysis care disparities over time.

Conclusions
In this national study of more than 1 million US adults with ESKD, racial and ethnic disparities in receipt of at least 12 months of predialysis nephrology care did not substantially improve from 2005 to 2015. These findings suggest that national strategies to address disparities in predialysis health care are needed. Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer:
The data reported herein have been supplied by the US Renal Data System. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the US Renal Data System or the US government.