Abstract
Approximately 70% of American Indian/Alaska Native (AI/AN) individuals reside in urban areas. Urban Indian Health Organizations (UIHOs) provide culturally engaged primary care for AI/AN patients and members of other racial and ethnic groups who have experienced disparities in diabetes and hypertension care, and are commonly affected by social and economic barriers to care. We assessed whether disparities were present between the racial and ethnic groups served by the largest UIHO in the USA. We developed retrospective cohorts of patients with hypertension or diabetes receiving primary care from this UIHO, measuring differences between AI/AN, Spanish-preferring Latinx, English-preferring Latinx, Black, and White patients in mean systolic blood pressure (SBP) and mean hemoglobin A1c (A1c) as primary outcomes. To assess processes of care, we also compared visit intensity, missed visits, and medication treatment intensity in regression models adjusted for sociodemographic and clinical characteristics. For hypertension (n = 2148), adjusted mean SBP ranged from 135.8 mm Hg among Whites to 141.3 mm Hg among Blacks (p = 0.06). For diabetes (n = 1211), adjusted A1c ranged from 7.7% among English-preferring Latinx to 8.7% among Blacks (p = 0.38). Care processes for both hypertension and diabetes varied across groups. No group consistently received lower-quality care. This UIHO provided care of comparable quality for hypertension and diabetes among urban-dwelling AI/ANs and members of other racial, ethnic, and language preference groups. Systematic assessments of care quality in UIHOs may help demonstrate the importance of their role in providing care and improve the quality of care.
Keywords: American Indians, Diabetes, Healthcare disparities, Hypertension, Quality of care
Introduction
American Indians and Alaska Natives (AI/ANs), along with members of other racial and ethnic minority groups, bear a disproportionate burden of chronic illness in the USA [1]. Disparities in care and outcomes for hypertension and diabetes between AI/ANs and White Americans have been well documented [2, 3]. Proposed strategies to eliminate these disparities include the provision of culturally engaged health care, improving access to care, and addressing the influence of social determinants of health on disease outcomes [4–6]. Strategies to provide culturally engaged care include the accessibility of racially and ethnically congruent staff and clinicians, language concordance, and provider and staff training in cultural humility, as well as culturally tailored and responsive care delivery programs. Chronic disease quality improvement strategies such as better integration and use of health technology, team-based care, use of evidence-based care guidelines, and alignment with the Chronic Care Model, an organization-wide, population-based approach to primary care for chronic illness, can also help to improve care [1, 6, 7].
Urban-dwelling AI/ANs constitute 70% of the AI/AN population in the USA [8]. Across the country, 33 Urban Indian Health Organizations (UIHOs) are supported by the Indian Health Service (IHS) to provide care for urban AI/ANs, although UIHOs receive only 1% of the IHS annual budget to fulfill this vital role [9]. Approximately 45% of UIHOs are also designated as Federally Qualified Health Centers (FQHCs) that also provide primary care to members of other racial and ethnic groups, many of whom receive Medicaid insurance. An analysis of US vital statistics from 1990 to 2000 found that AI/ANs who lived within the service areas of UIHOs were twice as likely to be poor or unemployed, more likely to have inadequate prenatal care, and subject to higher mortality from multiple causes than the general population in the same geographic areas [5].
While the quality of care for hypertension and diabetes has been studied in predominantly rural IHS clinics and tribal health programs [10, 11], little is known about the quality of care for these health conditions in UIHOs. One study using manual chart reviews of data from 2002 found few differences in the quality of diabetes care between UIHO and rural Indian health sites [12]. More current information about quality and equity of care from registry-based cohorts of individuals who receive care for hypertension and diabetes in UIHOs would be helpful in promoting quality improvement strategies within this critical component of the care delivery system for urban AI/ANs.
In preparation for a randomized controlled trial of an m-health intervention to improve visit-keeping, adherence and blood pressure (BP) control in First Nations Community HealthSource (FNCH) in Albuquerque, NM, the largest UIHO in the USA [13], we developed registries of all patients with hypertension or diabetes, or both, using data from the FNCH electronic health record. The purpose of this paper is to assess whether disparities in hypertension and diabetes outcomes and processes of care were present between the largest racial and ethnic groups served by FNCH.
Methods
Study Setting
The study took place at the First Nations Community HealthSource (FNCH), a nonprofit UIHO in Albuquerque, NM, founded in 1972. FNCH is authorized through Title V of PL 94–437 of the Indian Health Care Improvement Act to serve the health care needs of urban AI/ANs. Designated by the Health Resources and Services Administration (HRSA) as a Federally Qualified Health Center in 1997, FNCH also provides primary medical care to other socially disadvantaged residents of Albuquerque and surrounding areas [14]. Approximately 40% of FNCH patients are AI/AN; Diné (Navajo) is the most frequent tribal affiliation. Many other FNCH patients are undocumented immigrants from Mexico and Central America. Most patients speak English or Spanish as their primary language. Homelessness, food insecurity, and other social risk factors are common [15, 16].
FNCH patients are served by a bilingual staff of primary care physicians, nurse practitioners (NPs), physician assistants (PAs), on-site behavioral health clinicians, pharmacy practitioners, and an audiologist with the support of public health nurses, diabetes educators, and patient navigators. During the study period, the primary care medical staff (physicians, NPs, and PAs) consisted of 24 White clinicians, 4 AI/AN clinicians, and 1 Black clinician. (The sole Black clinician was AI and Black.) Ten of these clinicians were Latinx, and 17 were non-Latinx. Of 10 NPs, 7 were certified as family NPs, and 3 were adult NPs. The physician staff included 7 family medicine doctors, 1 internist, and 1 psychiatrist. FNCH also provides dental care, traditional healing, meals, and other social support services to address social needs and health disparities. In November 2015, services were further enhanced by the addition of an on-site pharmacy. FNCH clinicians have used the eClinicalWorks (Westborough, MA) electronic health record (EHR) since 2012.
Study Design and Participants
We conducted two concurrent retrospective cohort studies to assess the quality of care and identify disparities in care. To identify FNCH patients eligible for the hypertension cohort or the diabetes cohort, we included all patients who met criteria for either diagnosis using EHR data. Patients with coexisting hypertension and diabetes were included in both cohorts. These registries included FNCH patients who received primary care at FNCH between August 26, 2014 (the earliest date that allowed a 365-day look-back period for prior information in the FNCH electronic health record), and April 2, 2019. For each patient, the time in the study included the period between the first and last clinical encounters within that time range.
The hypertension cohort was developed in 2017 to facilitate recruitment into a randomized clinical trial (RCT) of an automated, text message-based intervention to improve hypertension control. The protocol for the trial and its main results have been published [13, 17]. Since the RCT found that the intervention was not effective in reducing blood pressure, improving medication adherence, or reducing missed primary care visits at FNCH, the 295 participants in the RCT are included in the hypertension cohort for the current analysis. The diabetes cohort was developed in 2018–2019 for the current analysis.
FNCH patients with hypertension were identified using a combination of diagnostic codes (ICD-9 codes 401 or 405, or ICD-10 codes I10, I13, I15, or R03), orders for antihypertensive medications, and elevated blood pressures recorded in the EHR [13]. Since FNCH patients continued to fill antihypertensive medications at outside pharmacies, we used medication orders, rather than pharmacy fills, to identify these medications. We identified individuals with hypertension as those who (1) had at least 2 visits with a hypertension diagnosis on different days; (2) had a single visit with a hypertension diagnosis and an antihypertensive medication order; (3) had a single visit with a hypertension diagnosis and 1 or more elevated blood pressure determinations (defined as blood pressure ≥ 140/90 mm Hg); or (4) had 2 consecutive elevated blood pressure determinations at visits on different days.
FNCH patients were identified as having diabetes using criteria adapted from the SUPREME-DM study [18]. These criteria were (1) two laboratory tests (fasting plasma glucose ≥ 126 mg/dl, random plasma glucose ≥ 200 mg/dl, or hemoglobin A1c ≥ 6.5%) on different days within a 2-year time period; (2) two orders for diabetes medications on different days within a 2-year period; (3) two ICD-9 codes (250, 357.2, 366.41, 362.01, 362.02, 362.03, 362.04, 362.05, 362.06, and 362.07) or ICD-10 codes (E10, E11) within a 2-year period on different days; or (4) any other combination of laboratory test, medication order, or diabetes diagnosis within a 2-year period. FNCH records did not differentiate between fasting and random measurements of serum glucose. Patients with only diagnostic codes for gestational diabetes mellitus were excluded, and their laboratory tests, diagnoses, and medication orders were removed from consideration during their pregnancy.
These broad criteria allowed inclusion of patients who received care intermittently from FNCH, either because they moved between the urban setting and reservation-based health care systems for AI/ANs, or because they were immigrants who received care in their country of origin as well as the USA. Individuals were excluded from either cohort if they received no primary care services from FNCH. Specific clinical encounters were excluded if they were not for primary care visits or were coded as cancellations.
Study Measures
The primary outcome to assess the quality of care for hypertension was the mean systolic blood pressure (SBP), calculated from readings obtained during primary care and other visits at FNCH during the period of participation in the cohort. Diastolic blood pressure (DBP) and hypertension control, defined by both less stringent (140/90 mm Hg) and more stringent (130/80 mm Hg) American Heart Association (AHA) criteria, were secondary outcomes [19]. For diabetes, the primary outcome measure was the mean hemoglobin A1c, assessed using all available laboratory results during the period of participation in the cohort. The secondary outcome was the level of diabetes control, defined as well controlled (A1c < 7.0%) or poorly controlled (A1c ≥ 9.0%).
The initial exposures of interest were race and ethnicity, based on self-identification from the FNCH EHR. We initially divided these responses into AI/AN, Latinx, non-Latinx Black, non-Latinx White, and other or missing. Since the group with other/missing race and ethnicity was small (6.0% of the hypertension cohort, 4.7% of the diabetes cohort), we excluded these individuals from each cohort. Because many FNCH patients were Spanish-preferring and the clinic staff had bilingual expertise to address their needs, we subdivided self-identified Latinx patients into English-preferring and Spanish-preferring subgroups, resulting in 5 subgroups in total for all analyses. Although the clinic staff did include individuals with bilingual expertise in Diné and English, too few patients preferred Diné to allow analysis of a Diné-preferring subgroup.
We assessed processes of care that might affect the primary and secondary outcomes in each cohort. For the hypertension cohort, process measures were as follows: (1) the number of visits with measurement of BP; (2) visit intensity, measured as the number of visits divided by the duration of cohort membership; (3) the proportion of missed visits recorded in the EHR (excluding visits cancelled by the patient); and (4) the intensity of treatment, defined as the maximum number of antihypertensive medications ordered at any single visit. Process measures for diabetes care were similar, and included (1) the number of visits with a diabetes diagnostic code, a laboratory test of hemoglobin A1c, plasma glucose, collection of at least one creatinine lab or urinary albumin/creatinine ratio, or a medication order for diabetes; (2) the intensity of care, measured as the number of visits divided by the duration of cohort membership; (3) the proportion of missed visits (excluding cancellations) recorded in the EHR; (4) the intensity of treatment, defined as the maximal number of oral hypoglycemic medications ordered at a single clinic visit; and (5) a medication order for insulin at any clinic visit.
Other covariates obtained from the FNCH EHR included demographic characteristics (age, sex), the presence of cardiovascular disease or depression, defined by the presence of at least one ICD-9 or ICD-10 code; chronic kidney disease, defined by the presence of an estimated glomerular filtration rate less than 60 l/min/1.73m2; the use of homelessness services from FNCH; smoking history; and BMI [13]. We excluded biologically implausible values for all variables, but otherwise used the values recorded in the EHR.
Statistical Analysis
Data analyses were conducted between December 2020 and May 2021. All analyses were conducted using SAS Studio 3.8 (Copyright 2012–2018, SAS Institute Inc., Cary, NC). We used the non-parametric Kruskal–Wallis test for continuous variables to assess differences between the 5 racial and ethnic and language preference groups because many study variables were not normally distributed. We used Chisquare test for categorical variables. We used Bonferroni adjustments to account for multiple statistical comparisons, and thus avoid an overemphasis on isolated differences across multiple racial and ethnic groups and multiple quality measures.
We developed multivariable models to compare process and outcome measures between the five groups. Candidate variables for covariate adjustment in each model included age (44 years and younger, 45–64 years, and 65 years and over), gender, cardiovascular disease, depression, chronic kidney disease, homelessness services ever used, smoking history, and BMI (< 25.0, 25.0–29.9, and ≥ 30.0). Covariates with p-values < 0.05 were then removed by backward selection.
Since each cohort member could have multiple readings for continuous variables (SBP, DBP, hemoglobin A1c, and creatinine values), we used linear generalized models with repeated measures, assuming a normal distribution. For the dichotomous outcomes (proportion of cohort members with BP control, good A1c control, and poor A1c control), we assumed a binomial distribution. Hypertension and diabetes process measures were all aggregated to the patient level with the exception of creatinine values for diabetes. For the number of visits per year, we used a Poisson model adjusting for person years and over-dispersion. For intensity of visits, we used a linear model assuming a log-normal distribution. For missed visits, we modeled the number of visits divided by total visits, assuming a binomial distribution. Finally, for intensity of treatment, we assumed a Poisson distribution accounting for over-dispersion. For the diabetes cohort, we assessed creatinine values, taking repeated measures into account and assuming a normal distribution. We also assessed insulin use, creatinine value obtained, and albumin/creatinine ratio obtained at the patient level, assuming a binomial distribution.
Ethical Oversight
This study was approved by the institutional review boards of Kaiser Permanente Colorado, the Colorado Multiple Institutional Review Board, and the Indian Health Service National Institutional Review Board. A waiver of informed consent for the current analysis was granted as the size of the dataset made individual informed consent impractical.
Results
Characteristics of Participants
The hypertension and diabetes cohorts combined included 2635 FNCH patients who received primary care at FNCH between August 26, 2014, and April 2, 2019, for hypertension (1424, 54.0%) or diabetes (487, 18.5%), or both (724, 27.5%). Table 1 depicts demographic and co-morbidity characteristics of the diabetes and hypertension cohorts separately. AI/AN patients were the largest group in both the hypertension and diabetes cohorts (29.2% and 32.9%, respectively), while Black patients were the smallest group (6.1% and 3.5%, respectively). AI/AN patients were the youngest group in both cohorts, with a mean age of 48.1 years (SD 12.8) in the hypertension cohort and 48.2 years (SD 11.5) in the diabetes cohort. Among AI/ANs, 21.4% had used FNCH homelessness support services, which was substantially higher than any other racial, ethnic, or language preference group (1.9% in Spanish-preferring Latinx, 6.5% in English-preferring Latinx, 8.4% in Black, and 3.7% in White patients, p < 0.001).
Table 1.
Characteristics of participants in hypertension and diabetes cohorts, First Nations Community HealthSource, 2014–2019
| Characteristic | Total | Non-Latinx American Indian/Alaska Native | Spanish-preferring Latinx | English-preferring Latinx | Non-Latinx Black | Non-Latinx White | p-value* |
|---|---|---|---|---|---|---|---|
| Hypertension cohort | |||||||
| N (%) of cohort | 2148 (100.0) | 627 (29.2%) | 412 (19.2%) | 522 (24.3%) | 131 (6.1%) | 456 (21.2%) | N/A |
| Total days in cohort, mean (SD) | 742.7 (518.2) | 748.3 (499.4) | 827.0 (532.2) | 696.0 (518.3) | 712.8 (468.4) | 721.0 (536.7 | 0.002** |
| Age at first visit, mean (SD) | 52.4 (12.9) | 48.1 (12.8) | 55.3 (12.1) | 51.7 (13.0) | 52.7 (12.3) | 56.2 (12.0) | < 0.0001** |
| Female, N (%) | 1113 (51.8%) | 320 (51.0%) | 234 (56.8%) | 263 (50.4%) | 69 (52.7%) | 227 (49.8%) | 0.24 |
| Concurrent hypertension and diabetes, N (%) | 724 (33.7%) | 228 (36.4%) | 170 (41.3%) | 167 (32.0%) | 36 (27.5%) | 123 (27.0%) | < 0.0001** |
| Total number of primary care isits, mean (SD) | 10.6 (10.2) | 10.5 (10.8) | 9.9 (8.4) | 10.3 (10.3) | 9.8 (8.4) | 11.7 (11.3) | 0.45 |
| Cardiovascular disease, N (%) | 50 (2.3%) | 12 (1.9%) | 11 (2.7%) | 11 (2.1%) | 1 (0.8%) | 15 (3.3%) | 0.4 |
| Depression, N (%) | 209 (9.7%) | 59 (9.4%) | 29 (7.0%) | 62 (11.9%) | 10 (7.6%) | 49 (10.8%) | 0.12 |
| Chronic kidney disease, N (%) | 141 (6.6%) | 23 (3.7%) | 27 (6.6%) | 35 (6.7%) | 15 (11.5%) | 41 (9.0%) | 0.001** |
| Smoking-related diagnosis or visit, N (%) | 173 (8.1%) | 27 (4.3%) | 20 (4.9%) | 43 (8.2%) | 16 (12.2%) | 67 (14.7%) | < 0.0001** |
| Body mass index***, mean (SD) | 32.2 (13.0) | 32.5 (7.4) | 31.1 (5.6) | 33.0 (8.2) | 31.3 (8.0) | 32.3 (24.4) | 0.0002** |
| Use of FNCH homelessness services, N (%) | 204 (9.5%) | 134 (21.4%) | 8 (1.9%) | 34 (6.5%) | 11 (8.4%) | 17 (3.7%) | < 0.0001** |
| Diabetes cohort | |||||||
| N (%) of cohort | 1211 (100.0) | 399 (32.9%) | 329 (27.2%) | 276 (22.8%) | 42 (3.5%) | 165 (13.6%) | N/A |
| Total days in cohort, mean (SD) | 726.2 (511.8) | 711.7 (504.1) | 810.2 (520.2) | 685.0 (495.6) | 573.1 (481.0) | 701.3 (530.1) | 0.004** |
| Age at first visit, mean (SD) | 51.3 (12.1) | 48.2 (11.5) | 52.0 (11.3) | 51.2 (13.6) | 55.1 (12.1) | 56.5 (10.4) | < 0.0001** |
| Female, N (%) | 676 (55.8%) | 252 (63.2%) | 186 (56.5%) | 135 (48.9%) | 19 (45.2%) | 84 (50.9%) | 0.002** |
| Concurrent diabetes and hypertension, N (%) | 724 (59.8%) | 228 (57.1%) | 170 (51.7%) | 167 (60.5%) | 36 (85.7%) | 123 (74.6%) | < 0.0001** |
| Total number of primary care visits, mean (SD) | 10.7 (10.3) | 10.7 (10.7) | 10.1 (8.3) | 10.6 (10.4) | 10.8 (11.8) | 12.2 (11.8) | 0.69 |
| Cardiovascular disease, N (%) | 30 (2.5%) | 9 (2.3%) | 8 (2.4%) | 9 (3.3%) | 0 (0.0%) | 4 (2.4%) | 0.76 |
| Depression, N (%) | 123 (10.2%) | 44 (11.0%) | 27 (8.2%) | 30 (10.9%) | 1 (2.4%) | 21 (12.7%) | 0.21 |
| Chronic kidney disease, N (%) | 75 (6.2%) | 15 (3.8%) | 18 (5.5%) | 21 (7.6%) | 6 (14.3%) | 15 (9.1%) | 0.01 |
| Smoking related diagnosis or visit, N (%) | 76 (6.3%) | 15 (3.8%) | 12 (3.7%) | 24 (8.7%) | 4 (9.5%) | 21 (12.7%) | 0.0001** |
| Body mass index, mean (SD) | 32.7 (7.3) | 33.5 (7.5) | 30.9 (5.8) | 33.7 (8.0) | 34.6 (8.9) | 32.5 (7.0) | < 0.0001** |
| Use of FNCH homelessness services, N (%) | 86 (7.1%) | 57 (14.3%) | 6 (1.8%) | 14 (5.1%) | 4 (9.5%) | 5 (3.0%) | < 0.0001** |
SD, standard deviation
5-group comparisons: Kruskal–Wallis tests for age, number of PC visits, BMI, and total time in cohort. Chi-square tests for categorical/dichotomous variables
Remains significant at 0.05 level after adjustment for multiple comparisons
Body mass index available for 2105 of 2148 participants (98.0%) in hypertension cohort and for 1188 of 1211 (98.1%) of diabetes cohort
Hypertension Process and Outcomes of Care
Table 2 depicts hypertension process and outcome measures across racial and ethnic groups, after adjustment for covariates. Mean adjusted SBP, our primary hypertension outcome, ranged from 135.8 mm Hg (SD 16.7) in White patients to 141.3 (SD 18.8) in Black patients (p = 0.06 for the 5-group comparison). Hypertension control, defined as BP < 140/90 mm Hg, ranged from 61.8% among Spanish-preferring Latinx to 50.3% among Black patients (p = 0.02 for the 5-group comparison).
Table 2.
Comparisons of process and outcome measures for hypertension between racial and ethnic groups
| Model | Non-Latinx American Indian/Alaska Native (N = 627) | Spanish-preferring Latinx (N = 412) | English-preferring Latinx (N = 522) | Non-Latinx Black (N = 131) | Non-Latinx White (N = 456) | p-value |
|---|---|---|---|---|---|---|
| Process of care measures | ||||||
| Adjusteda number of visits/year with BP measurement, mean (SD) | 4.5 (3.4) | 4.0 (2.9) | 4.7 (3.5) | 4.5 (3.2) | 5.1 (3.6) | < 0.0001* |
| Adjustedb intensity of visits, mean (SD) | 6.5 (6.5) | 6.4 (6.5) | 7.6 (6.3) | 6.2 (6.2) | 8.3 (6.4) | < 0.0001* |
| Adjustedc proportion of missed visits (%) | 12.60% | 10.80% | 10.80% | 12.20% | 7.10% | < 0.0001* |
| Adjusteda maximum number of antihypertensive medications, mean (SD) | 1.9 (1.7) | 2.4 (1.8) | 2.2 (1.7) | 2.6 (1.8) | 2.3 (1.8) | < 0.0001* |
| Outcome measures | ||||||
| Adjustedd,e systolic BP, mean (SD) | 135.9 (17.1) | 137.3 (20.8) | 136.1 (15.8) | 141.3 (18.8) | 135.8 (16.7) | 0.06 |
| Adjustedd,e diastolic BP, mean (SD) | 80.9 (11.6) | 78.3 (9.2) | 80.7 (10.1) | 86.4 (13.2) | 81.4 (12.0) | < 0.0001* |
| Adjustedc,e,f % with controlled BP of 140/90 | 60.00% | 61.80% | 59.90% | 50.30% | 60.00% | 0.02 |
| Adjustedc,e,f % with controlled BP of 130/80 | 30.30% | 35.30% | 29.80% | 23.70% | 29.30% | 0.0005* |
SD, standard deviation
Notes: Intensity of visits defined as N of visits divided by duration of time in cohort (in years)
Generalized linear model assuming a Poisson distribution accounting for over-dispersion
Generalized linear model assuming a log-normal distribution
Generalized linear model assuming a binomial distribution
Generalized linear model assuming a normal distribution
Model accounts for repeated measures
Since model is adjusting for covariates (from Table 1), counts are not presented since they will not be whole numbers when multiplying by the percent
Results remain significant after Bonferroni adjustment
All process of care measures differed across the 5 study groups. For example, visit intensity, measured as the number of visits divided by the duration of cohort membership, was highest for Whites (8.3, SD 6.4), and lowest for Non-Latinx Black patients (6.2, SD 6.2, p < 0.001 for the 5-group comparison). Intensity of treatment, defined as the maximum number of antihypertensive medications ordered at any single visit, was highest for Blacks (2.6, SD 1.8) and lowest for AI/ANs (1.9, SD 1.7, p < 0.001 for the 5-group comparison).
Diabetes Process and Outcomes of Care
Table 3 depicts diabetes outcome and process of care measures across the five analytic groups, after adjustment for covariates. Mean adjusted hemoglobin A1c, our primary diabetes outcome, ranged from 7.7 (SD 1.8) in English-preferring Latinx patients to 8.7 (SD 1.8) in Black patients (p = 0.38 for the 5-group comparison). Diabetes control, defined as hemoglobin A1c < 7.0%, ranged from a high of 42.2% among White patients to 34.2% among Black patients (p = 0.90 for the 5-group comparison). Intensity of visits ranged from 8.8, SD = 6.4, for Black patients to 6.3 adjusted, SD = 6.4, for Spanish-preferring Latinx patients (p = 0.002). The proportion of missed visits also varied across groups, ranging from 16.4% for Black patients to 7.9% for non-Latinx Whites (p < 0.001). None of the other differences in process of care measures for diabetes groups was significantly different.
Table 3.
Comparisons of process and outcome measures for diabetes between racial and ethnic groups
| Model | Non-Latinx American Indian/Alaska Native (N = 399) | Spanish-preferring Latinx (N = 329) | English-preferring Latinx (N = 276) | Non-Latinx Black (N = 42) | Non-Latinx White (N = 165) | p-value |
|---|---|---|---|---|---|---|
| Process measures | ||||||
| Adjusteda number of diabetes care visits/year, mean (SD) | 4.7 (4.0) | 4.1 (3.3) | 4.6 (1.9) | 6.1 (4.1) | 5.3 (3.9) | 0.0006 * |
| Adjustedb intensity of diabetes visits, mean (SD) | 7.5 (6.5) | 6.3 (6.4) | 7.5 (6.4) | 8.8 (6.4) | 8.5 (6.3) | 0.002 * |
| Adjustedc proportion of missed visits, % | 13.40% | 10.70% | 11.30% | 16.40% | 7.90% | < 0.0001 * |
| Adjusteda maximum number of oral hypoglycemic medications (SD) | 2.0 (1.3) | 2.0 (1.3) | 1.9 (1.3) | 1.7 (1.2) | 1.7 (1.2) | 0.12 |
| Adjustedc % of patients receiving insulin | 45.90% | 36.70% | 45.80% | 35.40% | 42.20% | 0.1 |
| Adjustedc proportion of participants with measurement of serum creatinine, % | 74.40% | 76.70% | 76.90% | 75.30% | 75.00% | 0.93 |
| Adjustedd,e serum creatinine among those with measurement, mean (SD) | 1.0 (0.7) | 1.1 (1.1) | 1.0 (0.7) | 1.0 (0.3) | 0.9 (1.9) | 0.63 |
| Adjustedc proportion of participants with measurement of urine albumin/creatinine, % | 42.90% | 47.30% | 36.60% | 47.50% | 43.30% | 0.11 |
| Outcome measures | ||||||
| Adjustedd,e hemoglobin A1c, mean (SD) | 8.0 (2.2) | 8.1 (2.3) | 7.7 (1.8) | 8.7 (1.8) | 7.9 (1.8) | 0.38 |
| Adjustedc,e hemoglobin A1c < 7, % | 38.00% | 38.90% | 42.00% | 34.20% | 42.20% | 0.9 |
| Adjustedc,e hemoglobin A1c ≥ 9, % | 27.60% | 25.80% | 24.10% | 34.60% | 24.30% | 0.85 |
SD, standard deviation
Notes: Intensity of visits defined as N of visits divided by duration of time in cohort (in years)
Generalized linear model assuming a Poisson distribution accounting for over-dispersion
Generalized linear model assuming a log-normal distribution
Generalized linear model assuming a binomial distribution
Generalized linear model assuming a normal distribution
Model accounts for repeated measures
Results remain significant after Bonferroni adjustment
Discussion
In this study assessing disparities in hypertension and diabetes care across five groups differentiated by race, ethnicity, and language preference within a UIHO, we found that our primary outcome measures, adjusted mean SBP for hypertension and adjusted mean hemoglobin A1c for diabetes, were comparable across groups. We also assessed processes of care for hypertension and diabetes that could provide insight into differences in health outcomes or identify special concerns of specific racial, ethnic, or language groups. These relationships were not consistent across patient subpopulations within FNCH, however.
Despite a higher intensity of medication treatment, DBP was highest and stringent hypertension control (< 130/80 mm Hg) was least common for Black patients in FNCH. Visit intensity was highest and missed visits lowest in White patients, but mean SBP, DBP, and BP control for Whites were comparable to those of AI/AN, Spanish-preferring, and English-preferring Latinx patients. White patients with diabetes also had higher visit intensity and lower rates of missed visits than members of most other racial or ethnic groups, but these care processes were not associated with significantly better diabetes control. Process and outcome measures were comparable between Spanish-preferring and English-preferring Latinx patients for both hypertension and diabetes, which may be related to the high level of language concordance between patients and their clinicians [20–22].
Comparisons of Care Quality with Other Populations and Settings
Little current information exists about the content or quality of care for hypertension and diabetes in UIHOs nationally. One study used a manual chart review to compare 2002 data on diabetes process and outcome measures between urban UIHOs and predominantly rural IHS sites and found few differences between urban and rural sites that provided care to AI/ANs [12]. Although FNCH is the largest site in the UIHO network, comparisons between UIHOs across a broader array of clinical conditions would be helpful to characterize the quality of care for urban AI/ANs more fully.
Our findings for primary outcomes in this clinical setting contrast with some prior population-based studies that document persistent disparities in meeting hypertension and diabetes treatment goals in Latinx, especially Mexican American and Black populations [2, 3, 23–27]. Consistent with our findings, data from the National Health and Nutrition Examination Survey (NHANES) indicate that non-Latinx Blacks have higher odds of using multiple antihypertensive agents but lower odds of blood pressure control when compared with non-Latinx whites [23]. Less is known about hypertension or diabetes control among AI/ANs from national data sources. AI/AN individuals are not separately identified in the NHANES, for example. The Indian Health Service has reported outcomes for diabetes care that meet or exceed national benchmark standards [28]. Such strict adherence to diabetes care clinical guidelines by providers at FNCH may also account for the lack of racial and ethnic disparities in the care of patients with diabetes. Moreover, the emphasis on language concordance and culturally congruent staff and providers by FNCH may be another reason for the lack of language disparities in hypertension seen among their Latinx patients. This is supported in the literature on cultural competency in the care of racial and ethnic group patients with diabetes, particularly Spanish-preferring Latinx patients [21]. Faced with language barriers, Spanish-preferring Latinx patients with diabetes are more likely to have poorer glycemic control than English-preferring Latinx patients with diabetes who have language-concordant physicians [22]. One recent study reported comparable quality of diabetes care with no linguistic disparities for glycemic control among Latinx patients receiving care in community health centers, consistent with our findings [29]. Another recent study examining racial and ethnic disparities in hypertension management and outcomes at HRSA-funded health centers using 2014 Health Center Patient Survey data found lower odds of taking hypertension control medication for AI/ANs and the highest likelihood of hypertension-related emergency department (ED) visits for AI/ANs followed by Blacks [30]. The authors concluded the higher likelihood of ED visits among Blacks could reflect a higher severity of hypertension in this group [30]. This conclusion is consistent with high treatment intensity for hypertension among Black patients in our study.
Limitations
Our study had several limitations. All information reported in this paper was from a single UIHO site, and our findings may not be generalizable to other UIHOs. We cannot attribute these outcomes to specific health policies or quality improvement interventions, or to structural features of the FNCH clinic. We lacked information about the quality or experience of care from the patient perspective. We lacked information on most social risk factors other than prior use of homelessness services. Differences in these risk factors between racial, ethnic, and language preference groups might explain some of the differences that we observed. Medication nonadherence plays a critical role in controlling and treating hypertension [31], but we were unable to assess medication adherence since FNCH patients filled their medications at multiple pharmacies.
Even though we included all patients from comprehensive hypertension and diabetes cohorts at the largest UIHO in the USA, the study remained statistically underpowered to detect clinically important differences in hypertension and diabetes outcomes between racial, ethnic, and language preference groups. Although there are no universally accepted definitions of a clinically important difference in BP or hemoglobin A1c, differences of 5 mm Hg in SBP or DBP, and 0.5% differences in hemoglobin A1c, have commonly been used [32, 33]. The differences in SBP, DBP, and hemoglobin A1c between Black patients and the other racial, ethnic, and language preference groups in FNCH approached these thresholds. Although Black patients constituted only 6.1% of the hypertension cohort and 3.5% of the diabetes cohort, they may represent a particularly vulnerable group in this setting.
Conclusions
Both the methodology and the findings of this study have implications for future research and clinical quality improvement efforts. Given the near-universal use of EHRs in UIHOs, the development of hypertension and diabetes registries in UIHOs that use a consistent set of variables with standardized definitions may facilitate quality improvement initiatives within individual sites, and comparisons between them. The lack of consistent disparities in processes and outcomes of care across racial, ethnic, and language preference groups suggests that the FNCH approach to addressing social needs of their patients, cultural competence, and commitment to fair and equitable care may have minimized disparities within the clinic. Strategies implemented at FNCH supporting a patient-centered approach to culturally effective care could potentially be adapted by other federally qualified health centers serving a “majority minority” population in meeting the public health challenge of reducing racial and ethnic health disparities.
Funding
This research was funded by an American Heart Association Strategically Focused Research Network Grant to the Centers for American Indian and Alaska Native Health (CAIANH) at the University of Colorado Anschutz Medical Campus (15SFDRN25710168). Drs. Steiner and Manson received additional support from the Center for Diabetes Translation Research at CAIANH, funded by the National Institute for Diabetes and Digestive and Kidney Diseases (NIDDK; P30DK092923). Dr. Schroeder also received support from a career development award from NIDDK (1K23DK099237).
Declarations
Competing Interests Dr. Moore was a consultant with Novo Nordisk, Inc. The other authors have no relevant financial or non-financial interests to disclose. Ethics Approval This study was approved by the institutional review boards of Kaiser Permanente Colorado, the Colorado Multiple Institutional Review Board, and the Indian Health Service National Institutional Review Board. The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki. Informed Consent A waiver of informed consent was granted.
References
- 1.Kruse G, Lopez-Carmen VA, Jensen A, et al. The Indian Health Service and American Indian/Alaska Native Health Outcomes. Annu Rev Public Health. 2022;43:559–76. 10.1146/annurev-publhealth-052620-103633. [DOI] [PubMed] [Google Scholar]
- 2.Yoon SS, Carroll MD, Fryar CD. Hypertension prevalence and control among adults: United States, 2011–2014. NCHS Data Brief. 2015;220:1–8. [PubMed] [Google Scholar]
- 3.McWilliams JM, Meara E, Zaslavsky AM, et al. Differences in control of cardiovascular disease and diabetes by race, ethnicity, and education: U.S. trends from 1999 to 2006 and effects of Medicare coverage. Ann Intern Med. 2009;150(8):505–15. 10.7326/0003-4819-150-8-200904210-00005. [DOI] [PubMed] [Google Scholar]
- 4.Fiscella K, Sanders MR. Racial and ethnic disparities in the quality of health care. Annu Rev Public Health. 2016;37:375–94. 10.1146/annurev-publhealth-032315-021439 [DOI] [PubMed] [Google Scholar]
- 5.Castor ML, Smyser MS, Taualii MM, et al. A nationwide population-based study identifying health disparities between American Indians/Alaska Natives and the general populations living in select urban counties. Am J Public Health. 2006;96:1478–84. 10.2105/AJPH.2004.053942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sequist TD. Improving the health of the American Indian and Alaska Native population. JAMA. 2021;325(11):1035–6. 10.1001/jama.2021.0521. [DOI] [PubMed] [Google Scholar]
- 7.Carey RM, Muntner P, Bosworth HB, et al. Prevention and control of hypertension: JACC Health Promotion Series. J Am Coll Cardiol.2018;72(11):1278–93. 10.1016/j.jacc.2018.07.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Indian Health Service. Urban Indian Health Program [fact sheet]. Rockville, MD: Indian Health Service. https://.www.ihs.gov/newsroom/factsheets/uihp/. Published October 2018. Accessed March 29, 2021. [Google Scholar]
- 9.Urban Indian Health Institute. About Urban Indian Health Programs.https://.www.uihi.org/urban-indian-health/about-urbanindian-health-organizations/ Accessed April 12, 2022.
- 10.Narva A. Population health for CKD and diabetes: lessons from the Indian Health Service. Am J Kidney Dis. 2017;71:407–11. 10.1053/j.ajkd.2017.09.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.O’Connell J, Reid M, Rockell J, et al. Patient outcomes associated with utilization of education, case management, and advanced practice pharmacy services by American Indian and Alaska Native peoples with diabetes. Med Care. 2021;59:477–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Moore K, Roubideaux Y, Noonan C, et al. Measuring the quality of diabetes care in urban and rural Indian Health Programs. Ethn Dis. 2006;16:772–7. [PubMed] [Google Scholar]
- 13.Schroeder EB, Moore K, Manson SM, et al. An interactive voice response and text message intervention to improve blood pressure control among individuals with hypertension receiving care at an Urban Indian Health Organization: protocol and baseline characteristics of a pragmatic randomized controlled trial. JMIR Res Protoc. 2019;8(4):e11794. 10.2196/11794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wakefield M. Federally qualified health centers and related primary care workforce issues. JAMA. 2021;325(12):1145–6. 10.1001/jama.2021.1964. [DOI] [PubMed] [Google Scholar]
- 15.Xu S, Goodrich GK, Moore KR, Manson SM, Gottlieb LM, Hessler D, Schroeder EB, Steiner JF. Identifying relative changes in social risk factors: an analytic approach. Med Care. 2021;59(2):e9–15. 10.1097/MLR.0000000001441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Steiner JF, Goodrich GK, Moore KR, Manson SM, Gottlieb LM, Kelly C. Schroeder EB. An approach to selecting single or multiple social risk factors for clinic-based screening. J Gen Intern Med. 2022;37:703–5. 10.1007/s11606-021-06740-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schroeder EB, Moore KR, Manson SM, et al. A randomized clinical trial of an interactive voice response and text message intervention for individuals with hypertension. J Clin Hypertens (Greenwich). 2020;22(7):1228–38. 10.1111/jch.13909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Nichols GA, Schroeder EB, Karter AJ, et al. Trends in diabetes incidence among 7 million insured adults, 2006–2011: The SUPREME-DM project. Am J Epidemiol. 2015;181(1):32–9. 10.1093/aje/kwu255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS /APhA/ASH/ASPC/NMA/PCNA Guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Hypertension. 2018;71(6):1269–324. 10.1161/HYP.0000000000000066. [DOI] [PubMed] [Google Scholar]
- 20.Traylor AH, Schmittdiel JA, Uratsu CS, et al. Adherence to cardiovascular disease medications: does patient-provider race/ethnicity and language concordance matter? J Gen Intern Med. 2010;25(11):1172–7. 10.1007/s11606-010-1424-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fernandez A, Schillinger D, Warton EM, et al. Language barriers, physician-patient language concordance, and glycemic control among insured Latinos with diabetes: The Diabetes Study of Northern California (DISTANCE). J Gen Intern Med. 2011;26(2):170–6. 10.1007/s11606-010-1507-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Parker MM, Fernández A, Moffet HH, et al. Association of patient-physician language concordance and glycemic control for limited-English proficiency Latinos with type 2 diabetes. JAMA Intern Med. 2017;177(3):380–7. 10.1001/jamainternmed.2016.8648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gu Q, Burt VL, Dillon CF, et al. Trends in antihypertensive medication use and blood pressure control among United States adults with hypertension: The National Health And Nutrition Examination Survey, 2001 to 2010. Circulation. 2012;126(17):2105–14. 10.1161/CIRCULATIONAHA.112.096156. [DOI] [PubMed] [Google Scholar]
- 24.Casagrande SS, Fradkin JE, Saydah SH, et al. The prevalence of meeting A1C, blood pressure, and LDL goals among people with diabetes, 1988–2010. Diabetes Care. 2013;36(8):2271–9. 10.2337/dc12-2258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Muntner P, Hardy ST, Fine LJ, et al. Trends in blood pressure control among US adults with hypertension, 1999–2000 to 2017–2018. JAMA. 2020;324(12):1190–200. 10.1001/jama.2020.14545. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Chobufo MD, Gayam V, Soluny J, et al. Prevalence and control rates of hypertension in the USA: 2017–2018. Int J Cardiol Hypertens. 2020;6:100044. 10.1016/j.ijchy.2020.100044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lee W, Lloyd JT, Giuriceo K, et al. Systematic review and metaanalysis of patient race/ethnicity, socioeconomics, and quality for adult type 2 diabetes. Health Serv Res. 2020;55(5):741–72. 10.1111/1475-6773.13326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Moore K, Jiang L, Manson SM, et al. Case management to reduce cardiovascular disease risk in American Indians and Alaska Natives with diabetes: results from the Special Diabetes Program for Indians Healthy Heart Demonstration Project. Am J Public Health. 2014;104(11):e158–64. 10.2105/AJPH.2014.302108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Leung LB, Vargas-Bustamante A, Martinez AE, et al. Disparities in diabetes care quality by English language preference in community health centers. Health Serv Res. 2018;53(1):509–31. 10.1111/1475-6773.12590. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sripipatana A, Pourat N, Chen X, et al. Exploring racial/ethnic disparities in hypertension care among patients served by health centers in the United States. J Clin Hypertens (Greenwich). 2019;21(4):489–98. 10.1111/jch.13504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Munger MA, Van Tassell BW, LaFleur J. Medication nonadherence: an unrecognized cardiovascular risk factor. Med Gen Med. 2007;9(3):58. [PMC free article] [PubMed] [Google Scholar]
- 32.Saiz LC, Gorricho J, Garjón J, et al. Blood pressure targets for the treatment of people with hypertension and cardiovascular disease. Cochrane Database Syst Rev. 2020;9(9):CD010315. 10.1002/14651858.CD10315.pub4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Campbell L, Pepper T, Shipman K. HbA1c: a review of nonglycemic variables. J Clin Pathol. 2019;72:12–9. 10.1136/jclinpath-2017-204755. [DOI] [PubMed] [Google Scholar]