Racial and ethnic disparities in ED use among older adults with asthma and primary care nurse practitioner work environments

Background Nurse practitioners (NPs) increasingly deliver primary care in the United States. Yet, poor working conditions strain NP care. We examined whether racial/ethnic health disparities in ED visits among older adults with asthma are moderated by primary care NP work environments. Methods Survey data on NP work environments in six states were collected from 1,244 NPs in 2018–2019. 2018 Medicare claims data from 46,658 patients with asthma was merged with survey data to assess the associations of all-cause and ambulatory care sensitive conditions (ACSC) ED visits with NP work environment and race/ethnicity using logistic regression. Results NP work environment moderated the association of race (Black patients versus White patients) with all-cause (odds ratio [OR]: 0.91; p-value = 0.045) and ACSC (OR: 0.90; p-value = 0.033) ED visits. Conclusions Disparities in ED visits between Black and White patients with asthma decrease when these patients receive care in care clinics with favorable NP work environments.


Introduction
Over 4 million adults age 65 and older in the United States (U.S.) have an asthma diagnosis. 1Asthma is often undertreated and underdiagnosed in older adults, partly because its symptoms can be confused with other chronic conditions common in older adults, such as chronic obstructive pulmonary disease, interstitial lung disease, eosinophilic bronchitis, heart failure, bronchiectasis, and sarcoidosis. 2,3In addition, the physiological and psychological effects of aging can complicate the diagnosis, treatment, and management of asthma among older adults. 4,5Asthma is often successfully managed by primary care clinicians 6 ; however, when patients lack access to available or high-quality primary care, they may experience asthma exacerbations and visit the emergency department (ED) for care. 7Older adults with asthma frequently visit the ED, at a rate of 18.6 visits per 10,000 population. 8This outcome is costly for the patient and the healthcare system and could be avoided via high-quality care in primary care settings. 7,9ithin the population of older adults with asthma in the U.S., individuals from certain racially and ethnically minoritized groups experience disproportionately higher asthma prevalence, morbidity, and mortality. 10For example, Black and Hispanic adults are twice as likely as White adults to report an asthma-related ED visit in the last 12 months. 113][14][15] Health care system interventions are needed to address these disparities. 15cess to high-quality primary care can help older adults from racially and ethnically diverse backgrounds improve asthma control and thus reduce asthma-related health disparities.An increasing proportion of primary care delivered in historically marginalized and primary care underserved communities is provided by nurse practitioners (NPs). 16,17Nurse Practitioners are advanced practice registered nurses with either a Master's or Doctoral degree. 18As licensed, independent clinicians, NPs can provide a wide range of health care services, including the diagnosis and management of diseases and health promotion and education.The NP workforce is experiencing rapid growth, reaching 385,000 NPs in 2023, with a projected 45% growth between 2022-2032. 19,20Research shows that NPs provide safe, highquality primary care, and the outcomes of their care are comparable to those of physicians. 17,21Yet NPs face many working challenges within their clinical settings, which limits their capacity to deliver care to their patients.The work environment of NPs, comprised of NP perceptions of working conditions in their organizations, plays a key role in NP delivery of care and outcomes. 22A favorable NP work environment consists of collegiality between NPs and physicians, clear communication between NPs and administrators, understanding of the NP capabilities within the practice, and support for the NP independent practice. 23In unfavorable work environments, NPs have limited administrative support, strained relations with staff, and inequitable access to clinic resources such as staff support. 24,25For example, staff support is often unavailable to NPs to perform patient care activities, such as in-o ce screening and diagnostic testing or lab draws 26 , activities that are essential to achieving and sustaining asthma control.
Unfavorable NP work environments are also associated with a lower quality of patient care, a lack of patient-centered care, and increased hospitalizations and ED utilization. 25,27The NP Health Disparities model, which shows the relationship between various community, policy, and organizational factors affecting NPs and health disparities, posits that an unfavorable work environment exacerbates health disparities by limiting NP's ability to deliver high-quality care to patients from historically marginalized communities in the U.S.. 28 Yet, it is unknown how the NP work environment contributes to disparities among older adults with asthma.As the population of older adults in the U.S. is becoming more racially and ethnically diverse and the costly burden of chronic conditions increases, 29,30 evidence on the impact of the NP work environment on reducing racial health disparities in unnecessary acute care utilization in historically marginalized communities is needed.The purpose of this study was to examine whether racial and ethnic health disparities in ED visits among older adults with asthma are moderated by the NP work environment in primary care practices.

Design
We used a cross-sectional design to collect survey data on work environments from NPs in primary care practices in six U.S. states in 2018-2019.The survey data was merged with 2018 Medicare claims data for patients from the same practices.This study received approval from the Institutional Review Board of [removed for review] Medical Center.Below, we provide an overview of the methods presented in detail and published elsewhere. 25,31udy Setting We conducted the study in six geographically diverse states across the US: Arizona (AZ), California (CA), New Jersey (NJ), Pennsylvania (PA), Texas (TX), and Washington (WA).Each state in the U.S. has policies determining the care and services NPs can provide to their patients. 32These state-level scope of practice regulations determining NPs' autonomy level in delivering patient care vary from state to state. 32s practicing in AZ and WA had a full scope of practice, they could treat patients and prescribe medications independently; in NJ and PA, they had a reduced scope of practice, NPs needed to collaborate with physicians to deliver care; and in CA and TX they had a restricted scope of practice, NPs required physician supervision to practice.

Study Sample Primary Care Practices
We de ned primary care practices as those in which 50% or more of the physicians had family practice, general practice, geriatrics, internal medicine, preventative medicine, or pediatrics specialties.We then selected primary care practices that employed at least one NP for selection in our study.

Nurse Practitioner Sample
To identify NPs in these primary care practices, we used IQVIA's OneKey database, which contains information on ambulatory-based clinicians in the U.S., including clinician and practice names, practice locations, contact information, and National Provider Identi cation (NPI) numbers. 33We requested the full sample of primary care NPs in AZ, NJ, and WA, a 75% random sample of NPs in PA, and a 50% random sample in CA and TX, given the difference in the NP workforce size across states, with CA and TX having a large number of NPs.Using a modi ed Dillman process 34 , we sent three separate survey mailings and two postcard reminders to collect data.Each mailing included an online link to the survey, giving NPs a choice to complete the survey online or on paper.In total, 5,689 NPs met the sampling criteria, and 1,244 NPs returned completed surveys (a response rate of 21.9%).More information on the survey, response rate, and nonresponse bias is reported elsewhere. 31

Patient Sample
We rst obtained Medicare claims for all patients (n = 1,123,861) who received care in study practices in 2018.Medicare is a federal insurance plan that offers insurance coverage to all adults 65 years and older.
Patients were attributed to the practices where they received care using a common approach 35 in which they were rst attributed to clinicians (NPs and physicians) by NPI and then to a speci c practice using the practice identi er from the OneKey database.Patients without a dominant clinician were excluded from the analysis (n = 642).We then selected only patients ages ≥ 65 who had been diagnosed with asthma based on the Centers for Medicare & Medicaid Services Chronic Condition Warehouse algorithm. 36Patients younger than 65 may also receive Medicare for certain qualifying reasons, including disability.Our nal sample included 46,658 patients with asthma attributed to 917 primary care practices, for which we had both NP surveys and patient data.

Variables Key Explanatory Variables
We measure the explanatory variable, race and ethnicity, using the Research Triangle Institute (RTI) race and ethnicity coding algorithm, which included seven categories: African American or Black (Black), Asian or Paci c Islander (Asian), Hispanic or Latinx (Hispanic), non-Hispanic White (White), American Indian/Native American, Unknown, and Other.This algorithm improves the sensitivity of identifying patients from various racial and ethnic groups compared to the Medicare race variable alone. 37s completed the Nurse Practitioner-Primary Care Organizational Climate Questionnaire (NP-PCOCQ)-a reliable and valid measure of the NP work environment. 23The NP-PCOCQ contains 29 items asking NPs how much they agree that certain work conditions are present in their practices on a 4-point Likert-like scale from "strongly agree" to "strongly disagree".The tool has four subscales measuring critical work environment domains: NP-Administration Relations (NP-AR), NP-Physician Relations (NP-PR), Independent Practice and Support (IPS), and Professional Visibility (PV). 23For example, the NP-AR subscale measures if practice managers take NP concerns seriously or make efforts to improve their working conditions, and the IPS subscale measures NPs' ability to practice independently and have support for their practice.The NP-PCOCQ and its subscales have strong psychometric properties with acceptable internal consistency reliability and construct, discriminant, and predictive validity. 23,38 rst computed individual-level mean scores on each NP-PCOCQ subscale.Then, we calculated practice-level mean scores by aggregating the responses of all NPs within each practice since the work environment is a property of the organization, not the individual NP. 39A global measure of practice-level work environment was computed by averaging scores of the four NP-PCOCQ subscales; higher scores indicate a more favorable work environment.We created a standardized work environment score for the regression models with an average score of 0 and a standard deviation of 1.

Outcome Variables
We measured two outcomes: all-cause and Ambulatory Care Sensitive Conditions (ACSC) ED visits, coded categorically (zero events or 1 or more events).Using Medicare Part B claims data, we identi ed all-cause ED visits as any visit for Healthcare Common Procedure Coding System codes 99281, 99282, 99283, 99284, and 99285. 40ACSC ED visits were unique ED visits with evidence of being avoidable or primary care treatable according to the "New York University ED Algorithm". 41For each ED visit, the algorithm assigns a probability, based on the primary International Classi cation of Diseases (ICD) 10-CM diagnosis, that the visit is in one of ve categories: 1-Non-Emergent; 2-Emergent, Primary Care Treatable; 3-Emergent, ED Care Needed, Preventable/Avoidable; 4-Emergent, ED Care Needed, Not Preventable/Avoidable; 5-All other.ED visits were counted as ACSC if the probability of belonging in the rst three categories was greater than 0, based on the principal ICD-10-CM diagnosis from the Medicare Part B claims data.

Covariates
We included patient demographic information (i.e., age, sex) and the type of patients' dominant primary care clinician (i.e., physician, NP) as patient-level covariates.In addition, because asthma and chronic obstructive pulmonary disease (COPD) are heterogeneous conditions that often overlap in older adults, 42 we included having COPD separately, as well as the number of 13 chronic conditions from the U.S. Department of Health and Human Services list of standard chronic conditions (i.e., hypertension, congestive heart failure, coronary artery disease, cardiac arrhythmias, hyperlipidemia, stroke, arthritis, cancer, chronic kidney disease, dementia, depression, diabetes, osteoporosis). 43To t the most parsimonious model, we counted chronic conditions.Additionally, the count of conditions is a better predictor of total Medicare expenditures than the cumulative duration of chronic conditions. 44 the organizational level, we controlled for the number of NPs and physicians in practice, practice type (i.e., physician o ces, hospital-based clinics, community health centers, other), location (rural vs. urban), and practices' structural capability index measuring the structural attributes (i.e., availability of the electronic health record, disease registries, weekend hours, performance feedback to clinicians, disease registries and reminder systems, community referrals, shared communication with patients) associated with the delivery of high-quality care. 45Fixed-effects for each state were included to account for differences at the state level, including scope of practice variation.Though missing data was < 5% of the survey data, we assessed patterns of missingness and found the missingness to be independent of NPs' demographic attributes.Thus, case-wise deletion was used.There was no missing data in the Master Bene ciary Summary File, from which patients' demographic information was extracted.

Data Analysis
We computed descriptive statistics for all patient-level and organizational-level characteristics.Bivariate associations between patient-level characteristics (i.e., age, dominant primary care provider type, having COPD, number of other chronic conditions, and sex) and race/ethnicity (i.e., White, Hispanic, Black, Asian, and all other) were calculated using analysis of variance or chi-square tests.Frequencies of outcomes (i.e., all-cause and ACSC ED visits) by race and ethnicity group were also calculated.Finally, we used mixed-effect logistic regression models to assess the associations of outcomes with NP work environment and race and ethnicity, controlling for patient-and organizational-level factors as described above.An interaction term between race and ethnicity and work environment score was created to assess whether the work environment moderated racial and ethnic disparities.Next, we estimated the odds ratio of each outcome at different levels of NP work environment score (from − 2 to a maximum score of 1.60) from the nal model, including the signi cant interaction effect to further examine how work environment impacted racial and ethnic disparities in ED and ASCS ED visits.To account for the clustering effect of 46,658 patients nested in 917 practices, we used random effect models and adopted a two-sided α level of .05.Our organizational-level sample size of 917 practices is greater than the recommended sample size of 50 at the second level to run random effect models. 46Our sample size allows us to detect a small effect size with 80% power.All analyses were conducted in SAS, version 9.4 (SAS Institute, Inc., Cary, NC).

Characteristics of Patients
Overall, we included 46,658 patients age 65 and older who were diagnosed with asthma and had received care at one of the primary care practices in our study (Table 1).Patients had a mean age of 74.8 years (SD = 7.2), and 70.1% were women.Black patients had the highest average number of chronic conditions of all racial and ethnic groups, 3.7 conditions (SD = 2.2), and the highest proportion of COPD (39.3%).Asian patients were less likely to have an NP as their primary care clinician (10.4%) than all other race and ethnicity groups.1, Fig. 1).As the work environment scores increased, both White and Black patients with asthma had lower odds of all-cause or ACSC ED visits (Fig. 1).Compared to White patients with asthma, we found that the NP work environment did not moderate the associations between other race and ethnicity groups (i.e., Asian and Hispanic patients) and all-cause or ACSC ED visits.Overall, Asian patients were equally likely to have an all-cause or ACSC ED visit compared to White patients.Hispanic patients were more likely to experience both all-cause (OR 1.50; 95% CI 1.38-1.62;p-value < 0.001) and ACSC (OR 1.63; 95% CI 1.50-1.78;p-value < 0.001) ED visits compared to White patients.

Discussion
In this study, we examined whether NP work environments in primary care practices affect racial and ethnic disparities in ED visits for older adults with asthma.Our ndings show a signi cant disparity; one in four White patients with asthma had at least one all-cause ED visit, while almost half of the Black patients with asthma (48%) had an all-cause ED visit.Similarly, the rates were higher for Black and Hispanic patients for ACSC ED visits, deemed preventable with high-quality primary care.Our ndings also show that disparities in ED visits between Black and White patients with asthma decrease when these patients receive care in primary care practices with favorable NP work environments.
Results are consistent with previous ndings that, overall, better NP work environments improve quality outcomes, enhance patient-centeredness, and reduce hospitalizations and ED utilization. 25,27Our study ndings, however, add that not all patient groups experience this bene t-and in fact, Black patients are disproportionately more likely to experience improved outcomes when the NP work environment is more favorable.Over the years, many interventions to reduce racial and ethnic disparities have been tested in the U.S. 47,48 , with an overall lack of robust effect.The results of this study suggest that improving the work environments of NPs in primary care may be an effective tool to address disparities in asthma outcomes.
Preventing unnecessary ED visits among older adults with asthma is a likely bene t of favorable NP work environments.NPs practicing in favorable work environments may be more likely to provide evidencebased care to their patients with asthma.In such environments, NPs also have adequate support and resources, favorable collegial relationships, and optimal use of their advanced skills. 25,27Thus, supporting favorable NP work environments may help patients receive high-quality care and prevent ED visits.
Also, asthma control can be improved when clinicians and patients engage in shared decision-making. 49is is important because other interventions to improve asthma control, such as coaching 50 , have not been effective.NPs may be particularly well-suited to be engaged in shared decision-making as their education typically includes training in therapeutic communication strategies and patient engagement.
Also, NP practice is characterized as more holistic in its orientation to care, allowing NPs to develop trusting relationships with their patients as they design treatment plans to accommodate the care that patients want and value.Favorable work environments may promote NPs' ability to engage in shared decision-making, translating it into better patient outcomes.
Despite the positive attributes associated with NP care, signi cant barriers to asthma control persist.
Without access to high-quality care, community resources, and pharmacologic treatments, patients will likely continue to suffer from poor asthma control that places them at risk for reduced quality of life, future acute health care utilization, and even death.While increased healthcare coverage remains out of reach for many Americans, NPs can have an immediate effect by delivering high-quality, patient-centered care to the communities that need it most, given that their numbers are growing rapidly.Our ndings present critically important policy and practice implications in reducing racial and ethnic healthcare disparities.As the NP workforce grows, they are vital to delivering care in communities underserved by primary care and helping narrow the health disparity gap.Yet, despite the importance of the NP work environment, many NPs report practicing in challenging environments that do not provide adequate time and resources for patient care.Investing in work environments to allow NPs to deliver high quality care is critically needed for the goal of eliminating racial and ethnic health care disparities in the U.S.
Our ndings also may have international implications.The healthcare workforce globally has undergone major changes in recent decades triggered by a growing prevalence of chronic conditions, shortage of primary care providers, and advancement of nursing education. 51,52As a result, countries increasingly rely on nurses with advanced roles, such as NPs. 53In many countries, policies are being designed to shift tasks from physicians to advanced practice nurses as an effective strategy to increase primary care capacity, meet the demand for care, and achieve the goals of universal health coverage. 54Thus, creating favorable work environments for this growing workforce should be a priority to achieve these goals.
Our study has limitations.We collected data in six U.S. states with a variable NP scope of practice regulations, so the ndings may not be generalizable nationwide.Future studies should be expanded nationally.Our ndings may also not be generalizable internationally given the different demographic distribution and healthcare system characteristics in the U.S. Future international studies should be conducted, given the rapid growth of NPs internationally.Some patients with asthma in the study also likely had COPD.While we adjusted separately for COPD, it can be challenging to distinguish clinically between COPD and asthma in older adult populations, given that the two conditions are heterogeneous airway diseases that often overlap. 55We relied on self-reported data from NPs, which are subject to selfreport bias.Our about 22% response rate is low yet comparable to other clinician survey responses. 56herefore, non-response bias might be an issue, yet responders were not signi cantly different from nonresponders, and the demographic characteristics of our sample were comparable to those in the national NP sample. 31Our study used a cross-sectional, observational design, and determining causation is not possible.While this study produces important insights on the potential impact of NP work environments on the outcomes of patients with asthma, more research is needed to understand better the aspects of the work environment that can enhance the quality of asthma care delivery.Furthermore, intervention studies should be designed to determine the impact of work environment on ED use among patients with asthma.

Conclusion
This is the rst study that investigated the association between NP work environment in primary care practices and racial and ethnic health disparities in ED use among older adults with asthma in the U.S.
Merging unique survey and patient data, we found that the NP work environment moderates the relationship between race and ethnicity and ED use.When patients receive care in practices with poor NP work environments, Black older adults are even more likely to have higher rates of all-cause and ACSC ED visits compared to White patients.Our ndings have key implications for practice, policy, and research innovations to enhance the NP work environment in primary care practices and achieve health equity.

Figures
Figures

Table 1
Patient Characteristics by Race and Ethnicity Groups a Includes: American Indian/Alaskan Native, Unknown Race and Ethnicity, and Other Race and Ethnicity Characteristics of Primary Care Practices

Table 2
total, 16,478 patients with asthma (35.3%) visited the ED, and 11,644 patients with asthma (25%) had an ACSC ED visit (Table1).Black and Hispanic patients were more likely than White patients to both have an all-cause (Black 48.1%; Hispanic 44%; White 33.9%) and ACSC ED visits (Black 38.3%; Hispanic 33.6%; ED visits among Medicare patients with asthma (Table3).Greater standardized work environment scores were associated with lower odds of all-cause and ACSC ED visits between Black and White patients (Supplementary Table1).When the standardized NP work environment measure reached its maximum score of 1.60-indicating the most favorable environment, the odds ratio for all-cause and ACSC ED visits for Black patients compared to White patients were the lowest (all-cause ED visits 1.42; 95% con dence interval [CI] = 1.18-1.70;p-value = 0.002 and ACSC ED visits 1.57; 95% CI 1.29-1.90;pvalue < 0.001).Conversely, when the standardized work environment measure reached its lowest score of NP = nurse practitionerAll-Cause and ACSC ED Visits In -2, the odds of all-cause and ACSC ED visits for Black patients compared to White patients were their highest (all-cause ED visits 1.97; 95% CI 1.64-2.36;p-value < 0.001 and ACSC ED visits 2.25; 95% CI 1.87-2.72;p-value < .001)(Supplementary Table

Table 3
Multi-level Regression Models Assessing the Moderation Effect of NP Work Environment on Associations of Race and Ethnicity with All-Cause and ACSC ED Visits among Medicare Patients with Asthma (n = 46,658)