Publication
Article
Author(s):
The proportion of allergists accepting Medicaid in the US varied significantly among and within states.
ABSTRACT
Objective: To determine the geographic variability of Medicaid acceptance among allergists in the US.
Study Design: Geospatial analysis predicted Medicaid acceptance across space, and a multivariable regression identified area-level population demographic variables associated with acceptance.
Methods: We used the National Plan & Provider Enumeration System database to identify allergists. Medicaid acceptance was determined from lists or search engines from state Medicaid offices and calls to provider offices. Spatial analysis was performed using the empirical Bayesian kriging tool. Multivariate logistic regression was used to identify county-level characteristics associated with provider Medicaid acceptance.
Results: Of 5694 allergists, 55.5% accepted Medicaid. Acceptance in each state ranged from 13% to 90%. Washington, Arizona, and the Northeast had lowest predicted proportion of both Medicaid acceptance and Medicaid acceptance per 10,000 enrollees. Overall, county-level characteristics were not associated with the likelihood of accepting Medicaid in multivariate analyses. Only the percentage of individuals living in poverty was associated with a higher likelihood of providers accepting Medicaid (OR, 1.245; 95% CI, 1.156-1.340; P < .001).
Conclusions: A barrier to accessing allergy-related health care is finding a provider who accepts a patient’s insurance, which is largely variable by state. Lack of access to allergy care likely affects health outcomes for children with prevalent atopic conditions such as food allergy.
Am J Manag Care. 2024;30(8):374-379. https://doi.org/10.37765/ajmc.2024.89588
Takeaway Points
Food allergy (FA) is a growing public health issue that affects 7.6% of children in the US. Allergy and immunology specialists are the primary group of clinicians who diagnose and manage FA, yet a known challenge for patients of lower socioeconomic status is finding a provider.
Food allergy (FA) is a growing public health issue that is estimated to affect 7.6% of children in the US.1 However, only 4.7% of US children have a confirmed FA diagnosis by any type of physician, and among Medicaid-enrolled children, this rate drops to 0.6%, suggestive of an underdiagnosis of FA in this population.2 Allergy and immunology specialists (hereafter allergists) are the primary group of clinicians who diagnose and manage FA. Previous research found that among children enrolled in Medicaid, those living in higher-poverty counties are less likely to visit an allergist for preventive measures, such as allergy testing or epinephrine autoinjectors, and more likely to have an FA-related emergency department (ED) visit.3 This highlights a difference in how families navigate the health care system and, therefore, FA-related health outcomes, based on the socioeconomic status in a geographic region.
A known challenge for patients of a lower socioeconomic status is finding a provider who accepts their insurance, creating distance and time barriers to accessing care.4 This may be attributed to the high provider refusal rate for Medicaid compared with Medicare and private insurance, which has been seen among general/family practice, pediatrics, general surgery, obstetrics and gynecology, and psychiatry.5-7 Reasons for the higher provider refusal rate for Medicaid-enrolled patients include a lower reimbursement rate and the presumed higher psychosocial and illness burden of Medicaid-insured patients.8 Although the refusal of Medicaid coverage is associated with higher rates of ED visits and delayed care overall, the impact of refusal on the receipt of FA-related care remains unknown.5
Furthermore, there is geographic variability of FA prevalence among states in the US, with lower prevalence associated with states having a lower urban index.9 It may be that rather than a lower prevalence of FA in more rural states, there are lower rates of FA diagnoses, possibly attributable to lower access to allergist providers. In fact, there is state variability in the prevalence of a physician-confirmed FA diagnosis among Medicaid-enrolled children, ranging from a low of 0.2% in Nevada to a high of 1.4% in Alaska.2
There is geographic variability in physicians that accept Medicaid among states in the US and within a state.10 Physicians who practice in metropolitan areas, defined by the US Office of Management and Budget’s Metropolitan Statistical Areas, and states with high Medicaid-to-Medicare fee ratios are correlated with a higher acceptance of Medicaid.6 Medicaid enrollees have a more difficult time scheduling appointments for specialty care, such as an allergist, compared with primary care.11 To our knowledge, there are no data on Medicaid acceptance by allergists. This article aims to identify the geographic variability of Medicaid acceptance by allergists in the US and its correlates, including urbanization and poverty, to assess the patterns of access to allergy care for Medicaid enrollees. We hypothesize large state-to-state geographic variability with lower allergist acceptance in rural counties.
METHODS
Data Collection
Individual allergists in 50 states and the District of Columbia (DC) were identified through the National Plan & Provider Enumeration System (NPPES) by taxonomy codes in October 2020. The NPPES database is updated daily with added or deactivated National Provider Identifiers (NPIs). NPPES relies on providers or someone acting on their behalf to update information in the database. Although Medicaid acceptance is included in the NPPES data, the identifier is optional and not updated regularly, rendering it inaccurate.
To determine Medicaid acceptance, we took a 3-pronged approach. First, state Medicaid offices were asked for electronic lists of providers who accept Medicaid. Next, if the office could not provide a list, its search engine was used to identify Medicaid-accepting providers by NPI number. Finally, if providers were unable to be searched by NPI number, direct calls to provider offices were made to confirm acceptance status. Phone numbers were found through the NPPES database and, if found to be inaccurate, were searched on Google. If a state had fewer than 30 providers or if a state-specific list or search engine was not found, every provider was called. For states with more than 30 providers and a state-specific list or search engine, every 10th provider was called. Calls were made to all providers for states with less than 50% congruence between Medicaid acceptance from calls and from a Medicaid office–provided list or a search engine. Calls and contacting state-specific Medicaid offices were completed between June 2021 and March 2023.
Providers were excluded from the final data set either due to confirmation that the provider was no longer practicing (eg, due to retirement, moving states) or when a provider was unable to be contacted (NPI deactivated or incorrect number). When a provider was excluded, the next NPI on the list was called for the states where every 10th provider was called.
Outcome Measures
The primary outcome was Medicaid acceptance. The “gold standard” for provider Medicaid acceptance reflected the 3-pronged approach, where the call result superseded all other results. For example, if a Medicaid office–provided list reported acceptance, but calling the provider resulted in no acceptance, our final input was no acceptance for that provider.
Statistical and Spatial Analysis
Geographic variability of Medicaid acceptance was represented by 2 maps: (1) the proportion of allergists who accept Medicaid and (2) the rate of allergists who accept Medicaid per 10,000 individuals enrolled. Using ArcGIS Pro 3.0.1 (Esri), a geographic information systems application, our data set with Medicaid acceptance by providers was joined with data on Medicaid enrollees from Area Health Resources Files (AHRF) 2019-2020 and a US Census county shapefile from the National Historical Geographic Information System by Federal Information Processing Standards (FIPS) codes.12 Once feature layers were projected, prediction trends were calculated using the empirical Bayesian kriging (EBK) tool, a geostatistical interpolation method that calculates unknown data locations between measured points through a process of subsetting and simulations.12,13 A predictive output surface type was selected and parameters were adjusted to provide the best-fit model.
A multivariate logistic regression identified demographic data associated with Medicaid provider acceptance. Five continuous county-level variables by FIPS codes were included from the AHRF data set (2019-2020), specifically, percentage Black/African American and percentage Hispanic/Latino in 2010, percentage of persons in poverty in 2018, and percentage of Medicaid enrollees in 2019. Each variable was scaled to represent a 5–percentage point increase. Rural/urban status was estimated using the rural-urban commuting area (RUCA) codes, created based on 2010 Census commuting data and 2019 zip codes.14 The RUCA code associated with a provider’s zip code was collapsed into a 3-tier categorical variable of urban, large town, or rural. ORs were calculated by exponentiating the maximum likelihood estimates of Medicaid acceptance for each of the covariates in the model. Regressions were performed using SAS 9.4 (SAS Institute Inc).
RESULTS
The number of allergists identified in the NPPES database was 6015; 321 of these were excluded due to the provider’s office confirming that the provider was no longer working at the office or unable to be contacted. Therefore, a total of 5694 providers remained in the data set, 23.2% of whom were called to confirm their Medicaid acceptance. In comparison, in 2021 the Association of American Medical Colleges reported a total of 5009 active allergists in the US.15 Every provider was called in 47% of the states (Table 1). Twelve states had a Medicaid office–provided list, and 16 states had a Medicaid office–provided search engine (Table 1). In the states where a Medicaid office–provided list or search engine was used, the average congruence between these acceptance results and results from calling every 10th provider on the list was 64%(eAppendix Table [available at ajmc.com]). Oregon and Washington had a congruence of less than 50%, so all providers were called (eAppendix Table).
A total of 55.5% of NPPES allergists were found to accept Medicaid. Thirteen of 51 states (50 states plus DC) were found to have less than or equal to a 50% acceptance rate (Table 2). The range of acceptance in each state was 13.4% to 89.5% (Table 2). The 3 states with the lowest numbers of allergy providers were Wyoming (n = 3), South Dakota (n = 6), and Vermont (n = 7) (Table 2). The 3 states with the highest rates of Medicaid acceptance were New Mexico (89.5%), Alaska (88.9%), and Oregon (85.0%), although the number of total allergy providers in these states ranged from 9 in Alaska to 19 in New Mexico to 40 in Oregon (Table 2). The states with the 3 lowest rates were New York (13.4%), Nevada (26.1%), and Washington and Wyoming (33.3%), although the total number of allergy providers in these states varied, with 500 in New York, 23 in Nevada, 126 in Washington, and 3 in Wyoming (Table 2).
The map of the proportion of allergists who accept Medicaid shows high predicted rates of acceptance in most of the West Coast and Mountain state regions and low predicted rates of acceptance in the Northeast region and Washington, Arizona, and Texas (Figure 1). In contrast, regions where predicted rates of allergists who accept Medicaid per 10,000 enrolled are highest include Virginia, the upper Midwest, and southeastern Idaho, and lowest include the Northeast, Washington, Arizona, and Florida (Figure 2).
There were 297 allergy providers who practice in rural neighborhoods in the US. There was no independent effect of rural or large town neighborhood status on the likelihood of accepting Medicaid (P = .7901) (Table 3). Each increase of 5 percentage points in the proportion of the population living below the federal poverty level was associated with a higher proportion of Medicaid-accepting providers (OR, 1.245; 95% CI, 1.156-1.340; P < .0001) (Table 3). Otherwise, no other variables were associated with provider Medicaid acceptance.
DISCUSSION
This study found that Medicaid is accepted by roughly half of the 5694 allergists in the US. Additionally, Medicaid was accepted by less than 50% of allergists in 13 of the 51 states and DC. The estimated prevalence of FA in children is 8%, and the number of children with FA enrolled in Medicaid is 3 million. Approximately 3 million Medicaid-enrolled, food-allergic children would need to see one of the few Medicaid-accepting allergists to diagnose and manage their condition, which may explain the significantly lower prevalence of 1% physician-diagnosed FA among Medicaid-enrolled children compared with 8% in the general pediatric population.
Not only is access to Medicaid-accepting allergists low, but our data also show large state-to-state variability in Medicaid acceptance. Although rural neighborhood status was not significantly associated with Medicaid acceptance in our multivariate models, we found only 297 allergy providers in these areas. Lower access to physicians in rural communities has been demonstrated when surveying physicians of all specialties in the US.6 Previous work based on Medicaid claims also found a lower odds of FA prevalence in rural vs urban areas.9 Our data confirm a disparity in access to allergy care for Medicaid enrollees based on location within and among states.
Medicaid enrollees struggling to access a Medicaid-accepting provider has been explained by many factors, most commonly Medicaid’s lower reimbursement rates and the long interval of time before reimbursements are received.5,7,8 The availability of specialists in the health system influences whether that specialty can provide Medicaid-accepting providers, as a scarcer resource is most accessible to patients without public insurance.8 Our rate of allergists’ Medicaid acceptance was lower than that previously found for pediatricians (78%), general surgeons (88%), and obstetricians/gynecologists (81%) and more than that found among psychiatrists (36%) and orthopedic surgeons (31%).7,16 In fact, the average Medicaid reimbursements for the same evaluation and management codes for mental health and substance use disorders were found to be less for psychiatrists compared with primary care physicians, which may explain the low rates of Medicaid acceptance by psychiatrists.17,18 The consequence of having fewer providers in a specialty, such as psychiatrists and allergists, is that fewer providers are willing to accommodate slow and low reimbursements, creating a disparity in access to specialty care in these fields for Medicaid-insured patients. As a result, individuals from a lower income stratum, some of whom have Medicaid, are more likely to acquire care from primary care providers rather than from specialists for allergies and other atopic conditions.19,20 However, primary care providers often do not diagnose or manage FA by performing skin testing or oral food challenges and will refer patients to allergy specialists, whom Medicaid-insured patients struggle to access.
One solution to improve access to allergy-related care among Medicaid-insured patients would be to increase the training and guidelines that primary care physicians and nurse practitioners receive in allergy-related health care so that there are more providers available to diagnose and manage FA. Partnerships can be made between primary care providers and allergists to collaborate on the diagnosis, management, and treatment of FA to streamline allergy-related health care. Such initiatives have been implemented in asthma-related care that have improved the diagnostic accuracy and management of asthma by primary care providers.21 The target for interventions would be the regions where Medicaid-accepting allergists are scarcest, such as rural communities, which have also been targeted for asthma-related care.22 In such rural communities, providers can also implement telehealth to reduce the access disparity based on proximity to a provider’s clinic.
Our analytic methods can be applied to other specialties to prove areas of deficit in the supply of providers for a given demand of Medicaid-insured patients, incentivizing states and national specialty organizations to bring more specialists to these areas. Additional projects should identify and address other barriers to accessing allergy-related care among Medicaid-insured patients, such as provider quotas of Medicaid-insured patients in a patient panel, longer wait times for appointments, or needing a primary care provider referral to schedule an allergist appointment. For example, policies should be created that enforce parity in reimbursement rates and efficiency between Medicaid and private insurance for allergy health care to incentivize more allergists to accept Medicaid.5,8,23
Limitations
Our geostatistical analysis predicted values from a limited number of known measured values, which assumes similarity between nearby observations. One limitation to this analysis is that values surrounded by clusters of significantly different values may display as higher or lower in the final visualization.11 For instance, our map predicted Nevada to have a higher proportion of acceptance than we calculated due to few data points and high neighboring values. Another limitation was the delay between obtaining the list of NPPES providers and when Medicaid offices were contacted and variation in how frequently the Medicaid offices updated data. We compensated by calling provider offices directly and removing providers from our data set whom we were unable to contact. The phone calls themselves represent another limitation as we were forced to assume that the information received through the one phone call was the truth. Lastly, hospital catchment and outreach often extend beyond a county, so our conclusions based on a county-level measure may not encompass patients’ experiences beyond their county.
CONCLUSIONS
A barrier to accessing allergy-related health care is the limited number of providers who accept Medicaid, which varies greatly within and among states. Such health inequity may be contributing to more Medicaid-insured patients seeking allergy care from primary care providers, who may be less equipped to provide specialty diagnosis and treatment, such as skin testing, oral challenges, etc.24 This difference may contribute to the established differences in health-related FA outcomes based on socioeconomic status, such as more FA-related ED visits and costs in low-income households compared with high-income households and lower FA diagnoses among Medicaid-enrolled children compared with all children in the US.2,5,19 Expanding the number of Medicaid-accepting providers who can deliver allergy-related health care, while also addressing other barriers the Medicaid population experiences to obtaining high-value care through additional research and policy changes, are likely to improve disparities in FA health outcomes.
Author Affiliations: Center for Food Allergy and Asthma Research, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine (FOH, CZ, SRN, RSG, LAB), Chicago, IL; Geospatial and Data Services, Northwestern University Libraries (MF), Evanston, IL; Division of Allergy & Immunology (SRN) and Division of Academic General Pediatrics (RSG), Ann and Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL.
Source of Funding: Thermo Fisher Scientific.
Author Disclosures: Dr Nimmagadda receives research grant support from the National Institutes of Health and Food Allergy Research and Education (FARE). Dr Gupta reports serving as a medical consultant/advisor for Genentech, Novartis, Aimmune LLC, Allergenis LLC, and FARE; receiving research support from the National Institutes of Health (R21 ID #AI135705, R01 ID #AI130348, U01 ID #AI138907), FARE, Sunshine Charitable Foundation, Novartis, and Genentech; and having an ownership interest in Yobee Care Inc. Dr Bilaver receives research grant support from the National Institutes of Health, Thermo Fisher Scientific, FARE, Genentech, National Confectioners Association, Novartis, Yobee Care Inc, and Before Brands Inc. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (FOH, RSG, LAB); acquisition of data (FOH, CZ, MF, LAB); analysis and interpretation of data (FOH, CZ, MF, SRN, LAB); drafting of the manuscript (FOH, CZ, SRN, RSG, LAB); critical revision of the manuscript for important intellectual content (FOH, SRN, RSG, LAB); statistical analysis (CZ, MF); administrative, technical, or logistic support (MF, SRN); and supervision (SRN).
Address Correspondence to: Lucy A. Bilaver, PhD, Center for Food Allergy and Asthma Research, Northwestern University Feinberg School of Medicine, 750 N Lake Shore Dr, Floor 6, Room 676, Chicago, IL 60611. Email: l-bilaver@northwestern.edu.
REFERENCES
1. Gupta RS, Warren CM, Smith BM, et al. The public health impact of parent-reported childhood food allergies in the United States. Pediatrics. 2018;142(6):e20181235. doi:10.1542/peds.2018-1235
2. Bilaver LA, Kanaley MK, Fierstein JL, Gupta RS. Prevalence and correlates of food allergy among Medicaid-enrolled United States children. Acad Pediatr. 2021;21(1):84-92. doi:10.1016/j.acap.2020.03.005
3. Kanaley MK, Dyer AA, Negris OR, et al. Guideline-informed care among Medicaid-enrolled children with food allergy. Am J Manag Care. 2020;26(12):505-512. doi:10.37765/ajmc.2020.88538
4. Arpey NC, Gaglioti AH, Rosenbaum ME. How socioeconomic status affects patient perceptions of health care: a qualitative study. J Prim Care Comm Health. 2017;8(3):169-175. doi:10.1177/2150131917697439
5. Bhandari N, Shi Y, Jung K. Patient experience of provider refusal of Medicaid coverage and its implications. J Health Care Poor Underserved. 2016;27(2):479-494. doi:10.1353/hpu.2016.0096
6. Decker SL. In 2011 nearly one-third of physicians said they would not accept new Medicaid patients, but rising fees may help. Health Aff (Millwood). 2012;31(8):1673-1679. doi:10.1377/hlthaff.2012.0294
7. Holgash K, Heberlein M. Physician acceptance of new Medicaid patients: what matters and what doesn’t. Health Affairs Forefront. April 10, 2019. Accessed July 30, 2023. https://www.healthaffairs.org/content/forefront/physician-acceptance-new-medicaid-patients-matters-and-doesn-t
8. Tipirneni R, Kieffer EC, Ayanian JZ, et al. Factors influencing primary care providers’ decisions to accept new Medicaid patients under Michigan’s Medicaid expansion. Am J Manag Care. 2019;25(3):120-127.
9. Gupta RS, Springston EE, Smith B, Warrier MR, Pongracic J, Holl JL. Geographic variability of childhood food allergy in the United States. Clin Pediatr (Phila). 2012;51(9):856-861. doi:10.1177/0009922812448526
10. Tipirneni R, Rhodes KV, Hayward RA, et al. Geographic variation in Medicaid acceptance across Michigan primary care practices in the era of the Affordable Care Act. Med Care Res Rev. 2018;75(5):633-650. doi:10.1177/1077558717697750
11. Hsiang WR, Lukasiewicz A, Gentry M, et al. Medicaid patients have greater difficulty scheduling health care appointments compared with private insurance patients: a meta-analysis. Inquiry. 2019;56:46958019838118. doi:10.1177/0046958019838118
12. What is empirical Bayesian kriging? ArcGIS Pro. Accessed May 25, 2023. https://pro.arcgis.com/en/pro-app/latest/help/analysis/geostatistical-analyst/what-is-empirical-bayesian-kriging-.htm
13. Krivoruchko K. Empirical Bayesian kriging. Esri ArcUser. 2012. Accessed May 25, 2023. https://www.esri.com/news/arcuser/1012/empirical-byesian-kriging.html
14. Rural-urban commuting area codes. US Department of Agriculture. Updated September 25, 2023. Accessed March 22, 2023. https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/
15. Physician Specialty Data Report. Association of American Medical Colleges. Accessed July 5, 2024.
https://www.aamc.org/data-reports/workforce/data/number-people-active-physician-specialty-2021
16. Shi W, Anastasio A, Guisse NF, et al. Impact of insurance and practice type on access to orthopaedic sports medicine. Orthop J Sports Med. 2020;8(7):2325967120933696. doi:10.1177/2325967120933696
17. Carlo AD, Basu A, Unutzer J, Jordan N. Acceptance of insurance by psychiatrists and other physicians, 2007-2016. Psychiatr Serv. 2024;75(1):25-31. doi:10.1176/appi.ps.202100669
18. Sterling S, Weisner C, Hinman A, Parthasarathy S. Access to treatment for adolescents with substance use and co-occurring disorders: challenges and opportunities. J Am Acad Child Adolesc Psychiatry. 2010;49(7):637-646. doi:10.1016/j.jaac.2010.03.019
19. Hurst K, Gerdts J, Simons E, Abrams EM, Protudjer JLP. Social and financial impacts of food allergy on the economically disadvantaged and advantaged families: a qualitative interview study. Ann Allergy Asthma Immunol. 2021;127(2):243-248. doi:10.1016/j.anai.2021.04.020
20. Siegfried EC, Paller AS, Mina-Osorio P, et al. Effects of variations in access to care for children with atopic dermatitis. BMC Dermatol. 2020;20(1):24. doi:10.1186/s12895-020-00114-x
21. Gilchrist FJ, Carroll WD, Clayton S, et al. Implementation of a primary care asthma management quality improvement programme across 68 general practice sites. NPJ Prim Care Respir Med. 2023;33(1):21. doi:10.1038/s41533-023-00341-y
22. Bender BG, Simmons B, Konkoly N, Liu AH. The Asthma Toolkit Bootcamp to improve rural primary care for pediatric asthma. J Allergy Clin Immunol Pract. 2021;9(8):3091-3097.e1. doi:10.1016/j.jaip.2021.03.058
23. Knudsen HK, Studts JL. Physicians as mediators of health policy: acceptance of Medicaid in the context of buprenorphine treatment. J Behav Health Serv Res. 2019;46(1):151-163. doi:10.1007/s11414-018-9629-4
24. Daniels L, Barker S, Chang YS, et al. Harmonizing allergy care-integrated care pathways and multidisciplinary approaches. World Allergy Organ J. 2021;14(10):100584. doi:10.1016/j.waojou.2021.100584
Addressing Gender Disparities in Secondary Heart Disease Prevention
Combatting the Opioid Epidemic: Insights From the Front Lines