Publication
Article
The American Journal of Managed Care
Author(s):
Emergency department (ED) visits and hospitalizations for ambulatory care–sensitive conditions (ACSCs) among Medicaid patients constitute almost 40% of all ED visits and hospitalizations, with lower rates observed in areas with greater proximity to urgent care facilities and density of rural health clinics.
ABSTRACT
Objectives: To compare rates and analyze health facility determinants of emergency department visits and hospitalizations for ambulatory care–sensitive conditions (ACSCs) among Medicaid patients by geographical location.
Study Design: Retrospective cross-sectional analysis of 48.3 million patients receiving Medicaid and their acute care visits across 34 states and the District of Columbia in 2019.
Methods: Descriptive analyses of county-level variations in emergency department visits and hospitalizations (acute care) for ACSCs, and multivariate regressions of proximity to and density of health facility infrastructure as correlates to utilization and spending. Regression models were adjusted for county-level poverty rates, chronic disease rates, and state fixed effects.
Results: Among the studied patient population receiving Medicaid, nearly 40% of acute care visits were for ACSCs, with variations across and within states. Rates ranged from 17.8 per 1000 member-months in Vermont to 39.0 in Mississippi, and from 5.9 to 77.9 between counties within states. Longer distances to the nearest urgent care center and primary care shortage area designation correlated to higher acute care visits for ACSCs (+ 4.3 per 1000 member-months for every 100 miles; 95% CI, 2.9-5.7; P < .001; +1.5 per 1000 member-months if shortage area; 95% CI, 0.4-2.6; P = .006). Counties with more rural health clinics had fewer acute care visits for ACSCs (–3.4 fewer visits per rural clinic per 1000 population; 95% CI, –4.6 to –2.2; P < .001). Among 6 states with additional spending data, 4.2% of total Medicaid spending was attributable to acute care visits for ACSCs.
Conclusions: Our evaluation revealed more than 13-fold variation in acute care utilization for ACSCs between Medicaid counties within the same state. Proximity to urgent care facilities and density of rural health clinics were major explanatory variables for these variations, underscoring the importance of local health infrastructure in reducing acute care utilization for ACSCs.
Am J Manag Care. 2024;30(11):In Press
Takeaway Points
State Medicaid administrators and health plans aim to decrease utilization and spending on emergency department (ED) visits and hospitalizations for ambulatory care–sensitive conditions (ACSCs).
Medicaid is the largest insurer in the US, covering approximately 80.6 million people as of 2022 (~24.2% of the US population1) and representing 17% of national health care spending as of 2021.1 Medicaid health plans have been tasked by state Medicaid administrators to reduce utilization of the emergency department (ED) or hospital (acute care visits) where possible.2
Medicaid plans have historically been challenged by the poor quality of Medicaid claims data when estimating how much opportunity exists to reduce acute care visits for ambulatory care–sensitive conditions (ACSCs), such that prior studies were limited to a few states, a single health plan, older data prior to major movements in Medicaid toward managed care administration, or a focus on a small subset of ED visits.3-5 To the best of our knowledge, no study has examined geographic variation in Medicaid spending on acute care visits for ACSCs.
Medicaid has recently benefited from collaborative efforts to improve the standardization, comprehensiveness, and uniformity of data across states (34 states in 2019).6 As both Medicaid plans and state agencies attempt to reduce acute care visits for ACSCs, the newer data harmonization efforts provide population-level insights into acute care visits and thereby could inform ongoing state and federal debates on which health facility investments may be most helpful to achieve reduced acute care visits for ACSCs.7 Ongoing debates have included whether and to what degree to invest in making urgent care facilities more widely available in addition to primary care facilities,8,9 yet major analyses on the issue to date have been limited to commercial or Medicare data, despite the higher rate of acute care visits in the Medicaid patient population.10 Furthermore, funding and support for expansion of federally qualified health centers (FQHCs) and rural health clinics have been perennially challenged, with debates about how the correlations between access to these clinics and higher acute care visit rates are due to selection bias (as sicker and poorer patients are disproportionately served by these facilities11), despite a lack of rigorous research to identify the population-level effects of these clinic types on Medicaid populations beyond single locations.11,12
Using newly harmonized national Medicaid claims data, we sought to analyze variations in acute care utilization for ACSCs among Medicaid patients across and within US states and counties and to understand the degree to which those variations correlate to differences in density and proximity to health facilities such as urgent care clinics, FQHCs, rural health clinics, and primary care generally.
METHODS
We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (eAppendix Table 1 [eAppendix available at ajmc.com]).13 We compared utilization and spending for acute care visits for ACSCs across states, counties, counties within states, and patients within counties. We then fit multivariate regression models to evaluate the associations between various measures of density and proximity to health care facilities and the observed variations in acute care event rates for ACSCs.
Data
We used 2019 US Medicaid claims data via the Transformed Medicaid Statistical Information System Analytic Files (the most recent available year not affected by COVID-19), which included demographic and eligibility data, geographic information, and claims for outpatient, inpatient, and ED health care services, including both fee-for-service and managed care. We transformed medical claims data to person- and county-level measures for acute care utilization rates for ACSCs.
Participants
We applied 4 inclusion/exclusion criteria. First, to ensure we captured recent improvements in data comprehensiveness and quality, we included data from states meeting minimum quality standards defined by Medicaid.gov’s Data Quality Atlas,14 which incorporated assessments of each state’s enrollment benchmarks, claim volume, and data completeness (eAppendix Methods, Section 4). This resulted in the inclusion of 34 states and the District of Columbia (DC) for ACSC acute care utilization rate calculations. Among the subset of states reporting spending data, we performed an additional analysis to verify spending data quality (eAppendix Methods, Section 4). This resulted in the inclusion of 15 states that met spending data quality criteria for ED visits, 8 states for hospitalizations, and 6 states for both ED visits and hospitalizations.
Second, we included individuals with any enrollment in Medicaid in 2019 to broadly capture utilization.
Third, similar to prior Medicaid studies,15-18 we excluded individuals who were dually enrolled in Medicare and Medicaid because the majority of their medical services are paid through Medicare claims, and dually enrolled persons typically have separate proactive elder care programs with a focus on different facilities, such as skilled nursing facilities (eAppendix Methods, Section 9).
Fourth, consistent with prior literature,19 we limited analysis to counties with at least 100 recipients to ensure stability in utilization estimates and reduce risk of identifiability.
Our final sample consisted of 48,251,536 patients in 2115 counties (34 states and DC) for the primary ACSC utilization rate analysis and, for the 6 states reporting spending data, 5,262,637 patients in 224 counties for the savings rate analysis (eAppendix Figure 5).
Outcomes
We assessed the overall rate of ACSC acute care visits per 1000 member-months; as a percentage of all acute care visits; and by state, county, across counties within states, and across members within counties. Further, among states reporting spending data, we also assessed total cost of acute care for ACSCs (2019 US$ per member per year [PMPY]). As additional secondary outcomes, we separately assessed each of ED visits and hospitalizations in a disaggregated manner.
We defined ED visits based on Current Procedural Terminology codes, revenue codes, and place-of-service codes. To count episodes of care, we linked ED and inpatient claim records for the same patient when dates of service were contiguous.20 Consistent with prior literature,21 ED visits labeled as for ACSCs were identified by the New York University ED Patch Algorithm, and hospitalizations labeled as for ACSCs were identified by the Agency for Healthcare Research and Quality (AHRQ) definition of a Prevention Quality Indicator (eAppendix Methods, Sections 5 and 6).22,23
Covariates
Previous studies focused on a behavioral health model of individual-level decision-making to evaluate predictors of acute care utilization for ACSCs,24-26 often correlating such utilization to race/ethnicity, age, and other nonmodifiable factors. Our hypotheses, however, related to understanding structural, facility-level factors influencing differences among populations in acute care utilization rates for ACSCs to inform population-level investments. Hence, in line with STROBE guidelines for reporting the theoretical concepts that underlie the testing of a hypothesis in observational epidemiology studies, we generated our hypotheses from a conceptual model related to the structural determinants of health care access to choose covariates for our multivariate models,27 focusing in particular on our hypotheses around the density and proximity to different types of health facilities and their relationship to acute care utilization for ACSCs.
Statistical Analysis
We first studied variations in unadjusted county-level rates of acute care visits for ACSCs per 1000 member-months (Figure 1, panel A) and as a percentage of total acute care visits (Figure 1, panel B), and for each of ED visits and hospitalizations individually (eAppendix Figures 1 and 2). We also measured PMPY spending for acute care visits for ACSCs and as a percentage of total medical spending, and for ED visits and hospitalizations individually. Further, we measured county-level spending for acute care visits for ACSCs by state (Figure 2). Finally, we estimated the cost savings if higher-spending counties in a given state had reduced their spending for acute care for ACSCs to the state’s median spending (Figure 3).
To measure the association between county-level utilization of acute care visits for ACSCs, we fit a linear regression model with acute care visits for ACSCs per 1000 member-months as the outcome and the county-level covariates described earlier. We performed the same analysis with spending as the outcome (2019 US$ PMPY) (eAppendix Table 9). Models included predictors of county-level poverty rate, mean number of chronic illnesses using the Elixhauser Comorbidity Index, primary care shortage areas, median distance (miles) to the nearest urgent care clinic and ED, and per capita number of FQHCs and rural health centers from the AHRQ Social Determinants of Health database.28
Analyses were performed in PySpark 3.2.1 (Apache Software Foundation). The Western Institutional Review Board (IRB) exempted this study from review. The data sets utilized in this study are not publicly accessible. They can be obtained from CMS. Accessing these data entails completion of an IRB process and the procurement of a seat on the CMS data portal. However, we did include county-level utilization and spending metrics used in our analysis. Per CMS guidelines, we excluded counties with fewer than 12 visits. Researchers can find the data and code necessary to replicate and extend our study findings on GitHub.29
RESULTS
Data from 34 states and DC, with a total of 48,251,536 unique patients, met comprehensiveness and quality metrics for inclusion in the study (Table 1). The majority of patients identified as female (54.0%), were younger than 18 years (47.1%), were US citizens (84%), were not married (46.5%), and did not indicate having a disability (92.5%). Less than half were White (37.9%) or living below the federal poverty line (37.2%).
Acute Care Visits
Among the study participants, we observed 25.8 acute care visits for ACSCs per 1000 member-months, which varied by a factor of 2.2 across states (range, 17.8 in Vermont to 39.0 in Mississippi; median, 25.3; IQR, 21.0-29.7), a factor of 166 across counties (range, 0.5-82.9; median, 25.3; IQR, 20.0-31.4), a factor of 13.2 across counties within states (range, 5.9 among counties in Hawaii to 77.9 in South Dakota; median, 31.3; IQR, 25.9-36.8), and a factor of 119 across patients within counties (range, 1.0-119.0 visits per patient; median, 12.0; IQR, 8.0-18.0) (Figure 1 and eAppendix Tables 2 and 3). Of all acute care visits, 39.6% were for ACSC visits, again widely varying across states (range, 35.1% in Pennsylvania to 44.2% in Mississippi; median, 39.0%; IQR, 37.4%-40.9%), counties (range, 9.4%-65.0%; median, 39.1%; IQR, 35.8%-42.1%), and across counties within states (range, 3.4% among counties in Hawaii to 47.5% in South Dakota; median, 20.4%; IQR, 16.3%-24.3%) (Figure 1).
ED Visits
There were 25.1 ED visits for ACSCs per 1000 member-months, which varied by more than a factor of 2.2 across states (range, 17.3 in Vermont to 37.9 in Mississippi; median, 24.5; IQR, 20.4-28.8), a factor of 164.2 across counties (range, 0.5-82.1; median, 24.6; IQR, 19.4-30.7), a factor of 12.9 across counties within states (range, 6.0 among counties in Hawaii to 77.1 in South Dakota; median, 31.3; IQR, 21.6-36.4), and a factor of 119.0 across patients within counties (range, 1.0 to 119.0 visits per patient; median, 12.0; IQR, 8.0-17.0) (eAppendix Figure 1 and eAppendix Tables 2 and 3). Of all ED visits, 44.7% were for ACSC visits, again widely varying across states (range, 39.3% in Pennsylvania to 51.0% in Mississippi; median, 44.6%; IQR, 42.8%-46.9%), counties (range, 18.8%-71.6%; median, 45.1%; IQR, 41.8%-48.3%), and across counties within states (range, 5.7% among counties in Hawaii to 44.6% in South Dakota; median, 21.1%; IQR, 15.7%-26.6%) (eAppendix Figure 1).
Hospitalizations
There were 0.7 hospitalizations for ACSCs per 1000 member-months, which varied by a factor of 4.0 across states (range, 0.3 in Maine to 1.2 in Missouri; median, 0.7; IQR, 0.6-1.2), a factor of 350.0 across counties (range, 0.01-3.5; median, 0.6; IQR, 0.4-0.9), and a factor of 16.0 across counties within states (range, 0.2 among counties in Maine to 3.2 in California; median, 1.2; IQR, 0.8-1.7), and a factor of 7300.0 across patients within counties (range, 0.01 to 73.0 visits per patient; median, 4.0; IQR, 2.0-7.0) (eAppendix Figure 2 and eAppendix Tables 2 and 3). Of all hospitalizations, 7.7% were for ACSCs, again widely varying across states (range, 3.6% in Kansas to 12.6% in DC; median, 7.3%; IQR, 6.1%-8.3%), counties (range, 0.0%-29.2%; median, 6.7%; IQR, 5.1%-8.3%), and across counties within states (range, 2.5% among counties in Delaware to 29.2% in North Dakota; median, 11.1%; IQR, 8.2%-15.0%) (eAppendix Figure 2).
Spending
Across 6 states, spending for acute care visits was 33.9% of total medical spending (5.0% for ED visits and 28.9% for hospitalizations). Spending for acute care visits for ACSCs was 4.2% of total medical spending (12.4% of total spending for acute care visits; ED visits made up 2.2% of total medical spending and 1.9% for hospitalizations) (eAppendix Table 4). Overall, spending for acute care visits for ACSCs was $124 PMPY, which varied by a factor of 2.5 across states (range, $107 in Florida to $264 in Alaska; median, $153; IQR, $112-$170), a factor of 44.1 across counties (range, $11-$485; median, $122; IQR, $99-$173), a factor of 4.4 across counties within states (range, $102 in counties in Maine to $452 in counties in Montana; median, $150; IQR, $106-$446), and a factor of 1184 across patients within counties (range, $1 to $1184; median, $24; IQR, $8-$55) (Figure 3 and eAppendix Tables 5 and 6). Subgroup analyses by ED visits and hospitalizations alone are detailed in the eAppendix (eAppendix Figures 3 and 4 and eAppendix Tables 5 and 6).
Statistical Analysis
Across the 2115 counties, there were 26.3 acute care visits for ACSCs per 1000 member-months per county. Higher acute care visits for ACSCs were more common in counties with higher chronic illnesses per member (increase of 11.9 ACSC acute care visits per 1000 member-months per 1-unit increase in the county-level mean number of chronic illnesses; 95% CI, 9.3-14.6; P < .001) and poverty rates (increase of 0.3 ACSC acute care visits per 1000 member-months per 1-unit increase in the percentage of individuals living 50% below the poverty line in a given county; 95% CI, 0.2-0.4; P < .001) (Table 2). Further, longer distances to the nearest urgent care were associated with higher use of acute care for ACSCs (increase of 4.3 acute care visits for ACSCs per 1000 member-months for each 100 miles to the nearest urgent care; 95% CI, 2.9-5.7; P < .001), whereas longer distances to the nearest ED were associated with reduced use (reduction of 8.8 acute care visits for ACSCs per 1000 member-months for each 100 miles to the nearest ED; 95% CI, –11.1 to –6.5; P < .001). Finally, counties with more FQHCs had a nonsignificant reduction (reduction of 1.2 acute care visits for ACSCs per 1000 member-months per 1 FQHC increase per 1000 population; 95% CI, –2.9 to 0.5; P = .18), whereas those with more rural health clinics had a significant reduction in acute care visits for ACSCs (reduction of 3.4 acute care visits for ACSCs per 1000 member-months per 1 clinic increase per 1000 population; 95% CI, –4.6 to –2.2; P < .001). Counties designated as whole-county primary care shortage areas had higher use (increase of 1.5 acute care visits per 1000 member-months; 95% CI, 0.4-2.6; P = .006) compared with counties with no primary care shortage areas. We found parallel results when regressing the covariates against the outcome of spending (eAppendix Table 9).
DISCUSSION
In this analysis of 48.3 million patients receiving Medicaid across 34 states and DC, we found that nearly 40% of acute care visits in 2019 were for ACSCs, with rates and spending widely varying across and within states. The percentage of total acute care visits for ACSCs ranged from 35% to 44% across states and 3% to 48% across counties within states. Our results were across a much larger number of states and patients compared with results from a prior study in Medicaid, which was limited to 4 states.3 Our results also highlight an increase in acute care visits for ACSCs compared with studies of older data sets (for example, 14%-27% of all ED visits were for ACSCs compared with our finding of 40%).30-32 One plausible explanation is the greater difficulty in scheduling health care appointments among Medicaid patients compared with commercially insured patients.33 Another is the high rate of unaddressed social needs affecting health outcomes among patients receiving Medicaid.34 Our results reflect high rates of acute care utilization among Medicaid-receiving populations. This rate is higher than those previously reported among the Medicare or commercial populations in prior years.35
Importantly, county-level proximity to urgent care facilities and density of rural health clinics were associated with decreased use of acute care visits for ACSCs, whereas being a primary care shortage area was associated with increased acute care use for ACSCs. Although some limited data have supported the association between urgent care and acute care utilization among commercial populations,9 results have been largely unavailable for the population covered by Medicaid. Additionally, although FQHC availability has been associated with both increased and decreased utilization, we find that controlling for poverty and number of chronic conditions may help adjust for confounding, and thus our results are more consistent with single-study Medicaid research that showed associations between FQHCs and lower utilization of services for ACSCs.36 Notably, our results across a larger national sample were not statistically significant for all FQHCs, only designated rural health clinics. Our findings also underscore the relationship between higher poverty and increased use of acute care visits for ACSCs because these communities may lack the social resources most associated with acute care visits for ACSCs, such as food and housing insecurity that relate to ACSCs such as hypertension, diabetes, and substance use.34,37-39
Limitations
Our analysis nevertheless has several limitations. First, we excluded 16 states with insufficient data comprehensiveness or quality, although our study is more inclusive than previous studies.3 Second, we used claims-based algorithms, relying on discharge diagnoses, to categorize acute care visits as being for ACSCs. Electronic health record data could provide additional context on whether the chief complaint required ED, hospital, or primary care treatment. Prior studies indicate a 90% concordance between chief complaint and discharge diagnosis, which could lead to an overestimation of acute care visits for ACSCs.40 Third, we used data from 2019 instead of 2020 because the COVID-19 pandemic led to an 18% to 25% reduction in ED visits compared with prior years, which would result in an underestimation of acute care visits for ACSCs during nonpandemic periods.41,42 Fourth, we excluded dually eligible Medicare and Medicaid patients, who make up 14% of the Medicaid population because their claims are mostly in Medicare data and they often have elder care–specific programs that are thought to profoundly affect use of acute care utilization. Utilization among Medicare-insured and dually eligible persons has been extensively studied, so we sought to fill the relative gap in research on Medicaid-only individuals. Dually eligible individuals are generally sicker and hospitalized more frequently for ACSCs, so their exclusion likely lowers our acute care visit estimates for ACSCs. Sixth, our statistical analysis did not account for physician practice styles, and literature indicates that primary care provider practice styles can explain a significant share of the variation in more distant types of utilization, such as ED visits and hospitalizations.43 Finally, our multivariate analysis may have residual confounding due to provider practice type (eg, FQHC practices) and variations in utilization, leading to potential unmeasured confounders in the estimated associations.
CONCLUSIONS
In the current analysis, nearly 40% of acute care visits among Medicaid recipients were for ACSC, with rates and spending widely varying across and within states. County-level availability of urgent care, rural health clinics, and overall primary care was associated with reduced use of acute care for ACSCs. Our study was limited to states with adequate Medicaid data quality, but nevertheless, it constitutes the largest study of acute care among Medicaid populations to date. Our findings highlight the need for cross-region learning for evidence-based geographic initiatives to improve care utilization and facility access for patients receiving Medicaid.
Author Affiliations: Clinical Product Development, Waymark Care (SYP, AB, SB), San Francisco, CA; School of Social Policy and Practice, University of Pennsylvania (SYP), Philadelphia, PA; Icahn School of Medicine at Mount Sinai (AB), New York, NY; Institute of Health Policy, Management and Evaluation, University of Toronto (SB), Toronto, Ontario, Canada; Center for Vulnerable Populations, San Francisco General Hospital/University of California, San Francisco (SB), San Francisco, CA.
Source of Funding: None.
Author Disclosures: Drs Patel, Baum, and Basu are employees of Waymark (a health care delivery organization), own stock options or shares in Waymark, and have patents pending for “Predicting Changes in Risk Based on Interventions.” Dr Basu also is a board member of Waymark, is an employee of HealthRIGHT360, has received grants from the National Institutes of Health (R01DK125406, P30DK092924, U18DP006526, and R01DK116852) and CDC, receives personal fees from the University of California, San Francisco, and has received patents for an occupational health testing protocol.
Authorship Information: Concept and design (SYP, AB, SB); acquisition of data (SYP, AB); analysis and interpretation of data (SYP, AB, SB); drafting of the manuscript (SYP); critical revision of the manuscript for important intellectual content (SYP, AB, SB); statistical analysis (SYP); administrative, technical, or logistic support (SYP); and supervision (SB).
Address Correspondence to: Sadiq Y. Patel, PhD, MS, Waymark Care, 2021 Fillmore St, Ste 1059, San Francisco, CA 94115. Email: sadiq.patel@waymarkcare.com.
REFERENCES
1. Medicaid: An Overview. Congressional Research Service. Updated February 8, 2023. Accessed December 1, 2023. https://crsreports.congress.gov/product/pdf/R/R43357
2. Gadbois EA, Gordon SH, Shield RR, Vivier PM, Trivedi AN. Quality management strategies in Medicaid managed care: perspectives from Medicaid, plans, and providers. Med Care Res Rev. 2021;78(1):36-47. doi:10.1177/1077558719841157
3. Giannouchos TV, Ukert B, Andrews C. Association of Medicaid expansion with emergency department visits by medical urgency. JAMA Netw Open. 2022;5(6):e2216913. doi:10.1001/jamanetworkopen.2022.16913
4. Hosseini H. Misallocation of demand and the persistent non-emergent use of the emergency department post-healthcare reform. Hosp Top. 2020;98(2):51-58. doi:10.1080/00185868.2020.1750325
5. Swaminathan S, Ndumele CD, Gordon SH, Lee Y, Trivedi AN. Association of Medicaid-focused or commercial Medicaid managed care plan type with outpatient and acute care. JAMA Intern Med. 2020;180(12):1672-1679. doi:10.1001/jamainternmed.2020.5408
6. Williams N, Mayer CS, Huser V. Data characterization of Medicaid: legacy and new data formats in the CMS Virtual Research Data Center. AMIA Jt Summits Transl Sci Proc. 2021;2021:644-652.
7. Strengthening primary care through Medicaid managed care. Center for Health Care Strategies. 2021-2022. Accessed February 12, 2024. https://www.chcs.org/resource/strengthening-primary-care-through-medicaid-managed-care/
8. Hartnett K. Investors see opportunity in high-needs Medicaid market. Modern Healthcare. November 30, 2023. Accessed February 12, 2024. https://www.modernhealthcare.com/medicaid/high-needs-medicaid-
investment-cityblock-carebridge-waymark
9. Wang B, Mehrotra A, Friedman AB. Urgent care centers deter some emergency department visits but, on net, increase spending. Health Aff (Millwood). 2021;40(4):587-595. doi:10.1377/hlthaff.2020.01869
10. Chow JY, Comulada WS, Gildner JL, Desmond KA, Leibowitz AA. Association between Federally Qualified Health Center usage and emergency department utilization among California’s HIV-infected Medicaid beneficiaries, 2009. AIDS Care. 2019;31(4):519-527. doi:10.1080/09540121.2018.1524112
11. Nath JB, Costigan S, Lin F, Vittinghoff E, Hsia RY. Access to federally qualified health centers and emergency department use among uninsured and Medicaid-insured adults: California, 2005 to 2013. Acad Emerg Med. 2019;26(2):129-139. doi:10.1111/acem.13494
12. Myong C, Hull P, Price M, Hsu J, Newhouse JP, Fung V. The impact of funding for federally qualified health centers on utilization and emergency department visits in Massachusetts. PLoS One. 2020;15(12):e0243279. doi:10.1371/journal.pone.0243279
13. Ghaferi AA, Schwartz TA, Pawlik TM. STROBE reporting guidelines for observational studies. JAMA Surg. 2021;156(6):577-578. doi:10.1001/jamasurg.2021.0528
14. DQ Atlas. Medicaid.gov. Accessed December 1, 2023. https://www.medicaid.gov/dq-atlas/
15. Kronick R, Gilmer T, Dreyfus T, Lee L. Improving health-based payment for Medicaid beneficiaries: CDPS. Health Care Financ Rev. 2000;21(3):29-64.
16. Weir S, Aweh G, Clark RE. Case selection for a Medicaid chronic care management program. Health Care Financ Rev. 2008;30(1):61-74.
17. Gilmer T, Kronick R, Fishman P, Ganiats TG. The Medicaid Rx model: pharmacy-based risk adjustment for public programs. Med Care. 2001;39(11):1188-1202. doi:10.1097/00005650-200111000-00006
18. Ettner SL, Frank RG, McGuire TG, Hermann RC. Risk adjustment alternatives in paying for behavioral health care under Medicaid. Health Serv Res. 2001;36(4):793-811.
19. Patel SY, Rose S, Barnett ML, Huskamp HA, Uscher-Pines L, Mehrotra A. Community factors associated with telemedicine use during the COVID-19 pandemic. JAMA Netw Open. 2021;4(5):e2110330. doi:10.1001/jamanetworkopen.2021.10330
20. Patel SY, Huskamp HA, Barnett ML, et al. Association between telepsychiatry capability and treatment of patients with mental illness in the emergency department. Psychiatr Serv. 2022;73(4):403-410. doi:10.1176/appi.ps.202100145
21. Galarraga JE, Pines JM. Costs of ED episodes of care in the United States. Am J Emerg Med. 2016;34(3):357-365. doi:10.1016/j.ajem.2015.06.001
22. Johnston KJ, Allen L, Melanson TA, Pitts SR. A “patch” to the NYU emergency department visit algorithm. Health Serv Res. 2017;52(4):1264-1276. doi:10.1111/1475-6773.12638
23. Guide to Prevention Quality Indicators: Hospital Admission for Ambulatory Care Sensitive Conditions. Agency for Healthcare Research and Quality; October 2001. Accessed February 1, 2024. https://www.ahrq.gov/downloads/pub/ahrqqi/pqiguide.pdf
24. Adigun AC, Maguire K, Jiang Y, Qu H, Austin S. Urgent care center and emergency department utilization for non-emergent health conditions: analysis of managed care beneficiaries. Popul Health Manag. 2019;22(5):433-439. doi:10.1089/pop.2018.0138
25. Tapia AD, Cuenca CM, Johnson SJ, et al. Assessing challenges with access to care for patients presenting to the emergency department for non-emergent complaints. Med J (Ft Sam Houst Tex). 2021;(PB 8-21-07/08/09):74-80.
26. Uscher-Pines L, Pines J, Kellermann A, Gillen E, Mehrotra A. Emergency department visits for nonurgent conditions: systematic literature review. Am J Manag Care. 2013;19(1):47-59.
27. Marmot M, Allen JJ. Social determinants of health equity. Am J Public Health. 2014;104(suppl 4):S517-S519. doi:10.2105/AJPH.2014.302200
28. Social Determinants of Health database: download data and documentation. Agency for Healthcare Research and Quality. Updated June 2023. Accessed February 1, 2024. https://www.ahrq.gov/sdoh/data-analytics/sdoh-data.html#download
29. Patel SY. ACS acute care Medicaid. GitHub. Accessed February 1, 2024. https://github.com/sadiqypatel/acs-acute-care-medicaid
30. Weinick RM, Burns RM, Mehrotra A. Many emergency department visits could be managed at urgent care centers and retail clinics. Health Aff (Millwood). 2010;29(9):1630-1636. doi:10.1377/hlthaff.2009.0748
31. Statewide Collaborative Quality Improvement Project: Reducing Avoidable Emergency Room Visits. Final Remeasurement Report: January 1, 2010 – December 31, 2010. California Department of Health Care Services; June 2012. Accessed December 1, 2023. https://www.dhcs.ca.gov/dataandstats/reports/Documents/MMCD_Qual_Rpts/EQRO_QIPs/CA2011-12_QIP_Coll_ER_Remeasure_Report.pdf
32. Chou SC, Gondi S, Weiner SG, Schuur JD, Sommers BD. Medicaid expansion reduced emergency department visits by low-income adults due to barriers to outpatient care. Med Care. 2020;58(6):511-518. doi:10.1097/MLR.0000000000001305
33. 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
34. Canterberry M, Figueroa JF, Long CL, et al. Association between self-reported health-related social needs and acute care utilization among older adults enrolled in Medicare Advantage. JAMA Health Forum. 2022;3(7):e221874. doi:10.1001/jamahealthforum.2022.1874
35. Moore BJ, Liang L. Costs of emergency department visits in the United States, 2017. Healthcare Cost and Utilization Project statistical brief 268. Agency for Healthcare Research and Quality. December 2020. Accessed December 5, 2023. https://hcup-us.ahrq.gov/reports/statbriefs/sb268-ED-Costs-2017.jsp
36. Myong C, Hull P, Price M, Hsu J, Newhouse JP, Fung V. The impact of funding for federally qualified health centers on utilization and emergency department visits in Massachusetts. PLoS One. 2020;15(12):e0243279. doi:10.1371/journal.pone.0243279
37. Levi R, Bleich SN, Seligman HK. Food insecurity and diabetes: overview of intersections and potential dual solutions. Diabetes Care. 2023;46(9):1599-1608. doi:10.2337/dci23-0002
38. Beltrán S, Pharel M, Montgomery CT, López-Hinojosa IJ, Arenas DJ, DeLisser HM. Food insecurity and hypertension: a systematic review and meta-analysis. PLoS One. 2020;15(11):e0241628. doi:10.1371/journal.pone.0241628
39. Pan Y, Metsch LR, Wang W, et al. The relationship between housing status and substance use and sexual risk behaviors among people currently seeking or receiving services in substance use disorder treatment programs. J Prim Prev. 2020;41(4):363-382. doi:10.1007/s10935-020-00597-x
40. Raven MC, Lowe RA, Maselli J, Hsia RY. Comparison of presenting complaint vs discharge diagnosis for identifying “nonemergency” emergency department visits. JAMA. 2013;309(11):1145-1153. doi:10.1001/jama.2013.1948
41. Molina M, Evans J, Montoy JC, et al. Analysis of emergency department encounters among high users of health care and social service systems before and during the COVID-19 pandemic. JAMA Netw Open. 2022;5(10):e2239076. doi:10.1001/jamanetworkopen.2022.39076
42. Melnick G, O’Leary JF, Zaniello BA, Abrishamian L. COVID-19 driven decline in emergency visits: has it continued, is it permanent, and what does it mean for emergency physicians? Am J Emerg Med. 2022;61:64-67. doi:10.1016/j.ajem.2022.08.031
43. Fadlon I, Van Parys J. Primary care physician practice styles and patient care: evidence from physician exits in Medicare. J Health Econ. 2020;71:102304. doi:10.1016/j.jhealeco.2020.102304