Identifying Social Need Trends Among Emergency Department Super-Users: A Latent Class Analysis.

Identifying Social Need Trends Among Emergency Department Super-Users: A Latent Class Analysis

Introduction

Frequent emergency department (ED) users, particularly those with high rates of use, often experience complex social determinants of health (SDOH) that contribute to their repeated visits and may lead to adverse health outcomes. Our objective in this study was to characterize the needs most commonly affecting this population, which may support the creation of tailored interventions with a higher chance of success.

Methods

This was a retrospective analysis of patients evaluated by an ED social medicine team in a large urban academic ED from May 2021–December 2023. Patients were identified either via frequent ED use (patients with the 10 highest number of visits each month) or referred by ED staff. We assigned 11 categories of medical, mental health, and social needs through chart review and team documentation. The primary outcome was the identification of distinct need-based subgroups among patients with high ED use, using latent class analysis of documented social, mental health, and medical needs.

Results

Of the 181 patients, representing approximately 0.4% of the ED patient population, 149 (82.3%) were frequent ED users, with a median of 14 ED visits annually (IQR 8–21), while the rest were referred by ED staff (median 2 visits (1–4.25). The majority were non-Hispanic Black (159, 87.8%) and publicly insured (163, 90.0%). The most common needs were related to housing instability (116, 64.1%) and limited access to outpatient care (112, 61.9%). Latent class analysis revealed four distinct subgroups, all of which included limited access to outpatient care: 1) primarily limited access to outpatient care (40, 22.1%); 2) housing instability and other basic needs (27, 14.9%); 3) housing instability, psychotic disorders, other basic needs, and inability to live independently (23, 12.7%); and 4) housing instability, substance use, and psychotic disorders (91, 50.3%).

Conclusion

This analysis identified four distinct need-based subgroups of frequent ED users, suggesting the need for tailored, multifaceted interventions. Housing instability, psychiatric illness, substance use, and outpatient care access were prevalent across groups, underscoring the importance of including resources for these basic needs in any programs developed for this population. Findings support prioritizing coordinated care models, particularly those that incorporate housing, behavioral health, substance use resources, and navigation support, for this vulnerable population.

INTRODUCTION

Social determinants of health (SDOH) have been shown to negatively impact the health outcomes of individuals.1,2 Defined by the World Health Organization as the “non-medical factors [such as income, education, employment status, housing, and access to proper care] that influence health outcomes,”3 it is estimated that more than half of the U.S. population is affected by at least one adverse SDOH.4,5 For example, around 26.4 million American households are uninsured,6,7 which prevents them from being able to seek the routine care they need, resulting in worse health outcomes,8,9 higher rates of chronic diseases,5,10–12 and increased reliance on the emergency department (ED) for care.13

Emergency departments play a unique role in the healthcare system because, in addition to treating emergent conditions, they serve as a public health safety net, an entry point into the medical system, and sometimes the sole source of medical care and social services for individuals with limited access to care. Frequent ED users, commonly defined as individuals who present to the ED more than four to six times in a given year,14 have been found to have a disproportionate burden of SDOH.14,15 It is estimated that frequent ED users account for 3.816% of all ED users but 14–47% of all ED visits.11 Prior studies have found that frequent ED users are more often publicly insured, have a disproportionate burden of chronic diseases and unmet primary care needs, perceive their own health status as poor, and have a higher likelihood of substance abuse, unemployment, mental health issues, and unstable housing.11,16

While a growing body of literature has examined frequent ED users, the population is somewhat ambiguously defined and generally includes individuals with visit rates more than four per year.11,13,17,18 However, given the definition’s inconsistency as well as the population’s unique characteristics, the subset of this population with the highest reliance on the ED for care tends to be underexamined.19 These “super-utilizers”18 are sometimes defined as those who visit the ED > 18 times in a year. However, while there are likely important differences between individuals who visit the ED four times a year and those who visit three or four times that number, a standard definition of the minimum visit rate for this group has not been established. A few studies have suggested that the individuals with the highest number of ED visits are more likely to be uninsured, have lower acuity visits,20,21 and tend to visit the ED in place of other healthcare sources.18 Despite their disproportionate number of ED visits, little is known about these super-users, how different their needs may be from patients with a lower ED use rate, and what types of interventions would be most beneficial for them.

In 2021, a large urban academic hospital ED on the South Side of Chicago, an area of high economic hardship, implemented an ED social medicine team as a quality improvement (QI) initiative, modeled after a similar program in San Francisco.22 The team included social workers, care coordinators, patient navigators, behavioral health specialists, pharmacists, and an emergency physician. Team members met monthly to discuss patient cases and coordinate interventions. This multidisciplinary team model allowed for more comprehensive interventions for individuals with greater needs, as identified by high visit rates or clinician assessment, and the monthly meetings provided an opportunity for in-depth review of charts and group discussion of all potentially relevant resources. This project aims to review this cohort of individuals, most of whom would be considered not just frequent ED users, but super-users, with ED visit rates well above four per year, to understand the drivers of their ED use, with the ultimate goal of developing more efficient and effective interventions to better meet the specific needs of the most frequent ED users.

Population Health Research Capsule

What do we already know about this issue?

Frequent emergency department (ED) users report substantial social and medical needs, but the highest users are poorly characterized and lack targeted interventions.

What was the research question?

What distinct need-based subgroups exist among frequent ED users?

What was the major finding of the study?

Latent class analysis identified four subgroups; housing instability (64%) and limited access to outpatient care (62%) were the most common needs.

How does this improve population health?

Identifying major needs and subgroups enables development of targeted, coordinated interventions to reduce ED use among the highest users.

METHODS

We conducted a retrospective review of all individuals followed by the social medicine team from its inception in May 2021 through December 2023. Patients were identified for team intervention either through frequent ED visits (among the top 10 most frequent ED users in the calendar month prior to the team meeting) or referral from ED staff. We performed case selection criteria, variable definition, abstraction forms, missing-data management planning, abstractor training, performance monitoring, and medical record identification in line with recommended medical record review criteria.23 We recorded interventions by the ED social medicine team for clinical care in a REDCap secure database,24,25 which was supplemented with manual chart review. Demographic information, reasons for individual ED visits, and SDOH were extracted for each patient.

For each individual, it was noted whether they had been referred to the team by ED staff or identified via the monthly frequent ED users report. And for each of them we recorded the number of ED visits at the primary project site during the 12 months after that individual’s initial evaluation by the ED social medicine team. Although some visits to other EDs are visible in the electronic health record (EHR), many local EDs do not use the same EHR system and, thus, visits outside the primary site ED could not be reliably tracked. Because of this uncertainty, only visits at the primary site were included. Of note, patients with ED visits solely due to unavoidable medical conditions and those visiting primarily for sickle cell pain crises were not eligible for the ED social medicine team intervention, because these visits were considered medically unavoidable and not related to SDOH.

Categorization of Patient Needs

We created 11 categories of patient needs (Table 1) based on our clinical experience, frequently occurring themes in the database, and review of the literature. Categories were assigned to each individual based on group discussion by the ED social medicine team, as documented in the database, as well as review of the medical chart by two independent reviewers. Individuals were assigned to a category if it was noted during at least one ED visit, and individuals could be assigned to multiple categories.

Statistical Analysis

We characterized individuals in the social medicine team’s database descriptively based on their sociodemographic information and their reported needs. We used medians and interquartile ranges (IQR) for continuous variables (such as visit rates) and frequencies and percentages to describe the population for categorical variables. We used a chi-square test of independence to determine statistical significance between groups.

Latent class analysis was used to determine groupings of needs in the patient population. We used poLCA v. 1.6.0.1,26 a latent class analysis package built for R (R Foundation for Statistical Computing, Vienna, Austria),27 and the model was run with 2–10 clusters. The optimal number of clusters was determined by the lowest value of the Akaike information criterion.28–30 The model was validated by observing lower off-diagonal values and higher diagonal values in the posterior probability matrix.29 We ran the poLCA model 1,000 times with five different random starting positions to account for cluster assignment variation among study subjects. Cluster assignment was determined by the frequency of cluster assignment. Model fit and validity were assessed using Akaike information criterion and Bayesian information criterion for class selection, maximum likelihood estimation with multiple random starting values to ensure convergence and avoid local maxima, and posterior class membership probabilities to evaluate classification certainty. All data analysis was performed using R software v4.0.3.

This project was formally determined to be QI, not human subjects research; therefore, it was not overseen by the institutional review board, per institutional policy.

RESULTS

The ED social medicine team evaluated 181 patients from May 2021–December 2023. Of these, 149 (82.3%) were frequent ED users, identified via the monthly ED use report, and 32 (17.7%) were referred by ED staff. Frequent ED users had a median of 14 visits (IQR 8–21) to the primary project ED in the initial 12 months after evaluation by the social medicine team. Many of these individuals had visits to other EDs during this period as well, but these could not be reliably quantified from the EHR and were not included. Among individuals referred by staff, the median was two visits (IQR 1, 4.25) to the primary project ED, with a maximum of 16. Overall, 51 (28.2%) of the population had more than 18 visits in 12 months.

Among all patients, the median age was 55 (IQR 36–65), and 58% were male (Table 2). Most of the patients identified by the ED social medicine team were non-Hispanic Black (87.8%), reflecting the racial demographics of the local ED population, and used public insurance (92.8%), with 6.1% uninsured. The distribution of demographics was similar between the two groups, except that frequent ED users notably had a higher percentage of males than females (61.7% vs 40.6%) compared to referred patients. At the end of the evaluation period, 13 (7.18%) individuals were confirmed deceased, 11 of whom were frequent ED users.

Table 3 shows the frequency of each recorded need in the population. The two most common needs observed in the population were housing instability (116, 64.1%) and limited access to outpatient care (112, 61.9%). Overall, 64 (35.4%) had a history of substance use disorder, 76 (42%) had difficulty meeting basic needs, and 64 (35.4%) had a history of a psychotic disorder.

Some differences were observed between frequent ED users and the individuals referred to the ED social medicine team, although most of these were not statistically significant. Frequent ED users were more frequently noted to have substance use disorder (37.6% vs 25.1%), psychiatric illness (psychotic disorder 38.3% versus 21.9%, anxiety/depression 23.5% versus 15.6%), and other basic needs such as lack of a cellphones or reliable transportation (43.6% versus 34.4%), while referred patients more frequently were unable to live independently (28.9% versus 43.8%).

We performed latent class analysis for the entire ED social medicine team population and identified four distinct subgroups characterized by the predominant needs of each group (Table 4), all of which included limited access to outpatient care: 1) primarily needed assistance accessing outpatient care; 2) housing instability and other basic needs; 3) housing instability, other basic needs, inability to live independently, and psychotic disorders; and 4) housing instability, substance use, and psychotic disorders. Subgroup 4 was the largest group, with 50.3% in this subgroup. Access to outpatient hemodialysis was noted in subgroup 1 (23%) and subgroup 4 (4%). Housing instability was highly prevalent across all subgroups, ranging from 40% in subgroup 1 to 96% in subgroup 2. The median number of visits in 12 months after involvement of the ED social medicine team was similar between the subgroups (14, 12, 12, 10, respectively), with a relatively high IQR (6–19.3, 6.5–20, 2.5–19.5, 6–19) in all subgroups, reflecting the presence of very high users in each subgroup. Frequent ED users in subgroup 4 had a median visit number of 12.5 (IQR 13 : [8,21]). Referred individuals comprised a larger proportion of subgroups 1 and 3 (22.5% and 26.1%, respectively) compared to subgroups 2 and 4 (7.41% and 16.5%, respectively).

DISCUSSION

To date, the literature has not focused on frequent ED users with extremely high ED visit rates. In this sample, frequent ED users had a median 12-month ED visit rate well above the threshold of four visits per year that is commonly used to define frequent ED use, and more than a quarter had over 18 visits per year. In one study of Medicaid data, the medical needs and SDOH of very high-use ED patients correlated significantly with those noted in the largest subgroup in this project, suggesting these may be common needs that will be key to address in future interventions. The combination of substance abuse, housing needs, and mental illness, which were major drivers identified in the largest subgroup in this study, has been shown to increase the likelihood of ED use in other studies as well.31–34 “Housing first” programs may decrease hospital and ED use, even without integrated substance use interventions,35,36 but they could be even more effective when using a holistic approach that includes substance use and mental health resources.37 These interventions can be implemented directly from the ED visit,38 which may be a critical opportunity for engagement in a housing-first model.

Interestingly, the group of individuals defined by unstable housing, substance use, and mental health needs, which constituted the largest subgroup, had a lower median ED visit rate than frequent ED users overall. While we were unable to examine resource use outside the primary site ED, members of this group were frequently documented to visit multiple EDs around the city, spend prolonged periods incarcerated, or occasionally be admitted to psychiatric hospitals. This may account for sporadic heavy use of the primary site ED, followed by periods of absence, creating the illusion of lower ED use when, in fact, this group’s overall use of the ED and system resource may have been much higher—an important area to consider in future research.

As has been reported in previous studies,11,14,39–41 most individuals in this QI project used either Medicaid or Medicare. Evidence suggests that Medicaid users experience disproportionate difficulty accessing care, including being less likely to have a usual source of care, facing long wait times and low appointment availability, and transportation issues,42–44 which is associated with higher ED use.44,45 While some studies have suggested that individuals with the highest ED use may not represent a burden on ED costs or crowding due to their relatively low prevalence and low-acuity visits,20 in this QI project some individuals were visiting the ED almost daily for a period of weeks to months. Further study is needed to understand the impact of this type of heavy ED use, in particular given low Medicaid reimbursement rates and bed use by uninsured and underinsured patients.

This analysis identified four major subgroups, or clinical phenotypes, of individuals who were evaluated by the ED social medicine team. While some characteristics, such as housing instability, psychiatric illness, and access to outpatient care, were present in all subgroups, differences between the subgroups suggest that tailored interventions could be developed to target the specific needs of each subgroup more efficiently. For instance, subgroup 4, the largest subgroup, experienced the confluence of mental health, housing instability, and substance use, suggesting that more resources should be put toward comprehensive interventions that target all these needs simultaneously, such as housing programs incorporating substance use treatment and mental health support on site. While these needs have been identified among patients with repeated hospitalizations,46 this study is novel in its focus specifically on the highest users of ED care, as well as the amount of data available regarding distinct SDOH. Existing literature generally focuses on those with more than four or five annual ED visits,47,48 which may not be representative of the needs of this very specific subset of individuals. Given the prevalence of housing instability, psychiatric illness, and access to outpatient care in all subgroups, all interventions for a similar population should consider including housing, mental health support, and community health workers or care navigators to assist individuals in accessing the care they need. Some existing interventions have already shown promise in improving outcomes for this group of individuals, but increased support for these programs is needed from health departments and other government agencies that could provide funding or infrastructure that individual EDs cannot develop or sustain alone. Implementation research examining service delivery strategies in a variety of hospital settings and providing evidence of feasibility and efficacy is urgently needed.

LIMITATIONS

This was a single-center QI initiative and, as such, results may not be generalizable to other locations. However, the findings are consistent with existing literature, and they are likely applicable to other urban EDs serving socioeconomically depressed populations. Additionally, because the establishment of an ED social medicine team was a QI initiative, team structure, record keeping, and interventions evolved over time. Identification of social needs is somewhat subjective, and this review relied on needs being recorded in a database or in the medical record, which may not always have been consistently done.

Individuals may not have disclosed certain SDOH to their medical team, and if not relevant to their ED visit, these needs may not have been recorded in the chart. Some individuals had more comprehensive social work evaluations, while others were unable or unwilling to participate in these evaluations, which may have affected the team’s ability to accurately identify their SDOH. Individuals identified to have only medical needs driving their ED visits, including sickle cell pain crises, were excluded, but it is possible that they, too, had unidentified SDOH. Lastly, because members of the ED social medicine team were not reliably able to access information about ED visits or medical admissions at other hospitals, psychiatric admissions, or incarceration, use of system resources could not be completely evaluated. However, the patterns observed in this QI project are likely similar to those of ED super-users elsewhere, and the findings about ED use at a single site are relevant to other EDs caring for similar populations.

CONCLUSION

Frequent users of the emergency department contribute to crowding and economic consequences; at the same time, they have worse health outcomes than other ED visitors. Needs identified in this analysis—most notably housing instability, psychiatric illness, substance use, and difficulty accessing outpatient care—could be used to develop targeted, comprehensive interventions that might have a higher likelihood of success than the current piecemeal system employed in most hospitals. As social safety nets face increasing challenges, placing additional strain on EDs and social service workers, it will be critical to advocate for additional, evidence-based resources that effectively address the most pressing needs of this vulnerable group.

REFERENCES

1Vo A Tao Y Li Y The association between social determinants of health and population health outcomes: Ecological analysis JMIR Public Health and Surveillance 2023 9 1 e44070 10.2196/44070 PMC10131773 36989028

2Krause TM Schaefer C Highfield L The association of social determinants of health with health outcomes Am J Manag Care 2021 27 3 e89 e96 10.37765/ajmc.2021.88603 33720674

3World Health OrganizationSocial determinants of health 2024 Available at: https://www.who.int/health-topics/social-determinants-of-healthAccessed December 16, 2024

4Kim EJ Abrahams S Uwemedimo O Prevalence of social determinants of health and associations of social needs among United States adults, 2011–2014 J Gen Intern Med 2020 35 5 1608 1609 10.1007/s11606-019-05362-3 PMC7210342 31749029

5Hacker K Social determinants of health and health-related social needs among adults with chronic diseases in the United States, behavioral risk factor surveillance system, 2022 Prev Chronic Dis 2024 21 10.5888/pcd21.240362 PMC11608007 39602222

6Briones EM Cohen RA Health insurance coverage: Early release of estimates from the National Health Interview Survey, January–June 2024 National Center for Health Statistics (U.S.)2024

7Keisler-Starkey K Bunch LN Health insurance coverage in the United States: 2023 U.S. Census Bureau Current Population Reports 2024 60 284

8Castaneda MA Saygili M The health conditions and the health care consumption of the uninsured Health Econ Rev 2016 6 1 55 10.1186/s13561-016-0137-z PMC5142170 27924584

9McWilliams MJ Health consequences of uninsurance among adults in the United States: Recent evidence and implications Milbank Q 2009 87 2 443 494 10.1111/j.1468-0009.2009.00564.x PMC2881446 19523125

10Komenda P Tangri N Klajncar E Patterns of emergency department utilization by patients on chronic dialysis: A population-based study PLoS One 2018 13 4 e0195323 10.1371/journal.pone.0195323 PMC5903639 29664922

11Giannouchos TV Kum HC Foster MJ Characteristics and predictors of adult frequent emergency department users in the United States: A systematic literature review J Eval Clin Pract 2019 25 3 420 433 10.1111/jep.13137 31044484

12Eisner MD Blanc PD Omachi TA Socioeconomic status, race and COPD health outcomes J Epidemiol Community Health 2011 65 1 26 34 10.1136/jech.2009.089722 PMC3017471 19854747

13Giannouchos TV Washburn DJ Kum HC Predictors of multiple emergency department utilization among frequent emergency department users in 3 states Med Care 2020 58 2 137 145 10.1097/MLR.0000000000001228 31651740

14LaCalle E Rabin E Frequent users of emergency departments: The myths, the data, and the policy implications Ann Emerg Med 2010 56 1 42 48 10.1016/j.annemergmed.2010.01.032 20346540

15Malecha PW Williams JH Kunzler NM Material needs of emergency department patients: A systematic review Acad Emerg Med 2018 25 3 330 359 10.1111/acem.13370 29266523

16Andriotti T Dalton MK Jarman MP Super-utilization of the emergency department in a universally insured population Mil Medicine 2021 186 9–10 e1010 e1016 10.1093/milmed/usaa399 33247301

17Iovan S Lantz PM Allan K Interventions to decrease use in prehospital and emergency care settings among super-utilizers in the United States: A systematic review Med Care Res Rev 2020 77 2 99 111 10.1177/1077558719845722 31027455

18Doupe MB Palatnick W Day S Frequent users of emergency departments: Developing standard definitions and defining prominent risk factors Ann Emerg Med 2012 60 1 24 32 10.1016/j.annemergmed.2011.11.036 22305330

19Pines JM Asplin BR Kaji AH Frequent users of emergency department services: Gaps in knowledge and a proposed research agenda Acad Emerg Med 2011 18 6 e64 e69 10.1111/j.1553-2712.2011.01086.x 21676051

20Ruger JP Richter CJ Spitznagel EL Analysis of costs, length of stay, and utilization of emergency department services by frequent users: Implications for health policy Acad Emerg Med 2004 11 12 1311 1317 10.1197/j.aem.2004.07.008 15576522

21Peabody C Gruber P Lam CN Medical necessity and resource use among frequent emergency department users Ann Emerg Med 2014 64 4 S10

22Chase J Bilinski J Kanzaria HK Caring for emergency department patients with complex medical, behavioral health, and social needs JAMA 2020 324 24 2550 2551 10.1001/jama.2020.17017 33351045

23Worster A Bledsoe RD Cleve P Reassessing the methods of medical record review studies in emergency medicine research Ann Emerg Med 2005 45 4 448 451 10.1016/j.annemergmed.2004.11.021 15795729

24Harris PA Taylor R Thielke R Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support J Biomed Inform 2009 42 2 377 381 10.1016/j.jbi.2008.08.010 PMC2700030 18929686

25Harris PA Taylor R Minor BL The REDCap consortium: Building an international community of software platform partners J Biomed Inform 2019 95 103208 10.1016/j.jbi.2019.103208 PMC7254481 31078660

26Linzer DA Lewis JB poLCA: An R package for polytomous variable latent class analysis J Stat Softw 2011 42 1 29

27R Core TeamR: A language and environment for statistical computing 2025 Available at: https://www.r-project.org/Accessed March 1, 2025

28Linzer DA Lewis JB poLCA: Polytomous variable latent class analysis, version 1.4 2013 Available at http://dlinzer.github.com/poLCAAccessed March 1, 2025

29Weller BE Bowen NK Faubert SJ Latent class analysis: A guide to best practice J Black Psychol 2020 46 4 287 311

30Lezhnina O Kismihók G Latent class cluster analysis: Selecting the number of clusters MethodsX 2022 9 101747 10.1016/j.mex.2022.101747 PMC9192797 35712641

31Moulin A Evans EJ Xing G Substance use, homelessness, mental illness and Medicaid coverage: A set-up for high emergency department utilization West J Emerg Med 2018 19 6 902 906 10.5811/westjem.2018.9.38954 PMC6225935 30429919

32Eyrich-Garg KM Cacciola JS Carise D Individual characteristics of the literally homeless, marginally housed, and impoverished in a US substance abuse treatment-seeking sample Soc Psychiat Epidemiol 2008 43 10 831 842 10.1007/s00127-008-0371-8 18504513

33Stablein GW Hill BS Keshavarz S Homelessness and substance use disorders Ritchie EC Llorente MD Clinical Management of the Homeless Patient: Social, Psychiatric, and Medical Issues Springer International Publishing2021 179 194

34Brennan JJ Chan TC Hsia RY Emergency department utilization among frequent users with psychiatric visits Acad Emerg Med 2014 21 9 1015 1022 10.1111/acem.12453 25269582

35Baxter AJ Tweed EJ Katikireddi SV Effects of Housing First approaches on health and well-being of adults who are homeless or at risk of homelessness: systematic review and meta-analysis of randomised controlled trials J Epidemiol Community Health 2019 73 5 379 387 10.1136/jech-2018-210981 PMC6581117 30777888

36Lusk HM Shaku D Hemrajani A Housing First: Harm reduction at the intersection of homelessness and substance use Hawaii J Health Soc Welf 2022 81 12 Suppl 3 6 11 PMC9783810 36660275

37Farkas KJ Romaniuk JR Baranowski M Beyond Housing First: rethinking neoliberal policies impacting homelessness J Public Health Policy 2025 46 1 180 192 10.1057/s41271-024-00537-7 39663385

38Formosa EA Kishimoto V Orchanian-Cheff A Emergency department interventions for homelessness: a systematic review CJEM 2021 23 1 111 122 10.1007/s43678-020-00008-4 33683611

39Zuckerman S Shen YC Characteristics of occasional and frequent emergency department users: Do insurance coverage and access to care matter? Med Care 2004 42 2 176 10.1097/01.mlr.0000108747.51198.41 14734955

40Ondler C Hegde GG Carlson JN Resource utilization and health care charges associated with the most frequent ED users Am J Emerg Med 2014 32 10 1215 1219 10.1016/j.ajem.2014.07.013 25154348

41Vinton DT Capp R Rooks SP Frequent users of US emergency departments: characteristics and opportunities for intervention Emerg Med J 2014 31 7 526 532 10.1136/emermed-2013-202407 24473411

42Cheung PT Wiler JL Lowe RA National study of barriers to timely primary care and emergency department utilization among Medicaid beneficiaries Ann Emerg Med 2012 60 1 4 10e2 10.1016/j.annemergmed.2012.01.035 22418570

43Hsiang WR Lukasiewicz A Gentry M Medicaid patients have greater difficulty scheduling health care appointments compared with private insurance patients: A meta-analysis Inquiry 2019 56 0046958019838118 10.1177/0046958019838118 PMC6452575 30947608

44Capp R Kelley L Ellis P Reasons for frequent emergency department use by Medicaid enrollees: A qualitative study Acad Emerg Med 2016 23 4 476 481 10.1111/acem.12952 26932230

45Allen H Gordon SH Lee D Comparison of utilization, costs, and quality of Medicaid vs subsidized private health insurance for low-income adults JAMA Network Open 2021 4 1 e2032669 10.1001/jamanetworkopen.2020.32669 PMC9377505 33399859

46Rinehart DJ Oronce C Durfee MJ Identifying subgroups of adult superutilizers in an urban safety-net system using latent class analysis: Implications for clinical practice Med Care 2018 56 1 e1 e9 10.1097/MLR.0000000000000628 PMC5406260 27632768

47Birmingham LE Cheruvu VK Frey JA Distinct subgroups of emergency department frequent users: A latent class analysis Am J Emerg Med 2020 38 1 83 88 10.1016/j.ajem.2019.04.029 31023586

48Chiu YM Dufour I Courteau J Profiles of frequent emergency department users with chronic conditions: a latent class analysis BMJ Open 2022 12 9 e055297 10.1136/bmjopen-2021-055297 PMC9528600 36175089