Article

Medicaid beneficiaries who continue to use the ED: a focus on the Illinois Medical Home Network

a b s t r a c t

Objectives: Frequent, nonurgent emergency department use continues to plague the American health care system through ineffective disease management and unnecessary costs. In 2012, the Illinois Medical Home Network (MHN) was implemented to, in part, reduce an overreliance on already stressed emergency departments through better Care coordination and access to primary care. The purpose of this study is to characterize MHN patients and compare them with non-MHN patients for a preliminary understanding of MHN patients who visit the emergency department. Variables of interest include (1) frequency of emergency department use during the previous 12 months, (2) demographic characteristics, (3) acuity, (4) disposition, and (5) comorbidities.

Methods: We performed a retrospective data analysis of all emergency department visits at a large, Urban academic medical center in 2013. Binary logistic regression analyses and analysis of variance were used to analyze data.

Results: Medical Home Network patients visited the emergency department more often than did non-MHN patients. Medical Home Network patients were more likely to be African American, Hispanic/Latino, female, and minors when compared with non-MHN patients. Greater proportions of MHN patients visiting the emergency department had asthma diagnoses. Medical Home Network patients possessed higher acuity but were more likely to be discharged from the emergency department compared with non-MHN patients.

Conclusions: This research may assist with developing and evaluating intervention strategies targeting the reduction of health disparities through decreased use of emergency department services in these traditionally underserved populations.

(C) 2015

Introduction

Frequent emergency department (ED) use continues to present a significant problem for the American health care system [1]. Many fre- quent ED visits are made by patients seeking care for nonurgent issues [1]. Although the ED provides essential care for acute health care prob- lems, it is not effective in providing Preventive care and continuity of disease management better found in primary care environments. For nonurgent health ailments, ED visits can cost the health care system three times the cost of primary care visits for the same problem [2].

? Prior presentations: N/A.

?? Funding sources/disclosures: N/A.

* Corresponding author. Department of Preventive Medicine, Rush University Medical Center, 1700 West Van Buren, Suite 470, Chicago, IL. Tel.: +1 312 942 1923.

E-mail addresses: [email protected] (C.M. Glover),

[email protected] (Y.A. Purim-Shem-Tov), [email protected] (T.J. Johnson), [email protected] (S.C. Shah).

1 Tel.: +1 312 947 0229.

2 Tel.: +1 312 942 7107.

3 Tel.: +1 312 942 7926.

The issue is compounded when patients visit multiple EDs which do not have real-time information about care from other providers [1]. Patients’ medical care may spread across multiple, disjointed medical systems. With the existing disconnect between patients, primary care providers (PCPs), and EDs, providers have become less effective in preventing and managing disease [1].

In 2012, the State of Illinois formed Medical Home Network (MHN)

[3] to serve 170 000 residents enrolled in Medicaid of the South and West Sides of Chicago. Medical Home Network is a Chicago-based non- profit organization uniting local providers with a common goal of improving care coordination and health outcomes for Medicaid recipi- ents. A key component of MHN is linking patients to an exclusive PCP to (1) reduce overreliance on already stressed EDs, (2) provide a contin- uous source of medical care, (3) improve disease management, and

(4) reduce Health care costs. Medical Home Network tracks patient activity throughout the network and electronically connects local pro- viders across settings and organizations through real-time notifications on clinical and other activity. By 2013, Medicaid recipients were enrolled into MHN if they had a PCP who was a member of MHN.

http://dx.doi.org/10.1016/j.ajem.2015.10.011

0735-6757/(C) 2015

Table 1

Basic comparisons between MHN and non-MHN populations for all variablesa

Variables

Non-MHNb

MHNb

Significance

Agec

21.46 +- 17.79

21.49 +- 19.29

.94

Acuityc

% of high-acuity visitisits

0.09 +- 0.26

0.09 +- 0.24

.46

% of medium-acuity visits

0.48 +- 0.46

0.49 +- 0.44

.77

% of low-acuity visits

0.04 +- 0.16

0.03 +- 0.16

.92

Day of the weekc

% of weekday visits

0.73 +- 0.4

0.73 +- 0.38

.76

% of weekend visits

0.27 +- 0.4

0.27 +- 0.38

.76

Time of the dayc

% of visits between 7 AM and 3 PM

0.41 +- 0.45

0.4 +- 0.43

.63

% of visits between 3 PM and 11 PM

0.44 +- 0.45

0.45 +- 0.43

.28

% of visits between 11 PM and 7 AM

0.15 +- 0.33

0.15 +- 0.31

.41

ED dispositionc

% of visits with discharges

0.81 +- 0.36

0.82 +- 0.34

.58

% of visits with admissions

0.13 +- 0.32

0.13 +- 0.3

.36

% of visits with transfers

0 +- 0.04

0 +- 0.03

.83

% of visits with LWBS/AMA/Absconded

0.05 +- 0.18

0.05 +- 0.18

.50

Insurancec

Commercial

166 (3.88%)

146 (3.41%)

.51

Medicare

18 (0.42%)

22 (0.51%)

Medicaid

4048 (94.62%)

4071 (95.16%)

No data/self-pay

46 (1.08%)

39 (0.91%)

Primary care physician statusd

No

21 (0.49%)

20 (0.47%)

.88

Yes

4257 (99.51%)

4258 (99.53%)

Employment statusd

No

3777 (88.29%)

3767 (88.06%)

.74

Yes

501 (11.71%)

511 (11.94%)

Diabetes statusd

No

4034 (94.3%)

4017 (93.9%)

.44

Yes

244 (5.7%)

261 (6.1%)

Hypertension statusd

No

3617 (84.55%)

3606 (84.29%)

.74

Yes

661 (15.45%)

672 (15.71%)

Asthma statusd

No

3368 (78.73%)

3350 (78.31%)

.64

Yes

910 (21.27%)

928 (21.69%)

Cancer statusd

No

4187 (97.87%)

4181 (97.73%)

.66

Yes

91 (2.13%)

97 (2.27%)

Heart problem statusd

No

4091 (95.63%)

4061 (94.93%)

.13

Yes

187 (4.37%)

217 (5.07%)

Sex

Male

1427 (33.36%)

1505 (35.18%)

.08

Female

2851 (66.64%)

2773 (64.82%)

Ethnicityd

Not Hispanic or Latino

2855 (66.74%)

2905 (67.91%)

.25

Hispanic or Latino

1423 (33.26%)

1373 (32.09%)

Raced

White

803 (18.77%)

766 (17.91%)

.54

Black or African American

2676 (62.55%)

2719 (63.56%)

Others

799 (18.68%)

793 (18.54%)

Total samples per group = 4278.

a Comparison between MHN and non-MHN populations for Patient demographic characteristics, Employment status, ED visit details, existing comorbidities, acuity, disposition, PCP status, and insurance status.

b Mean and SD or number and percentage provided.

c t Test conducted.

d ?2 Test conducted.

Despite implementation of MHN, it is unknown if and how the num- ber of ED visits varies according to whether or not a patient belongs to MHN. The purpose of this study is to determine whether number of ED visits differs among MHN patients in comparison to non-MHN pa- tients who visited an ED at a large, urban academic medical center in 2013. We hypothesize that MHN patients will have more ED visits in comparison to non-MHN patients. A secondary goal is to explore which related variables potentially modify the relationship between MHN status and number of ED visits. More specifically, we will explore how patient demographic characteristics, employment status, existing comorbidities, insurance and PCP statuses, details of the ED visit, acuity, and disposition potentially modify the relationship between MHN status and number of ED visits. We are not evaluating an intervention such as patient care coordination but aim to provide baseline characteris- tics of MHN patients when compared with their non-MHN counterparts.

Methods

We analyzed patient data for all ED visits at a large, urban academic medical center from January 1, 2013, through December 31, 2013. Clin- ical and billing data were retrieved from the hospital electronic medical record and data warehouse. Our data set excluded patients with errone- ously entered data and those transferred to labor and delivery as

indicated in ED disposition. This study was reviewed and approved by the organization’s institutional review board as an exempt study.

A dichotomous variable labeled “MHN patient status” served as the primary predictor and indicated whether or not a patient belonged to MHN. The main dependent variable was the total number of ED visits during the study time frame. We also included patient demographic characteristics such as age, sex (male/female), race (white, African American, and other), and ethnicity (Hispanic/Latino and non- Hispanic/Latino) as well as existing comorbidities including asthma, heart problems, diabetes, and hypertension as moderating variables. In addition, employment status and PCP status were also included in the study.

As the analysis was conducted on patient level data, the visit level variables such as acuity, ED disposition (discharged, admitted, trans- ferred, and Left without being seen/absconded), day of the week of the ED visit (Monday through Sunday), time of the day of arrival in the ED (ie, 7 AM-3 PM,3 PM-11 PM, and 11 PM-7 AM) and insurance status (com- mercial/private, Medicare, Medicaid, and self-pay/no data) were con- verted into percentage of total ED visits per variable value for each patient. For example, acuity per ED visit was collapsed from 5 levels to 3, namely, low (ie, nonurgent or minor), medium (ie, moderate), and high (ie, severe illnesses). Thereafter for each patient, we calculated the percentage of total ED visits that were of low acuity (ie, number of visits with low acuity/total number of visits). Thus, if a patient had 5

Table 2

Descriptive statistics and bivariate results for the outcome of total number of emergency visits by MHN status and other predictorsa

Variables

No. of ED visits,

mean +- SD or correlation coefficient

Significance

No. of ED visitsb

Non-MHN

1.63 +- 2.13

b.001?

MHN

1.88 +- 1.66

Primary care physician statusb

No

1.59 +- 1.2

.58

Yes

1.75 +- 1.92

Employment statusb

No

1.76 +- 1.95

.74

Yes 1.73 +- 1.65

Diabetes statusb

No

1.7 +- 1.82

b.001?

Yes

2.57 +- 2.94

Hypertension statusb

No

1.63 +- 1.69

b.001?

Yes

2.39 +- 2.76

Asthma statusb

No

1.66 +- 1.48

b.001?

Yes

2.1 +- 2.98

Cancer statusb

No

1.74 +- 1.91

.002?

Yes

2.26 +- 2.29

Heart problem statusb

No

1.71 +- 1.8

b.001?

Yes

2.6 +- 3.37

Sex

Male

1.73 +- 1.88

.33

Female

1.77 +- 1.94

Ethnicityb

Not Hispanic or Latino

1.86 +- 2.2

b.001?

Racec,d

Hispanic or Latino

White

1.54 +- 1.1

1.63 +- 1.26

b.001?

Black or African American

1.87 +- 2.23

Others

1.49 +- 1.11

Insuranced

Commercial

1.59 +- 1.2

Medicare

1.96 +- 1.51

.11

Medicaid

1.76 +- 1.95

Agee

No data/self-pay

1.39 +- 1.32

0.14

b.001?

Acuitye

% of High Acuity visits

0.09

b.001?

% of medium-acuity visits

0.03

0.02?

% of low-acuity visits

0.01

.34

Day of the weeke

% of weekday visits

0.01

.33

% of weekend visits

-0.01

.33

Time of daye

% of visits between 7 AM and 3 PM

0.00

.84

% of visits between 3 PM and 11 PM

-0.01

.35

% of visits between 11 PM and 7 AM

0.01

.30

ED dispositione

% of visits with discharges

-0.11

b.001?

% of visits with admissions

0.11

b.001?

% of visits with transfers

0.02

.03?

% of visits with LWBS/AMA/Absconded

0.02

.09

a Descriptive statistics and bivariate results for the outcome of total number of emergency visits by MHN status, patient demographic characteristics, employment status, ED visit details, existing comorbidities, acuity, disposition, PCP status, and insurance status.

b t Test conducted.

c Analysis of variance with post hoc Bonferroni analysis was performed.

d Black or African American different with both white and Others.

e Correlation analysis was performed.

* Statistically significant variables.

ED visits and 3 were of low acuity, then the percentage of low-acuity visits equaled 3/5, or 60%. Similarly, separate variables were created to account for percentage of total ED visits that were of medium and high acuity.

To ensure that the MHN and non-MHN patients were similar, we conducted propensity score matching. We ran a binary logistic regres- sion analysis with MHN status as the dependent variable and all other variables listed above, except total number of ED visits as independent variables to generate the propensity score. Based on actual MHN status, predicted MHN status and propensity score (ie, score +-0.1); MHN pa- tients were matched with non-MHN patients. Medical Home Network patients without appropriate matching non-MHN patients were exclud- ed from further analysis. To test whether matching was appropriately performed, ?2 and independent-samples t tests between MHN and non-MHN groups with all moderating variables were conducted.

On the matched samples, we conducted independent-samples t tests and analysis of variance with post hoc Bonferroni tests, as well as correlation analysis to investigate the relationship between various predictor and moderating variables with number of ED visits. We used Linear regression analysis to investigate the effects of MHN status on total number of ED visits after controlling for the moderating variables.

All data analyses were performed using IBM SPSS, version 18 (IBM SPSS, Armonk, NY).

Results

For the study time frame, there were a total of 64 077 ED visits with 37 342 individual or unique patients. Of these patients, 4326 (11.6%) were MHN patients and 33 016 (88.4%) were non-MHN patients. After performing the propensity score matching, 4278 MHN patients (98.89% of all MHN patients) along with matched non-MHN patients were included for further analysis. Medical Home Network-and non- MHN-matched patient groups were not significantly different when compared with all moderating variables: (1) percentage of low-, medium-, and high-acuity visits; (2) percentage of visits during 3 PM- 11 PM and 11 PM-7 AM; (3) percentage of ED Visit disposition as dis- charge, admitted, and transferred; (4) PCP and employment statuses;

(5) race; (6) ethnicity; (7) sex; and (8) all existing comorbidities (Table 1). MHN patients (mean [SD] ED visits, 1.88 [1.66]) had significantly more ED visits (ie, 0.25 visits) to the same ED in the study time frame than matched non-MHN patients (mean [SD] ED visits, 1.63 [2.13]; P b .001). The number of ED visits in the study time frame was

statistically significant for all existing comorbidities, ethnicity, race, age, percentage of visits with high acuity, percentage of visits with medium acuity, percentage of visits with discharges, percentage of visits with admissions, and percentage of visits with transfers (Table 2). If patients were Hispanic or Latino, the number of ED visits decreased by 0.32 (1.54 vs 1.86), whereas being African American increased the number of ED visits by 0.24 (1.87 vs 1.63) as compared with whites. With diabetes, the number of ED visits increased by 0.87 (2.57 vs 1.7). Percentage of discharge visits was negatively correlated (r = -0.11) with the total number of ED visits.

Linear regression analysis indicated that age, all existing comor- bidities except cancer, percentage of visits with low-Acuity levels, percentage of visits with transfers, and MHN status were statistically

Table 3

Linear regression results to compare total number of ED visits between MHN and non- MHN populationsa

Variables No. of ED visits

Intercept 0.92 (0.08 to 1.75)?

Age 0.01 (0 to 0.01)?

Acuity % of high-acuity visits 0.19 (0 to 0.39)

% of medium-acuity visits 0.04 (-0.06 to 0.14)

% of low-acuity visits 0.29 (0.03 to 0.54)?

Day of the week % of weekend visits -0.01 (-0.11 to 0.09) Time of the day % of visits between 7 AM to 3 PM -0.04 (-0.13 to 0.06)

% of visits between 11 PM to 7 AM 0.07 (-0.06 to 0.2) ED disposition % of visits with discharges -0.01 (-0.56 to 0.53)

% of visits with admissions 0.31 (-0.24 to 0.87)

% of visits with transfers 1.27 (0.03 to 2.51)?

significant in determining the total number of ED visits (Table 3). Med- ical Home Network patients had 0.25 (confidence interval [CI], 0.17- 0.33; P b .05) additional ED visits as compared with matched non-

MHN patients. Similarly, patients who had diabetes, asthma, and heart

Primary care physician status

% of visits with LWBS/AMA/Absconded

0.24 (-0.34 to 0.82)

0.22 (-0.36 to 0.79)

problems had 0.34 (CI, 0.16-0.53; P b .05), 0.32 (CI, 0.22-0.42; P b .05),

and 0.48 (CI, 0.28-0.67; P b .05) additional ED visits, respectively. With every 10% increase in the number of visits with low acuity, the total number of ED visits increased by 0.29 (CI, 0.03-0.54; P b .05). Similarly, with every 10-year increase in age, the total number of ED visits increased by 0.01.

Discussion

Our analysis supplements the knowledge of Medicaid enrollees be- longing to the MHN by describing them in terms of frequency of ED use. Although the average number of ED visits for both MHN (mean, 1.88 +- 1.66) and matched non-MHN (mean, 1.63 +- 2.13) patients do not meet the common standard of frequent ED use (>= 4 visits per year) [4,5], MHN patients were more likely to use the ED in comparison to matched non-MHN patients while controlling for patient demographic characteristics, employment status, details of the ED visit, acuity, dispo- sition, PCP status, and insurance status. To gain more understanding of ED use by MHN patients, we conducted a pilot qualitative study (the completed qualitative findings are currently being analyzed for manu- script publication). Medical Home Network patients answered inter- view questions regarding why they visit the ED. Preliminary analyses found that MHN patients were familiar with the ED and liked the care that they have received. In addition, MHN patients explained that they did not feel engaged in care with their MHN PCP. Although we did not interview non-MHN patients, our preliminary qualitative findings offer the MHN patient perspective and may offer ideas as to why MHN patients visit the ED more than matched non-MHN patients.

In addition, MHN patients with existing diagnoses of asthma, diabe- tes, heart problems, and hypertension were more likely to use the ED in comparison to their matched non-MHN counterparts. The MHN pa- tients, who continued to use the ED, arrived with lower acuity status and had higher percentages of transfers outside of the institution. How- ever, we did not specifically explore the role of Psychiatric diagnoses in patients who were transferred outside the ED.

Previous research has shown 4 themes related to frequent, nonur- gent ED use. First, Medicaid-eligible and Medicaid-insured patients are more likely to visit the ED for health care compared with privately in- sured patients [6]. Second, people living with chronic physical condi- tions, such as diabetes [7] and asthma [8], also present more often in the ED when compared with those not having these ailments. Third, people with symptoms related to substance abuse [2] and Mental illness [6,9] are more likely to frequent the ED compared with other individ- uals. Finally, despite a common assumption that people who frequent the ED do not have primary care, these individuals exhibit patterns of frequent but discontinuous use across health care environments such as PCPs [10,11]. With new Medicaid initiatives such as MHN, the profile of the current study’s MHN patients who continue to use the ED mirrors that of known Medicaid populations as described in the above literature.

Employment status -0.07 (-0.2 to 0.06)

Comorbidities Diabetes status 0.34 (0.16 to 0.53)? Hypertension status 0.31 (0.17 to 0.45)?

Asthma status 0.32 (0.22 to 0.42)?

Cancer status 0.08 (-0.2 to 0.36)

Heart problem status 0.48 (0.28 to 0.67)?

Race Black or African American 0.13 (-0.02 to 0.28)

Others -0.12 (-0.25 to 0.02)

Ethnicity Hispanic or Latino -0.03 (-0.18 to 0.12)

Sex Female -0.01 (-0.1 to 0.08)

Insurance Medicare -0.37 (-1.02 to 0.29)

Medicaid 0.11 (-0.09 to 0.32)

No data/self-pay -0.25 (-0.7 to 0.19)

MHN patient status 0.25 (0.17 to 0.33)?

a Linear regression results to compare total number of ED visits between MHN and non- MHN populations while controlling for patient demographic characteristics, acuity, disposition, ED visit details, PCP status, and insurance status.

* Statistically significant variables (P b .05).

Once we matched patients according to existing comorbidities, all except cancer resulted in increased ED use across all patients. Hence, MHN and other providers of Medicaid enrollees may also consider addi- tional resources for particularly vulnerable patients such as those diag- nosed as having asthma, diabetes, heart problems, and hypertension. One way this is currently being addressed is through a health disparities intervention conducted in the ED where all children with asthma are provided with asthma-specific action plans in conjunction with follow-up home visits by patient care coordinators. A similar interven- tion may be developed for MHN patients where, once stabilized in the ED and discharged, patients’ PCPs may consistently institute disease- specific action plans to facilitate a reduction in patients’ subsequent visits to the ED [12].

Patient care coordinators based in EDs and primary care offices may also represent an essential link for underserved patients, thus reducing the need for Subsequent ED visits. In response to current study results, ED-basED patient care coordinators consult with patients during their ED visits to assist with follow-up visits with their PCPs. The addition of patient care coordinators based in the community may identify commu- nity resources related to health care services and serve to connect the patient, the MHN, and designated PCPs. Patient care coordination inter- ventions can continue to reduce health disparities through decreased use of ED services in these traditionally underserved patient populations.

Limitations

We did not include psychiatric diagnoses in our analyses. Understanding the role of psychiatric diagnoses in ED visits made by MHN patients will assist with the determination of necessary resources for various MHN patients. In addition, we were unable to assess

whether patients visited other EDs; hence, we do not know the extent of patients’ overall ED use.

Conclusions and next steps

Despite belonging to MHN, patients continued to use the ED and more frequently when compared with their non-MHN counterparts. In- terventions such as patient care coordination and disease-specific ac- tion plans can be essential to better care for MHN patients while reducing health care costs. After intervention implementation, the MHN patient population should be reevaluated to see if any interven- tions decreased ED use. Another next step would be to understand the patient perspective regarding facilitators of ED use and barriers to primary care through qualitative methods. By expanding knowledge regarding the correlates of ED use in the context of the MHN, we may inform researchers and practitioners implementing interventions and practices to reduce unnecessary ED use.

Acknowledgments

This study has grant support from the Building Healthy Urban Communities Project, which is funded by BMO Harris Bank. This study also is part of the Rush Center for Urban Health Equity, which is funded by the National Institutes of Health through the National Institute for Heart Lung and Blood, Grant No. 1P50HL105189-01. The content of this manuscript is solely the responsibility of the authors and does not

necessarily represent the official views of the National Institutes of Health, National Institute for Heart Lung and Blood, or BMO Harris Bank.

References

  1. LaCalle E, Rabin E. frequent users of emergency departments: the myths, the data, and the policy implications. Ann Emerg Med 2010;56(1):42-8.
  2. Simonet D. Cost reduction strategies for emergency services: insurance role, practice changes and patients accountability. Health Care Anal 2009;17(1):1-19.
  3. http://mhnchicago.org/overview.html.
  4. Byrne M, Murphy AW, Plunkett PK, McGee HM, Murray A, Bury G. Frequent attenders to an emergency department: a study of primary Health care use, medical profile, and psychosocial characteristics. Ann Emerg Med 2003;41(3):309-18.
  5. Hansagi H, Olsson M, Sjoberg S, Tomson Y, Goransson S. Frequent use of the hospital emergency department is indicative of high use of other health care services. Ann Emerg Med 2001;37(6):561-7.
  6. Capp R, Rosenthal MS, Desai MM, Kelley L, Borgstrom C, Cobbs-Lomax DL, et al. Characteristics of Medicaid enrollees with frequent ED use. Am J Emerg Med 2013;31(9):1333-7.
  7. Liu X, Song P. Is the association of diabetes with uncontrolled blood pressure stron- ger in Mexican Americans and blacks than in whites among diagnosed Hypertensive patients? Am J Hypertens 2013. http://dx.doi.org/10.1093/ajh/hpt109.
  8. Gold LS, Yeung K, Smith N, Allen-Ramey FC, Nathan RA, Sullivan SD. Asthma control, cost and race: results from a national survey. J Asthma 2013;50(7):783-90.
  9. Doupe MB, Palatnick W, Day S, Chateau D, Soodeen RA, Burchill C, et al. Frequent users of emergency departments: developing standard definitions and defining prominent risk factors. Ann Emerg Med 2012;60(1):24-32.
  10. Baicker K, Taubman SL, Allen HL, Bernstein M, Gruber J, Newhouse JP, et al. The Oregon experiment–effects of medicaid on clinical outcomes. N Engl J Med 2013;368:1713-22.
  11. Chen AH, Rittenhouse D. The un-managed system of Medicare referrals. J Gen Intern

    Med 2012;27(5):487-9.

    Gill JM, Mainous AG, Nsereko M. The effect of continuity of care on emergency department use. Arch Fam Med 2000;9:333-8.

Leave a Reply

Your email address will not be published. Required fields are marked *