Evaluation of an Innovative “Emergency Department Surge” Program for the Care and Treatment of Patients with Low-Acuity Conditions

Abstract

Background:

Emergency Department (ED) crowding negatively impacts healthcare. Numerous strategies have been created to improve treatment of low-acuity conditions to reduce ED crowding and better utilize ED resources.

Methods:

We evaluated whether our novel “ED Surge” program within our institution’s Care Transition Clinic (CTC), which triages patients with lower-acuity conditions from the ED to our same-day on-site ambulatory CTC, impacted future ED and hospital utilization. Demographics, visit characteristics, and clinical outcomes were evaluated. Bivariate and multivariate analyses assessed associations between patient, visit, and outcome characteristics. Significance was defined as p<0.05.

Results:

The ED Surge program saw 727 unique patients and 762 scheduled visits with 76% completion across 31 months of operation. Patients were significantly younger (p<0.01) and insured with Medicaid (p<0.01) than the general ED population. This population also experienced a significant decrease in 30-day post-appointment ED utilization (0.08 post- versus 0.2 pre- mean visits, p=0.01) than those who did not complete an appointment. We also identified several diagnoses that may be better served in the ED as opposed to the ambulatory setting.

Conclusions:

Innovative clinics such as our “ED Surge” program held within ambulatory care sites such as the CTC can be a useful tool to impact future ED utilization of patients with low-acuity conditions. Care should be taken to design the clinic to an ED’s specific population needs (such as underserved patient groups) and refine it according to how the clinic’s patient outcomes develop to optimize patient care and safety.

Introduction

The emergency department (ED) serves as a critical point of entry to healthcare systems for patients with life-threatening issues. ED crowding has been proven to not only worsen patient satisfaction and logistical operations but also healthcare outcomes and mortality for patients presenting to the ED.[1–5] A contributing cause of ED crowding is related to patient care in the ED for ambulatory-sensitive conditions, as it not only adds patients to the ED patient load, it consumes time and resources of the ED that could be instead dealt with in a lower acuity setting.[1, 4, 6] Patients have been shown to present to the ED with ambulatory-sensitive conditions for a variety of reasons, including limited access to primary care, convenience (e.g., clinics not open 24/7 and many individuals cannot make clinic hours due to work and/or caregiving responsibilities), and need for improved communication with patients about optimal use of the healthcare system.[6, 7]

Healthcare systems have shown to reduce ED crowding through adding extra workforce and fast-tracking care and have simultaneously begun to initiate interventions to reduce ED crowding through increased utilization of outpatient settings. [8, 9] One published example of a program to reduce ED crowding is a program where patients are seen initially in the ED, then get rapid outpatient follow-up in 2–5 days to safely assess the patient’s improvement, avoiding additional ED visits or even inpatient care.[10] Other published examples include general practice urgent care clinics that are located close to EDs that helped to divert low-acuity patients, although these patients were not specifically triaged there.[11, 12] The most similar examples to our Surge Program involve specialized clinics (e.g., COVID-19 or geriatric foci) that work collaboratively with EDs to transfer patients with lower-acuity conditions to an ambulatory setting after their initial evaluation in the ED. However, it is less well known if this approach can work for the general adult population.[13, 14]

In this study, we evaluated whether our innovation of an “ED Surge” program, that aimed to transfer patients with appropriate lower acuity care needs to an on-site, same-day ambulatory setting after initial ED evaluation in our Care Transitions Clinic (CTC), improved outcomes for patients attending their CTC visit. Specifically, we aimed to describe the population that attended the program both in demographic and diagnosis data, define differences in the populations that were triaged and attended the clinic versus those that were triaged and did not attend, and finally determine if specific diagnoses were better served in the ED compared to our program based on post-CTC visit healthcare utilization. This clinic differs from prior examples by providing an alternative triage decision at close proximity for a wide variety of adult conditions. Here we present how we facilitated an opportunity to provide ongoing care and evaluation during the active visit at an alternative ambulatory site for a range of low acuity conditions to prevent ED utilization.

Methods

This prospective cohort study included patients who completed ambulatory care visits in the CTC after being triaged from the Emergency Department (ED) via our “ED Surge” program. This project was formally determined to be quality improvement by our quality improvement determination committee and not human subject research. Therefore, according to institutional policy, the project was not overseen by the Institutional Review Board.

Clinic Model

The CTC is an ambulatory care clinic that was initially established in 2020 to provide testing and care during the height of COVID-19 pandemic. It, however, has evolved to develop and implement clinical programs with the mission of reducing acute care utilization by reducing ED visits and hospitalizations. The clinic was open Monday through Friday from 7:30–3:30 pm (spring 2020 to spring 2022) with expanded hours to 7:00 pm as of spring 2022. The CTC hosts a range of programs in addition to the “ED Surge” program, including an ED follow up program,[10] sickle cell day program, and post-surgical follow-up programs that provide focused ambulatory care to individual populations that would have otherwise required urgent care in the ED.The “ED Surge” program created in 2021, utilizes outpatient clinics, including the CTC, and their clinicians to care for patients deemed appropriate for outpatient-level care (Figure 1). All patients sent to the “ED Surge” program within the CTC initially presented to the ED, were assessed with a medical screening examination according to The Emergency Medical Treatment and Labor Act (EMTALA) [15] by trained clinicians (physicians or advanced practice nurses), and were triaged within the ED based on their acuity and reason for visit. Patients with low acuity conditions flagged by the ED (e.g., respiratory, urinary, skin infections, chronic disease exacerbations, electrolyte abnormalities, dehydration, etc) were then screened by advanced practice nurses (APN) within the CTC based on their chief complaint and preliminary ED triage data, then transferred if clinically appropriate for a same day appointment as soon as the patient could arrive to the CTC. The ED team and/or CTC team screened each patient individually as not all those that fall into the low acuity category of standardized triage systems meet CTC criteria (i.e., stable vitals, enough time to complete visit before CTC closing, agrees to transfer, and available resources and expertise to provide required care). Resources that the CTC is able to provide and/or obtain include: medications (oral, intravenous, intramuscular), vaccinations, intravenous fluids, blood product transfusions, wound care, ECGs, x-rays, some CTs; we cannot obtain specialty consultations or MRIs and we do not have telemetry. Patients with and without established primary care relationships were eligible for the “ED Surge” program when ambulatory care was clinically appropriate. Patients were usually transported from the ED to the CTC by hospital staff with rare exceptions for patients well enough and with familiarity on how to get to the clinic themselves. After completing the CTC-based visit, patients were then transitioned back to their existing continuity care clinicians and/or set up with further community-based continuity care by our patient advocates, however the data on handoffs to continuity care are out of scope for this analysis.

Figure 1.

Clinical Flow of the CTC ED Surge Program

Main Outcomes

For this study, ED Surge Clinic patients were compared to the general ED population only within demographic data. The bulk of the clinical and diagnostic comparisons were based on the population that was triaged to the CTC, comparing the attended versus not attended populations. This was done as we felt the triage population represented such a distinct population from the general ED that it would be useless to compare the ED Surge Clinic population back to the general ED population.

1. Describe the ED Surge Clinic Population through demographic, visit and diagnostic data. Compare demographics to the general ED population.

2. Define differences in the populations of patients that were triaged to the ED surge clinic and attended their visit versus those that were triaged but did not attend and high/low acuity appointments in both demographic and post-CTC clinic ED/Hospital Utilization.

3. Determine if certain diagnoses are better addressed in the ED rather than our ED Surge Clinic based on post-CTC clinic visit ED/Hospitalization

Data Collection and Measurements

Data from each program visit were extracted from electronic medical records (EMR) from March 2021 until July 2023. Extracted data included patient age, gender (male, female), race (Black/African-American, White, American Indian or Alaskan Native, Asian/Mideast Indian, More than one race, Declined), ethnicity (Hispanic or Latino, Not Hispanic or Latino, Declined), insurance status (no insurance, Blue Cross Blue Shield, Medicaid, Medicare, private payors), insurance network coverage (out of network or in-network for UCM care), appointment status (no show, completed, canceled by clinician, canceled by patient, canceled due to scheduling error), disposition from the CTC visit (home, returned to the ED, directly admitted) and dates of ED visits and inpatient admissions that occurred before the CTC visit. We chose to compare patients that that completed versus did not complete their CTC visit to compare two populations that likely had the same clinical characteristics as they were triaged for a low acuity condition. ED visit demographics including age, gender, race, ethnicity, and insurance status were also collected during this same time frame for comparison to illustrate the demographic population that the “ED Surge” program or similar programs could benefit. Post-CTC visit acute care utilization data were also collected regarding the number and dates of ED visits and/or hospitalizations that occurred within the 30-day period after the ED Surge visit. All patients’ charts were then manually reviewed for patients’ primary CTC visit diagnosis. The diagnoses used in this manuscript reflect the completed appointment diagnosis, not the diagnosis at triage. Each completed visit had up to three diagnoses that aligned with 29 different diagnosis categories. Diagnosis categories were amended from previously published lists of ambulatory-sensitive conditions to align with the wide range of both higher and lower-acuity conditions seen in the CTC.[16–20] These visits were categorized after the visit occurred and the ambulatory-sensitive condition category was not a part of the triage decision. A full list of diagnoses used with highlighted changes to previously described lists is reported in Supplementary Table 1.

Data Analysis

Frequencies were calculated for all demographic variables except for age, for which mean and standard deviation (SD) were calculated.

Significant differences between groups were assessed with either Pearson’s or Fisher's Exact chi-squared (categorical variables) or 2-sample t-test analysis (continuous variables) using 2-sided for demographics and 1-sided for post-CTC acute utilization. Additional dichotomous variables were created for visit status (complete appointments versus no-show and patient canceled appointments) and disposition acuity (high disposition acuity: the patient was sent back to the ED or directly admitted to the hospital from their ED Surge visit versus low-disposition acuity: discharged home). Logistic regression evaluated associations between disposition acuity and diagnosis. Both adjusted and unadjusted analysis was used, adjusting for the demographic variables of gender, race, ethnicity, age, insurance, out-of-network status and if they had had an inpatient stay or ED visit in the last year prior to their CTC visit. Survival analysis with Kaplan Meier analysis and Log-Rank testing evaluated significant differences existing between groups when examining days until the next ED visit (failure=ED visit) and days until the next inpatient admission (failure=hospital admission) limited to 30 days after their ED Surge visit. Stata data analysis software (College Station, TX) was used for analyses. Power calculations were done for all comparisons using the known incidence of ambulatory sensitive conditions in an ED and then Cohen’s guidelines for effect size for the last. [21]. Reporting followed the Standards for QUality Improvement Reporting Excellence (SQUIRE) guidelines for reporting quality improvement studies.[22]

Results

Patient and Visit Characteristics

The CTC-based ED Surge program had a total of 727 unique patients and 762 scheduled ED Surge visits between March 2021 and July 2023.

Patients

Of the 727 unique patients (Table 1), the mean age was 35.2 years (SD 15.0, min=16, max =93). Most patients were female (59.8%), identified as Black race (86.5%), were not of Hispanic, Spanish, or Latino/a ethnicity (90.5%), were insured with Medicaid (64.1%) and had out of network insurance coverage for their UCM care (76.5%). Many of them also had been to the ED (41.4%) or had an inpatient stay (25.4%) within the last year before their CTC encounter. For comparison, the population that visited the ED during this same time had an average age of 46.0 years (SD 20.0, min=16, max=122), most patients were female (58.0%), identified as Black/African American race (81.7%), were not of Hispanic, Spanish, or Latino/a ethnicity (91.9%), and were insured with Medicaid (47.9%). The populations differed based on age (p<0.01), race (p<0.01), ethnicity (p<0.01), and insurance coverage (p<0.01), but not gender (p=0.3)

Table 1.

ED Surge Program compared to Emergency Department visit patient demographics from March 9, 2021 to July 13, 2023

Variable	No. of ED Surge Patients, n (%) (n=727)	No. of Emergency Department Patients, n (%) (n=274,455)	p-value
Gender			0.3
Female	435 (59.8)	159,084 (58.0)
Male	292 (40.1)	115,330 (42.0)
Unknown or None of the above‡	0 (0)	41 (0.2)
Race ‡‡			<0.01 * †
Black	629 (86.5)	224,132 (81.7)
White	34 (4.7)	31,450 (11.5)
More than one race	19 (2.6)	5,322 (1.9)
American Indian or Alaskan Native	1 (0.1)	449 (0.2)
Asian/Mideast Indian	2 (0.3)	2,529 (0.9)
None of the above§	1 (0.1)	4,077 (1.5)
Declined	4 (0.6)	1,184 (0.4)
Unknown or Patient Unable to Respond	35 (4.8)	12,124 (4.4)
Missing	2 (0.3)	0 (0)
Ethnicity			<0.01 * †
Not Hispanic, Spanish, or Latino/a Origin	658 (90.5)	252,242 (91.9)
Hispanic, Spanish or Latino/a Origin∥	25 (3.4)	13,624 (4.9)
Declined	5 (0.7)	1,104 (0.4)
Unknown or Patient Unable to Respond	36 (5.0)	5,950 (2.2)
None of the above	3 (0.4)	1,535 (0.6)
Mean age in years (SD)	35.2 (15.0)	46.0 (20.0)
Insurance Status
Medicaid¶	466 (64.1)	131,450 (47.9)	<0.01 *
Blue Cross Blue Shield (BCBS)	75 (10.3)	35,485 (12.9)
None of the above††	70 (9.6)	68,169 (24.9)
Medicare#	62 (8.5)	68,169 (24.9)
Other Private Payors**	54 (7.4)	20,435 (7.4)
Network Coverage			--
In-network for UCM	171 (22.4)	--
Out of Network for UCM	556 (76.5)	--
Prior Health Visit within 1 year			--
At least 1 ED visit	301 (41.4)	--
At least 1 hospitalization	69 (25.4)	--
Missing	1 (0.1)

Categories for the Emergency Department demographics were amended due to the larger number of categories contained in the total ED encounter dataset, Pearson’s or Fischers (^†) chi-squared or 2-sample t-test analysis (for age) used

^*significance defined as p<0.05

^‡‡Race category allowed more than one response

^‡Included combined categories of unknown and none of the above

^§Included categories containing <0.2% of the population and none of the above

^∥Included combined categories of Hispanic or Latino, Other Hispanic or Latino, Mexican or Mexican-American, Puerto Rican, Cuban

^¶Included combined categories of Medicaid Illinois HMO, Medicaid Illinois, Medicaid out-of-state

^#Included combined categories of Medicare, Medicare Advantage, Medicare-Medicaid

^**Included combined categories of United, Aetna, Cigna, General Commercial, Veteran’s administration, workman’s compensation

^††Includes seven categories with <0.1% of the population as well as those with no category listed (presumed uninsured)

Visits

Of the 762 scheduled visits of triaged patients to the CTC, 580 (76.1%) were completed (Table 2). Of the completed visits, 22 (3.8%) were considered “high disposition acuity” visits, with patients being directly admitted to an inpatient admission or sent back to the ED for further evaluation due to the severity of their complaints. Visits were considered incomplete if they were completed in the ED before transfer to the CTC (canceled by clinician), the patients left the ED before being seen in the ED (no show), they preferred to forgo care rather than be transferred to the CTC (canceled by patient), there was a scheduling error, or an employee firewall that prevented us from determining the visit status. Within the 30-day period after the ED Surge visit, 40 (6.9%) patients returned to the ED a mean of 1.2 times (SD 0.6 min=1, max=3) and 9 (1.5%) returned to be admitted a mean of 1 time (SD 0, min=0, max=1).

Table 2.

ED Surge Program Visit Characteristics

Characteristic	No. of Visits, n (%)
Visit Status (n=762)
Completed	580 (76.1)
Canceled by Clinician	81 (10.6)
No Show	21 (2.8)
Canceled by Patient	40 (5.3)
Canceled due to Scheduling Error	16 (2.1)
Employee Firewall	24 (3.2)
Completed Visit Characteristics (n=580)
Directly Admitted	12 (2.1)
Sent Back to ED for Further Workup	10 (1.7)
Post-Visit 30-Day Acute Care Utilization (n=580)
Emergency Department (SD)	40 (6.9)
Inpatient Admission (SD)	9 (1.6)

Patients who were triaged to the CTC and completed their appointments versus those who were triaged but did not show or canceled their appointments differed in insurance coverage (p=0.03) (Table 3). Of patients who completed their appointment, a higher proportion were insured by Medicaid or other private payers compared to those who had an incomplete visit; however, this did not differ by out of network or in network for UCM status. Otherwise, there was no difference in race, ethnicity, gender, or age or if they had visited the ED or had an inpatient admission in the last year. Patients who completed their ED Surge visit presented to the ED less frequently in the 30 days following their visit (0.08 mean visits versus 0.2, p=0.01), but there were no differences in the number of hospitalizations. Patients who had high disposition acuity ED Surge visits, requiring transfer back to the ED or direct admission to the hospital, were not different in age, gender, race, ethnicity, insurance, or network coverage or having a visit to the ED within 1 year. However, they were more likely to have a hospitalization in the last year. (Supplementary Table 2).

Table 3.

ED Surge Program completed appointments by patient characteristics

Characteristic	No. of Completed Visits, n (%) (n=580)	No. of Incomplete Visits, n (%) (n=61)	p-value
Gender			0.4
Female	349 (60.2)	33 (54.1)
Male	231 (39.8)	28 (45.9)
Race ‡			0.5†
Black/African American	508 (87.6)	52 (85.2)
White	23 (3.9)	1 (1.6)
More than one race	14 (2.4)	2 (3.3)
American Indian or Alaskan Native	0 (0)	0 (0)
Asian/Mideast Indian	2 (0.3)	0 (0)
None Listed	1 (0.2)	0 (0)
Declined	3 (0.5)	1 (1.6)
Unknown or Patient Unable to Respond	29 (5.0)	5 (8.2)
Missing	0 (0)	0 (0)
Ethnicity			0.4†
Not Hispanic, Spanish or Latino/a Origin	530 (91.4)	53 (86.9)
Hispanic, Spanish or Latino/a Origin	15 (2.6)	2 (3.3)
Declined	5 (0.9)	0 (0)
Unknown or Patient Unable to Respond	30 (5.2)	6 (9.8)
Missing	0 (0)	0 (0)
Mean age in years (SD)	34.9 (15.1)	37.0 (15)	0.3
Insurance Status			0.03 † *
Medicaid	388 (66.9)	32 (52.5)
Blue Cross Blue Shield (BCBS)	57 (9.8)	8 (13.1)
No Insurance Coverage	48 (8.3)	11 (18.0)
Medicare	46 (7.9)	8 (13.1)
Other Private Payors	41 (7.1)	2 (3.3)
Network Coverage			0.4
In-network for UCM	128 (22.1)	16 (26.2)
Out of Network for UCM	452 (77.9)	45 (73.8)
Prior Health Visit within 1 year
At least 1 ED visit	249 (42.9)	28 (45.9)	0.6
At least 1 hospitalization	45 (7.8)	9 (14.8)	0.06
Missing	1 (0.2)	0 (0)_

Pearson’s and Fisher’s (^†) Chi-squared or 2-sample t-test analysis (for age) used

^*significance defined at p<0.05

^‡Race category allowed more than one response

Diagnoses

The most common diagnosis category were ear, nose, and throat (ENT) infections with 151 visits (26%) and sexually transmitted infections (STI) or testing with 144 visits (24.8%), followed by vaccine-preventable infections with 66 visits (11.4%), and musculoskeletal issues without trauma (MSK) with 58 visits (10%). The other categories each were each less than 10% of visits (Table 4). However, its important to note that most of these categories did not meet the appropriate sample size for appropriate power which is further discussed in our limitations.

Table 4.

Diagnoses Codes of Completed Visits

Diagnosis	No. of Visits Total, n (%) (n=580)	No. Usual Acuity Visits, n (%) (n=558)	No. of High Acuity Visits, n (%) (n=22)	p-value
Ear Nose and Throat (ENT) Infections	151 (26.0)	146	5	0.8
Sexual Transmitted Infections (STI) or Screening	144 (24.8)	143	1	0.02 *
Vaccine-Preventable Infections	66 (11.4)	64	2	0.5
Musculoskeletal Issues (MSK) without Trauma	58 (10.0)	57	1	0.7
Cellulitis or other Skin Infections	48 (8.3)	46	2	0.7
Gastrointestinal Issues	44 (7.6)	38	6	<0.01 *
Kidney or Urine Infection (UTI)	33 (5.7)	31	2	0.4
Minor Trauma	32 (5.5)	30	2	0.4
Dehydration & Electrolyte Issues	23 (4.0)	17	6	<0.01 *
Hypertension (HTN)	18 (3.1)	18	0	>0.9
Asthma	17 (2.9)	17	0	>0.9
Dental Issues	16 (2.8)	16	0	>0.9
Allergies	15 (2.6)	15	0	>0.9
Post-Procedural Care	11 (1.9)	11	0	>0.9
Neurologic Issue	11 (1.9)	10	1	0.4
ENT Issues without infection	9 (1.6)	9	0	>0.9
Eye Issues	9 (1.6)	9	0	>0.9
Preventative Care	6 (1.0)	6	0	>0.9
Malignancy Concerns or Complications	5 (0.9)	3	2	0.01 *
Acute Kidney Injury (AKI)	5 (0.9)	3	2	0.01 *
Psychiatric Issues	4 (0.7)	3	1	0.1
Pneumonia and Bronchitis	4 (0.7)	4	0	>0.9
Congestive Heart Failure (CHF)	4 (0.7)	3	1	0.14
Diabetes Complications	4 (0.7)	2	2	<0.01 *
Angina	2 (0.3)	1	1	0.07
Pregnancy Testing or Diagnosis	2 (0.3)	2	0	>0.9
Chronic Obstructive Pulmonary Disease (COPD)	1 (0.2)	1	0	>0.9
Nutrition	1 (0.2)	0	1	0.04 *

Fisher’s Chi-squared testing used

^*significance defined at p<0.05

Bivariate Analysis

Visits for gastrointestinal issues (p<0.01), dehydration and electrolyte issues (p<0.01), diabetes complications (p<0.01), nutrition issues (p=0.04), acute kidney injuries (AKI) (p=0.01), and malignancy concerns or complications (p=0.01) were more likely to be a high disposition acuity visit. Visits with STI or screening (p=0.02) were statistically significantly less likely to be a high disposition acuity visit. Other conditions including ENT infections, vaccine-preventable infections, and MSK without trauma, among others, were not associated with high disposition acuity visits (p>0.05). (Table 4)

Multivariate Analysis

After adjusting for patient characteristics, STI or screening (OR 0.12; 95%CI [0.02, 0.94] and p=0.04), angina (OR 34.38; 95% CI [1.89, 624.55] and p=0.02), gastrointestinal issues (OR 5.46; 95%CI [1.94, 15.35] and p<0.01), dehydration and electrolyte issues (OR 12.92; 95%CI [3.92, 42.59] and p<0.01), diabetes complications (OR 108.81; 95%CI [7.96, 1486.14] and p<0.01), acute kidney injury (OR 12.56; 95%CI [1.57, 100.82] and p=0.02), and malignancy concerns or complications (OR 15.14; 95%CI [1.87, 122.16] and p=0.01) were significantly associated with disposition acuity using logistic regression. All other diagnoses remained insignificant in both unadjusted and adjusted analyses or did not have any high acuity visits to be analyzed, as shown in supplemental materials (Supplementary Table 3).

Post-CTC Visit ED and Hospital Utilization

Patients with ENT infections (p=0.048), as shown in Figure 2, and neurologic issues (p=0.01) returned to the ED sooner compared to those who had other diagnoses. Additionally, patients with MSK issues without trauma (p=0.02) and post-procedural care (p=0.01) had inpatient admissions sooner than those who had other diagnoses. No demographic criteria were significantly associated with returning to the ED or inpatient stays in the following 30 days after their ED surge program appointment (Supplementary Figures 1–24).

Figure 2.

Kaplan Meier Curve for days until next ED Visit after ED Surge appointment for patients with Ear Nose and Throat (ENT) infections.

Discussion

Our CTC ED Surge program successfully transferred patients with appropriate lower-acuity care needs to our on-site, same-day ambulatory setting after an initial ED evaluation. We saw over 700 patients within the first two years of operation and found that those who completed their “ED Surge” appointment saw a decrease in 30-day ED utilization. This demonstrates a small impact in reduction of future ED visits in addition to same day ED utilization that the ED “Surge” program and CTC clinic provide. As discussed in the introduction, this clinic had the potential for contributing to the improvement of ED crowding, which could impact patient satisfaction, health outcomes and specific financial impacts by reducing total costs to payers, especially in value-based care contracts. Additionally, we also found differences in the population demographics of those who were triaged to the CTC compared to the general ED population and those who completed their appointment after they were triaged compared to those who did not complete it (i.e., no-showed or canceled) after they were triaged. The patients cared for by the CTC ED Surge program reflect a population who are younger and more likely to be Medicaid-insured when compared to prior studies, who may utilize the ED for both ambulatory sensitive conditions and emergent/urgent care needs due to access barriers to primary care, providing an intermediate option for this population who may be experiencing these barriers. Finally, we were able to identify diagnoses that may be better served in the ED as opposed to triage to the CTC such as angina. Importantly, while known angina was an exclusion factor for transfer to the CTC, sometimes this symptom was reported after transfer. Further, at times angina was the presenting symptom to the ED with emergent causes ruled out prior to CTC transfer. With these data, we have refined our program to better serve our patients safely and efficiently. We hope that our experience might inform the development of similar programs to serve other unique population needs.

Patients who completed an ED Surge visit after being triaged to the CTC, when compared to those with an incomplete visit after being triaged, were 2.5 times less likely to return to the ED within the 30-day period after the complete or incomplete visit. This signals that the clinic was effective in caring for patients with low-acuity conditions triaged from the ED. Across the literature, similar clinics have been implemented with conflicting results. While individual programs demonstrated improvements in ED wait times, crowding, and overall utilization [11, 12, 14, 23–26], other published data have shown low-quality evidence and little improvement when evaluated together.[27–29] These conflicting findings are likely the result of these types of clinics being useful when tailored to the particular patient populations treated. Additionally, while not specific to the Surge program, our CTC clinic also connects patients to outpatient primary care if needed, following their appointment. Therefore, some of this decrease in utilization is likely due to improved primary care continuity, though we do not have the data to show that.

Patients who completed their visit after being triaged to the CTC differed significantly in insurance status from those who did not complete their visit after being triaged. A much higher proportion of patients who completed their appointment were insured with either Medicaid or other private payors, which likely contributed to the difference. It is also interesting to note that nearly 30% more Medicaid patients and nearly 60% less Medicare patients were triaged to the CTC compared to the ED population. This difference is also likely reflected by the nearly 10-year age difference between the ED Surge population and the ED. We believe that this is the most likely explanation behind the population differences, with the CTC seeing a larger amount of Medicaid for both completed appointments and when compared to the ED population. With large insurance changes within the United States such as the Affordable Care Act, ED crowding and use have been shown to correlate with improved access to health insurance.[30, 31] Within our regional community, patients insured by Medicaid may be more likely to present to the ED for lower-acuity issues due to myriad of barriers to obtaining ambulatory care. Therefore, our clinic may represent an important community resource, as it provides a type of intermediate care for an underserved population. We also examined the disposition and follow-up care for specific diagnoses. We found that several diagnoses were associated with early return to the ED or direct admission including angina, gastrointestinal issues, dehydration and electrolyte issues, diabetes complications, acute kidney injury, and malignancy concerns or complications. We also identified that ENT infections and neurologic issues returned to the ED sooner and MSK issues without trauma and post-procedural care returned with an inpatient admission sooner than other conditions. Notably, visits for STI or testing were protective. While some diagnoses such as angina or dehydration are clear as to why they could require a higher level of care, specific diagnoses within gastrointestinal issues such as abdominal pain diagnosed as pancreatitis and ENT infections such as ear or peritonsillar abscesses demonstrate how a seemingly ambulatory condition can be more serious in specific circumstances (Supplementary Table 1). As a result, we have continuously refined our triage process and have experienced a decrease in transfers back to the ED with higher sensitivity to these higher-risk diagnoses.

Our project has several limitations. First, our CTC ED surge program has a modest capacity with around 350 patient appointments available. However, there are additional ED surge program clinics that extend the impact of this process, though evaluation of those clinics is outside the scope of this analysis. While only a modest sample size, we found that our impact is not trivial and is meant to serve as one mechanism to address ED crowding, not intended to suggest it as the primary solution to the ED crowding. Given the annual restraints, our modest sample led to some limitations with sub-population analyses. Second, for our race category, patients in the larger hospital group were able to select more than one category, compared to our study dataset which only considered one answer for race. Therefore, the results comparing the differences between the ED population compared to the CTC population may have been skewed by race. Additionally, while race was included as a co-variate, we would like to address that it is a social construct and often when differences are found by race these differences are better explained by confounding factors. However, as our patient population experiences health and health-care related racism due to structural and implicit biases we aim to address some of these factors and included it in our analysis. Third, all patients’ insurances were billed if insured. For this evaluation however, it is out of scope to comment on visit billing/payments. We did not collect patient-reported reasons for incomplete visits or evaluate billing-related barriers. While insurance status was associated with appointment completion, cost and access barriers were not assessed in this study. Fourth, we also did not assess the post-CTC link to primary care in this analysis. Though this is a component to the clinic, the data were not complete enough to include in the manuscript. Fifth, there are likely a range of unobserved confounders including selection bias, indication, health related social needs and social determinants of health that were unable to be appropriately described through our dataset. Finally, our study examined only one health system, where this clinic structure was able to care for a specific patient population—one much younger and with a higher predominance of patients who were of Black race and insured with Medicaid. Therefore, generalizability of our findings may be limited.

In summary, our “ED Surge” program created an innovative transitional clinic that contributed to improved future ED utilization within lower-acuity patients presenting to the ED as well as providing care to a group of patients that may have been otherwise underserved within our community based on the evidence that our clinic was predominantly Medicaid insured. While the clinic examined was small, by providing alternative care paths to offload the ED during times of ED “surge” may have contributed to improving patient satisfaction, health outcomes and decreasing total cost of care to payors. Similar innovative clinics can be a useful tool to offload ED crowding but care should be taken to design the clinic to an ED’s specific population needs and outcomes. Future work can help investigate some of the findings and limitations of this paper, including being able to establish causal relationships, as well as identify specific clinic designs that may be more beneficial to a healthcare system and populations.