The Impact of the Critical Care Resuscitation Unit on Quaternary Care Accessibility for Rural Patients: A Comparative Analysis

Background Previous research suggests that patients from rural areas who are critically ill with complex medical needs or require time-sensitive subspecialty interventions face worse healthcare outcomes and delays in care when compared to those from urban areas. The critical care resuscitation unit (CCRU) at our quaternary care center was established to expedite the transfer of critically ill patients or those who need time-sensitive intervention. This study investigates if disparities exist in treatments and outcomes among patients transferred to the CCRU from rural versus urban hospitals. Methods This is a retrospective study of adult, nontrauma patients admitted to the CCRU via interhospital transfer from outside facilities from January 1 to December 31, 2018. Patients transferred from within our institution or with missing clinical data were excluded. Multivariable logistic regressions were performed to measure the association between patients' demographic and clinical factors with in-hospital mortality. Results We analyzed 1381 nontrauma patients, and 484 (35%) were from rural areas. Median age was 59 [47–69], and 629 (46%) were female. Median sequential organ failure assessment was 3 ([1–6], p=0.062) for both patients transferred from urban and rural hospitals. There was no significant difference between groups with respect to most demographic and clinical factors, as well as types of interventions after CCRU arrival, including emergent surgical interventions within 12 hours of arrival at the CCRU. Rural patients were more likely to be transferred for care by the acute care emergency surgery service than were patients from urban areas and were transferred over a significantly greater distance (difference of 53 kilometers (km), 95% CI: –58.9–51.7 km, P < 0.001). Transfer from rural areas was not associated with increased odds of in-hospital mortality (OR: 0.90, 95% CI: 0.60, 1.36; P=0.63). Conclusion Thirty-five percent of patients transferred to the CCRU came from rural areas, which house 25% of the state population of Maryland. Patients transferred from rural counties to the CCRU faced greater transport distances, but they received the same level of care upon arrival at the CCRU and had the same odds of in-hospital mortality as patients transferred from urban hospitals.


Introduction
Te provision of advanced medical care to critically ill patients poses signifcant challenges, especially for those residing in rural areas [1].Disparities in healthcare access between urban and rural regions are well documented and often result in delayed or suboptimal care for rural populations [2].In the context of this study, rural areas are defned by lower population density and greater distance from healthcare facilities, while urban areas are characterized by higher population density and proximity to comprehensive healthcare resources [3].Tis disparity is particularly pronounced in access to specialized critical and quaternary care resources, which are usually concentrated in urban centers [2].Interhospital transfers over signifcant distances are frequently required to connect rural patients to these resources, as oftentimes transfer to tertiary care centers leads to reduced mortality [4], and the time needed to coordinate and complete these transfers can have signifcant implications for patients with critical and time-sensitive conditions.
Maryland is a relatively small state, and thus its rural areas are in much closer proximity to urban resources compared to larger and more rural states in the US.However, rural populations in smaller states, primarily along the East Coast, still face signifcant barriers to accessing advanced medical care.Tese populations are often overlooked in discussions about rural healthcare disparities, which tend to focus on larger, more remote rural areas largely located geographically towards the midwest [2].Tis study specifcally focuses on rural communities in smaller states, highlighting their unique challenges and the importance of addressing healthcare access for this population.
Maryland's diverse geographical landscape presents unique challenges in providing equitable access to advanced medical care.Te state's two quaternary care centers are located in Baltimore, creating signifcant distance barriers for patients in rural areas, particularly on the eastern shore and in the western regions [5].Te eastern shore is separated from Baltimore by the Chesapeake Bay, with the Bay Bridge serving as a critical yet potentially congested route.Te distance from Worcester County to Baltimore is approximately 130 miles, a journey that can be prolonged by trafc conditions.Similarly, patients from Garrett County face a distance of around 180 miles to reach these centers.Tese logistical barriers highlight the need for a well-coordinated system of interhospital transfers and other innovative solutions to bridge the gap in access to specialized medical services [6].Te reliance on Baltimore's quaternary care centers by the state's rural population underscores the importance of addressing these challenges to ensure equitable healthcare access for all Maryland residents.
Te critical care resuscitation unit (CCRU) at the University of Maryland Medical Center (UMMC) was designed to improve Marylanders' access to critical and quaternary care by expediting interhospital transfers, coordinating advanced preparation for emergent procedures, and advancing the provision of specialized critical care from the moment a patient enters the hospital or is picked up by the transport team at an outside facility [7].Tis study examines whether the CCRU reduces disparities in access to specialized medical care and hospital outcomes for patients with specialized or critical care needs transferred from rural settings throughout the state.By investigating the presence of previously identifed rural/urban health disparities within the CCRU's operations, this study aims to determine if the CCRU model can potentially improve accessibility and reduce disparities in healthcare access for Maryland's rural population.Te results of this investigation could shape future strategies for managing interhospital transfers and enhancing the quality of care for patients from rural areas, who often face unique challenges in accessing advanced medical services.

Study Setting.
Te CCRU is a 6-bed intensive care unit (ICU)-based resuscitation unit and is located in the R Adams Cowley Shock Trauma Center at the University of Maryland Medical Center (UMMC) in Baltimore, Maryland.Tis unit is responsible for the triage, initial resuscitation, and coordination of care for patients transferred from other hospitals across the state of Maryland.When a physician from an outside hospital identifes a patient as a potential candidate for advanced interventions (such as advanced mechanical life support or surgical or endovascular therapies) or has critical care needs exceeding the capacity of their current institution, they can consult the CCRU and the relevant specialists through the Maryland Access Center, a 24/7 centralized hub capable of coordinating services across the University of Maryland Medical System.Te patient's appropriateness for transfer and an initial care plan is determined based on a joint conversation between the referring physician, CCRU physician, and specialists; transfer priority is determined by the CCRU physician based on the acuity of the patient's needs, available resources, and other incoming transfers.On arrival to the CCRU, the patient is reassessed by the CCRU team, who admits the patient and identifes any needed workup or interventions, and by the accepting specialist.Patients requiring emergent interventions are rapidly prepared for and transported to interventions as indicated.Certain patient populations, such as those requiring evaluation for venoarterial or venovenous extracorporeal membrane oxygenation (VA or VV ECMO), neurosurgical or cardiac surgery interventions, and mechanical thrombectomy for cerebrovascular accident due to large vessel occlusion, are transferred to UMMC exclusively through the CCRU, where others, such as those requiring subspecialty medical intensive care, are transferred through the CCRU only when a bed is not immediately available to accept the patient in their "destination unit" (such as the medical or surgical ICU).From 2013 to 2018, 25% of all transfers to UMMC occurred through the CCRU [8].Prior research has shown that patients transferred through the CCRU faced lower in-hospital mortality compared to those transferred to a traditional ICU within UMMC.[9].Other specifcs regarding the CCRU physician, advanced practice provider (APP), and nurse stafng have been previously described in detail [10].Te aim of the study is to identify disparities in access to specialized medical care and hospital outcomes among patients transferred to the CCRU from rural versus urban hospitals throughout the state of Maryland.In a prior publication, we demonstrated that the CCRU successfully facilitates the timely transfer and access to care based on patient acuity and time sensitivity [12].Based on this fnding, we hypothesize that patients transferred from rural hospitals would receive comparable treatment and face similar outcomes compared with those from urban hospitals.However, we do expect to see some diferences in the types of conditions, services, and interventions that these two groups are transferred for, based on the local availability of services.Since most laboratory evaluations were part of the standard of clinical care for CCRU patients, we anticipated minimal missing data.

Statistical Analysis.
We did not perform a formal sample size calculation; we anticipated enrollment of approximately 1500 patients for the full calendar year, with an estimated 25% from rural areas (in line with the state's population distribution).We expected this would provide a sufcient sample size to compare urban and rural patients.
Patients' demographic and clinical information were presented using descriptive analyses.Prior to analyses, histograms of continuous independent variables were inspected to determine their patterns of distribution.Continuous independent variables were presented with mean (±standard deviation (SD)) or median (interquartile range (IQR)) according to their distributions and were compared by the t-test or Mann-Whitney U test.Categorical variables were expressed as N and percentage and were compared with chi-square tests.Comparisons of independent variables between groups (urban vs. rural) were also expressed with diferences and their associated 95% confdence intervals (95% CIs).
We conducted multivariable logistic regressions to determine the association between nontrauma patients' demographic (including initial presentation to rural hospitals) and clinical factors and in-hospital mortality.We selected independent variables (Appendix A) a priori as those identifed by previous literature as predictors of in-hospital mortality [13].Results from multivariable logistic regressions were expressed as odds ratio (OR), 95% CI, and p value.Multicollinearity was assessed using the variance infation factor (VIF).Factors with VIF >5 were considered to have a high collinearity and were eliminated from the models.Te goodness of ft of the models was assessed with the Hosmer-Lemeshow analysis, of which p value >0.05 indicated good ft of the data.Te performance of the models was evaluated with the area under the receiver operating curve (AUROC).A model with AUROC approaching 1.0 would indicate excellent discriminatory capability between dichotomous outcomes (survivor vs. nonsurvivor).
Sensitivity analysis was performed using a multivariable ordinal logistic regression with the outcome of patients' disposition at hospital discharge.Discharge dispositions were ranked in the order of 0 (discharge home directly), 1 (any rehabilitation center), 2 (skilled nursing home/facility), and 3 (hospice/death).Results from the ordinal regression were expressed as OR, 95% CI, and correlation coefcient (corr.coef).A positive correlation indicates an increased likelihood of the lowest rank outcome (0, discharge home), while a negative correlation coefcient indicates an increased likelihood of the highest rank outcome (3, dead/ hospice) more likely.
Te majority of missing data pertained to the following laboratory markers: two bilirubin measurements, two white blood cell counts, one hemoglobin measurement, and 220 troponin measurements.Most of these markers are routine for all patients in the CCRU, but troponin is ordered Critical Care Research and Practice primarily for patients with cardiovascular symptoms.Our analysis centered on patient conditions assessed through the SOFA score; since troponin measurements are not included in the SOFA scoring system, their absence did not impact the score.Of the laboratory markers contributing to the SOFA score (bilirubin, white blood cell count, and hemoglobin), only fve values were missing.Tese missing values were imputed as "normal." All descriptive analyses and multivariable logistic regressions were performed with Minitab version 20 (https:// www.minitab.com,State College, Pennsylvania, USA).All statistical analyses with p value <0.05, except the Hosmer-Lemeshow test as discussed above, were considered statistically signifcant.

Demographics.
Tere were a total of 1731 transfer requests during our study period; we included 1381 nontrauma patients in the fnal analysis (Figure 1) who were transferred from other hospitals.Eight hundred and ninety-seven (65%) patients were transferred from urban areas, while 484 (35%) were from rural areas (Table 1).Te median age of the population was 59 (47-69) years, and 629 (46%) patients were female.Most of the patients' demographic and clinical factors did not vary signifcantly between groups (Table 1).However, patients from rural areas were transported over a signifcantly longer ground distance during interhospital transfer to the CCRU (difference of 53 kilometers, 95% CI: 51.7-58.9km, p < 0.001) and faced longer transport times (diference of 33 minutes, 95% CI: 22-45 min, p < 0.001).A higher percentage of patients from rural hospitals were transferred for evaluation and treatment by the acute care emergency surgery service when compared to those from urban hospitals (16% vs 8%, diference of 8%, 95% CI: 4%-12%, p < 0.001) (Figure 2).
While higher percentages of patients transferred from urban areas were treated with continuous infusions prior to arrival at the CCRU (diference of 8%, 95% CI: 3-13%, p � 0.002), there were no signifcant diferences in interventions after CCRU arrival (Table 2(a, b)).Te percentage of patients undergoing emergent surgical interventions within 12 hours of arrival at the CCRU was similar between groups: 23% (205) of urban patients and 25% (123) of rural patients (p � 0.29, Table 2(a, b)).Inhospital mortality rates and hospital length of stay were also similar between groups (Table 2(a, b)).

In-Hospital Mortality.
Te prevalence of in-hospital mortality was 14% among rural patients, compared to 16% among patients transferred from urban areas (p � 0.43), and patients transferred from rural areas did not face higher odds of in-hospital mortality (OR: 0.90, 95% CI: 0.60-1.36,p � 0.63; Table 2(a, b)).However, longer transport distance from the sending facility of transport to the CCRU was also associated with higher odds of unfavorable discharge disposition, including death or discharge to hospice (corr.coef: −0.01, OR: 1.00, 95% CI: 0.99-1.00,p < 0.001; Appendix C).

Discussion
Tis single-center, retrospective study investigated the outcomes of nontrauma patients who were transferred from rural areas within the state of Maryland to a specialized ICUbased resuscitation unit in downtown Baltimore, MD, and compared them with those of patients transferred from urban parts of the state.Our study found no signifcant diference between the two groups with respect to inhospital mortality or hospital length of stay.Te CCRU was designed to expand and expedite access to critical and quaternary care statewide.Prior studies from our group have demonstrated that the unit has increased the number of patients transferred to our institution while decreasing overall times from transfer request to arrival at UMMC, and for patients requiring urgent surgical intervention, to arrival in the operating room, and has decreased mortality for patients transferred through the CCRU when compared to those transferred to a traditional ICU [7,14,15].We cautiously interpret the fndings of this study to suggest that the CCRU may also function to reduce disparities in access to quaternary and specialized critical care faced by residents of rural counties in our state.
Given the highly and increasingly specialized and resource-intensive nature of critical care and subspecialty surgical care, it is not unexpected to see a large urban-rural divide, nor to expect that this divide will continue to deepen.Tis is likely to be compounded by increasing emergency department boarding of critically ill patients, which has been characterized as stretching small critical access hospitals well past their capacity to provide comprehensive and highquality care [1].Prior research has demonstrated that patients presenting to rural hospitals for trauma [8] and a variety of medical conditions [9,10,16,17] are more likely to require interhospital transfer and to face worse outcomes than their urban counterparts.Our fndings further support these claims.While approximately 25% of Maryland's population resides in rural regions, 35% of patients transferred to the CCRU were transferred from hospitals in rural counties.Patients were most often transferred for specialized care and interventions only available at quaternary centers, such as neurosurgery (intracranial hemorrhage), cardiac surgery (acute aortic disease), or neurology interventional radiology (ischemic stroke requiring thrombectomy; Appendix C).A signifcantly higher proportion of patients from rural areas were transferred to be cared for by the acute care emergency surgery service; this directly highlights a likely disparity in access to emergency general surgical care between our urban and rural counties.Tis disparity has been suggested by prior studies as well [18,19].
Rapid and coordinated transfer of rural patients to quaternary and subspecialty centers is a key component to addressing these disparities.Prior studies have shown that, primarily due to longer transport distances and times, patients from rural areas often face worse outcomes than those from urban areas, even when cared for at the same facilities.A study of trauma patients in Western Australia found that patients from rural areas waited an average of 11.6 hours from the time of injury to defnitive care, compared with approximately 1 hour for patients in urban areas [8].Within the U.S., it has been shown that the time from EMS activation to hospital arrival for patients with STEMI was signifcantly higher among patients in rural areas, even after accounting for total mileage [20].Our study found that patients transferred from rural areas faced longer transport distances, and our sensitivity analysis demonstrated that longer transport distance was associated with higher odds of in-hospital mortality.Tis fnding is in line with those of the studies discussed above, which examined the impact of time  and distance from the patient in the feld to defnitive hospital care.However, our study did not identify a disparity in outcomes between patients transferred from rural or urban areas.Tis may refect a lack of adequate power in our study, or the impact of utilization of air transportation and/ or coordination with transport medical providers.
To our knowledge, this is the frst study directly comparing the outcomes of patients from urban and rural areas undergoing interhospital transfer for defnitive care.Prior studies have investigated care provided at and enroute to rural and urban hospitals, the need for IHT among patients at rural and urban hospitals [16,21], or the outcomes more  globally of patients from rural and urban areas at the same tertiary (often urban) medical centers [22].Te comparison of outcomes of patients from rural areas who arrived at a tertiary center via IHT with those of patients from urban areas arriving via direct admission introduces a high risk of bias, as patients arriving via IHT have already been selected as those requiring high care intensity or subspecialty care or intervention, and are thus likely to be more critically ill (and potentially "behind" on their need for intervention) than those directly admitted [23].
Tis study provides an important frst foray into a more applesto-apples comparison of quaternary care and outcomes for patients from rural and urban areas within a single state and suggests that a specialized critical care resuscitation unit may play a role in optimizing that care.

Limitations.
Our study setting and patient population were unique, such that our fndings may not be directly generalizable.While patient transportation within the state of Maryland is subject to limitations stemming from the state's geography and its inclusion of the Chesapeake Bay, which limits access to a large section of the state, it is a relatively small state, and many of these challenges can be (at least partially) alleviated by judicious use of air transportation, weather permitting.Our fndings may not be applicable to areas of the country in which transport distances for rural patients are signifcantly longer than those within our state.Furthermore, our institution is one of two major referral centers within our area, and we are unable to determine how referral and transfer patterns to the other center may difer in comparison to those described here.Te data for our study were from 2018, and in the last year, we had full access to patients' clinical information prior to the COVID-19 pandemic.At that time, the nursing staf for the CCRU was at an optimal level and enabled a higher volume of transfers.As has been the case across the country, stafng levels declined within the CCRU and throughout our institution during the COVID-19 pandemic and in the period immediately following, which may have impacted patient transfers and outcomes.During the COVID-19 pandemic and extending until 2023, the CCRU faced stafng shortages nationwide.As a result of the high acuity, mixed pathology, and unique process of admitting within the CCRU, new patient intakes are and remain restricted to CCRU-trained nurses exclusively.Tis policy meant traveling nurses and other ICU personnel were unable to perform this critical function.Consequently, the number of yearly CCRU admissions  fell signifcantly when compared to 2018.In addition, due to the inability to manage multiple patient intakes simultaneously, the time from request to arrival was prolonged.However, stafng at our institution and within the CCRU specifcally has improved since late 2023, nearing prepandemic levels.Finally, our investigation of patient outcomes ended with hospital discharge, and thus did not account for potential disparities in access to postacute care, including rehabilitation and primary and subspecialty outpatient care.Each of these components of care plays a role in long-term patient recovery and outcomes, and access to these services has been previously highlighted as important contributors of health disparities between urban and rural Americans [13,24,25].

Conclusion
Patients who were transferred from a rural county within the state of Maryland to the critical care resuscitation unit at the University of Maryland Medical Center did not have signifcantly diferent hospital outcomes than those transferred from urban counties, despite facing longer transport distances and times.Patients transferred from rural areas comprised a higher proportion of transfers relative to the overall proportion of rural residents of the state, highlighting potential disparities in local access.Further research is needed to confrm our observations.

Figure 2 :
Figure2: Heat map of rural counties transferring patients to the critical care resuscitation unit (CCRU).Te numbers on the map correspond to the county identity in the table.Heat map was generated from our own dataset for this study using Microsoft Excel.

10 Critical Care Research and Practice Table 6 :
Total number of patients and number of patients with the top 5 diagnoses being transferred to the CCRU per each rural county in the state of Maryland.

Table 1 :
Demographics of the full population of patients transported from rural or urban areas to the CCRU.Bolded values indicate statistical signifcance. 1 Only the top 5 accepting services were listed here.UMMC, University of Maryland Medical Center CCRU, critical care resuscitation unit; CI, confdence interval; SOFA score, the sequential organ failure assessment score; Mon, Monday; Fri, Friday; Sat, Saturday; Sun, Sunday; mmol/L, millimole per liter; ug/uL, microgram per microliter; WBC, white blood cell; g/dL, gram per deciliter; IQR, interquartile range.Critical Care Research and Practice p < 0.001), and higher serum lactate (corr.coef: −0.12, OR: 0.89, 95% CI: 0.84-0.94,p < 0.001; Appendix C).Multivariable logistic regressions demonstrated that each increment in SOFA score at CCRU arrival was associated with 21% increased odds of in-hospital mortality (OR: 1.21, 95% CI: 1.15-1.28,p < 0.001; Appendix C).SOFA score was not signifcantly diferent between patients transferred from urban vs rural areas (Table

Table 2 :
(a) Clinical interventions and outcomes of the full population of patients who were transported from rural and urban areas to the CCRU.(b) Multivariable logistic regression assessing the association between patients' demographic and clinical factors and hospital disposition of dead/hospice.

Table 3 :
List of all variables being used in the multivariable logistic and multivariable ordinal regressions.

Table 4 :
Multivariable logistic regression assessing the association between patients' demographic and clinical factors and hospital disposition of dead/hospice.

Table 5 :
Results from ordinal logistic regression assessing the association between patients' demographic and clinical factors and the likelihood of clinically signifcant discrepancy in the primary outcome of disposition, where 0 � home, 1 � acute rehabilitation center, 2 � skilled nursing home/facility, and 3 � hospice/death.
1Te signifcance is that they are top 5 specialties.Te bold values are signifcant values.