Achieving regionalization through rural interhospital transfer

a b s t r a c t

Regionalization of Emergency medical care aims to provide consistent and efficient high-quality care leading to optimal clinical outcomes by matching patient needs with appropriate resources at a network of hospitals. Re- gionalized care has been shown to improve outcomes in trauma, myocardial infarction, stroke, cardiac arrest, and acute respiratory distress syndrome. In rural areas, effective regionalization often requires interhospital transfer. The decision to transfer is complex and includes such factors as capabilities of the presenting hospital; capacity at the receiving hospital; and financial, geographic, and patient-preference considerations. Although transfer to a comprehensive center has proven benefits for some conditions, the transfer process is not without risk. These risks include clinical deterioration, limited resource availability during transport, vehicular crashes, time delays for time-sensitive care, poor communication between providers, and neglect of patient preferences. This article reviews the transfer decision, financial implications, risks, and considerations for patients undergoing rural interhospital transfer. We identify several strategies that should be considered for development of the re- gionalized emergency health care system of the future and identify areas where further research is necessary.

Introduction

In 2006, the Institute of Medicine published a report on the Future of Emergency Care in the United States Health System, which recommended the development of a “coordinated, regionalized, accountable [emer- gency care] system” [1]. According to the conference, “regionalization is an active process by which patients are appropriately matched to ap- propriate resources” [2]. The goal of regionalization is to provide opti- mal care and enhance outcomes [3-6]. This should be compared to centralization, which is a more unplanned process in which patients are transferred to larger medical centers for a variety of medical, finan- cial, and legal reasons. In many cases, centralization is the de facto result of limited service availability and outright closure of Rural hospitals due to economic factors and population shifts. Typically, intentional regionali- zation and uncoordinated centralization processes occur concurrently. Both processes often involve interhospital transfer, and both can improve patient care, health outcomes, and economic efficiency (Fig. 1).

? Meeting presentations: None.

?? Financial support: Dr Mohr is supported by a grant from the Emergency Medicine

Foundation.

? Conflicts of interest: The authors have no financial conflicts of interest to disclose.

* Corresponding author at: University of Iowa Carver College of Medicine, 200 Hawkins Drive, 1008 RCP, Iowa City, IA 52242. Tel.: +1 319 353 6360.

E-mail address: [email protected] (N.M. Mohr).

More than 125 million patients are treated in US emergency depart- ments (EDs) annually. Nationally, 1.5% (8.5 million) of ED patients are transferred to another acute care hospital. In rural America, this transfer rate is doubled [7]. Patients are transferred for a myriad of medical diag- noses via networks of both established and ad hoc hospital relationships [8]. When these transfers match patient needs to necessary resources, then regionalization optimizes medical care; when needs and resources diverge, these transfers constitute waste.

Transfer partnerships between hospitals are both explicit (ie, for- malized networks for specific conditions) and implicit (ie informal prac- tices). Examples of both explicit and implicit regionalized health care networks are illustrated in Table 1. Explicit regionalized systems of care have improved outcomes for specific patient groups such as trauma and ST-elevation myocardial infarction [9,10]. Although these diseases have published triage criteria, many others do not, and identi- fication of patients for transfer remains variable [11].

In this article, we will discuss the medical and financial implications of patient transfer, risks of interhospital transfer, and the future of regional- ized emergency care. We will further identify areas for future research.

Factors affecting decision to transfer

The decision to transfer a patient to another hospital for emergency care is dependent on many factors including the capabilities of the pre- senting hospital, capacity at the receiving hospital, and financial factors.

http://dx.doi.org/10.1016/j.ajem.2015.05.032

Fig. 1. Regionalization vs centralization. These processes occur concurrently and have both overlapping and unique effects.

Capability of presenting hospital

Most transfers are initiated either because of condition-specific rec- ommendations [12,13] or because a provider assesses that a patient re- quires a level of care or specialized service not available at the local hospital [14]. Services frequently referred for interhospital transfer are cardiac revascularization, neurology and Neurosurgery consultation, trauma surgery, and critical care [15]. The growing lack of specialty cov- erage available in rural US EDs and the concentration of specialty ser- vices to Tertiary care centers have made interhospital transfer a critical component of patient management (Fig. 2) [16,17].

Transfer of patients occurs for a wide variety of medical problems, and regionalization of care has been demonstrated to have superior outcomes for several specific conditions. The first efforts at regionaliza- tion of emergency care involved patients with traumatic injuries. Established by the EMS Systems Act in 1973, Trauma systems were seen as a method of “distributing resources more equitably while expanding access to health care systems” and improving national trau- ma care [18]. Multiple studies have demonstrated improved outcomes for patient treated at an American College of Surgeons-certified level I trauma center vs nontrauma centers [19-23]. Based partly on the effec- tiveness of developing robust trauma systems [24], STEMI care was re- gionalized into transfer networks locally to shorten the time to cardiac catheterization–another time-sensitive intervention. Mortality is re- duced when percutaneous coronary intervention (PCI) is performed rapidly in STEMI patients [9,25-28], and guidelines recommend this therapy within 90 minutes of presentation [29,30]. In 2000, the Brain Attack Coalition published guidelines to establish Primary Stroke Cen- ters to improve care of patients with acute ischemic stroke [31]. Orga- nized systems of stroke care [32] and dedicated stroke units have both been shown to improve mortality and functional outcomes. Stroke center care is associated with decreased 1-year mortality of 11% to 38% [33-35]. Improved outcomes have also been demonstrated when

patients with Acute respiratory distress syndrome are trans- ferred to a center capable of extracorporeal membrane oxygenation . Interestingly, the clinical benefit of the transfer appeared to be independent of the use of ECMO, reinforcing that disease-specific care in high-volume centers may have intrinsic value [36]. Following the 2009-2010 H1N1 influenza pandemic, a cohort study of case- matched patients with H1N1-related ARDS demonstrated a significant Mortality benefit to patients transferred from Tertiary care hospitals to one of four ECMO centers in the UK [37].

Rural workforce

Many low-volume rural hospitals are unable to offer comprehensive procedural or Medical specialty services. However, rural hospitals serve the role of stabilizing and beginning time-sensitive therapies (eg, tissue plasminogen activator and antibiotics) for critically ill patients who present to their facility. Many of these EDs lack residency-trained/ board-certified emergency physicians (EPs). More than 70% of counties with less than 20000 residents have no board-certified EPs [38]. Instead, physicians trained in Family Medicine, internal medicine, and advanced practice providers (eg, physician assistants and nurse practitioners) provide much of the care in EDs in these communities. In one rural state, only 12% of EDs were staffed exclusively by EPs, and 61% were staffed by advanced practice providers in solo practice for at least a portion of the week [39]. In addition, a recent report of ED care using a national data set reported 10% lower hospital mortality for patients admitted from high-volume EDs compared with low- volume departments [40]. This combination of non-EP staffing and a low volume of critically ill patients unveils a potential problem for patient safety, particularly for patients with critical illness with time-sensitive interventions.

Pharmacy services

Although not often identified as a primary reason for transfer, Hospital-based services affect transfer decisions as well. For example, clinical pharmacy services have integrated themselves into all aspects of Multidisciplinary teams across the nation. Prior reports have shown that clinical pharmacy services can significantly decrease hospital costs and improve mortality [41,42]. Unfortunately, few hospitals pro- vide onsite 24-hour clinical pharmacists, and most have a pharmacist available for only a few hours per day [43]. With such limited availability of pharmacists, specialized clinical pharmacy services are rare in rural locations [44] and may represent 1 hospital resource that distinguishes high-volume hospitals from low-volume hospitals. Furthermore, rural hospitals that rarely use specialized or time-sensitive medications may not stock these medications for emergent use. Formulary restrictions may limit availability of costly medications, such as some rarely used toxicologic antidotes [45], antibiotics, antihemophilic factors, and Anticoagulant reversal agents. Thus, medication availability can drive transfer of some patients.

Table 1

Transfer networks can be categorized based on their level of organization and support

Classification Definition Examples

Explicit transfer networks A network established for a defined diagnosis or condition that recommends care through prehospital triage or interhospital transfer to a specific medical center based on a prespecified set of criteria and/or center designation. These networks require coordination by a central organization, surveillance and maintenance to optimize their function.

Implicit transfer networks A network that exists in a region and drives regional practice for patients meeting general, informally defined criteria that lead to either prehospital triage or interhospital transfer of patients who exceed the capabilities or specialty resources of a local medical center. These networks are not centrally maintained or governed and may evolve over time based on the resources, capabilities, and ease of transfer of various organizations within the system.

Trauma system STEMI networks Stroke networks

Cardiac arrest

Severe respiratory failure (ECMO) NeuroSurgical emergencies Transplant centers

Critical care services, including neonatal and pediatric

Fig. 2. Rural trauma hospital availability. This map shows trauma hospitals, levels I and II (black) and levels III to V (light gray). Shading shows areas that are within 60 minutes by air or ground transport to a level I or II trauma center. Map courtesy of TraumaMaps.org from the University of Pennsylvania.

Available capacity at receiving hospital

The growing scarcity of beds at tertiary Regional hospitals can lead to significant problems. Patients can wait for days at a referring hospital for an open bed or can be transferred to a full hospital only to spend hours or days waiting for a bed to become available. In some cases, patient volume exceeds capacity and EDs divert ambulance traffic. This may be feasible in urban areas with multiple capable hospitals, but in rural states, diversion is often implausible. Hospitals stretched to capacity are forced to use alternative staffing models and strain surge capacity, even as rural EDs and Hospital beds sit empty.

The problem of capacity is especially pronounced for psychiatric pa- tients. In the 1950s, there were an estimated 339 state hospital psychia- tric beds per 100000 population, but by 2000, that number had reduced to 22 per 100 000 [46]. The relatively long duration of hospitalization for Psychiatric patients also reduces bed turnover, which translates to long waits for hospitalization. A recent study found that the average length of stay for a patient in the ED before transfer to a psychiatric facility was

11.75 hours for transfer to a private psychiatric facility, vs 19 hours for transfer to a public facility. This finding was strongly associated with in- surance status [47]. Often, psychiatric patients board in the ED, which reduces the capacity of the ED to treat new arrivals; lengthens time to treatment; diverts nursing staff; reduces ED revenue; and, most impor- tantly, delays the psychiatric care that these patients need [48].

Emergency medical treatment and Act”>Regulatory and financial factors in interhospital transfer

Emergency Medical Treatment and Active Labor Act

Interhospital transfer of ED patients is mediated, in large part, by le- gislative mandate. The Emergency Medical Treatment and Active Labor Act (EMTALA) was designed as an “anti-dumping law” and revolution- ized the transfer process between hospitals. Before 1986, it was legal for hospitals to refuse to care for patients primarily for economic reasons [49]. Since passage of EMTALA, however, hospitals with EDs are re- quired to evaluate and treat all patients presenting for emergency care and to facilitate transfer for any patients requiring care beyond local ca- pabilities [7,50]. The EMTALA legislated an obligation between a physi- cian, a hospital, and a patient with an emergency medical condition. Hospitals were mandated to provide care, and referral centers were

required to accept transfers of emergency patients regardless of a patient’s ability to pay for services.

Insurance and payment factors

Before EMTALA, several landmark studies revealed disproportionate transfer rates from private to public hospitals, presumably for economic reasons [49,51,52]. Although emergency medical systems and EPs have adopted interfacility transfer protocols since EMTALA, imbalances re- main [53-61]. Uninsured patients transferred to tertiary centers for femur fractures [62] or hip arthroplasty [63] were shown to be unlikely to require tertiary orthopedic care, and this trend may be exacerbated among trauma patients [59]. Furthermore, the disproportionate rate of uninsured patients being transferred to trauma centers places addition- al Financial burden on tertiary centers [54]. In a California study, trans- ferred patients who were expected to be “poor payers” based on the expected level of insurance reimbursement were admitted to public fa- cilities more often than to other hospitals, despite adjustment for the total number of statewide admissions [58]. Studies of minor trauma pa- tients, Minor head injury, and Rural patients have all demonstrated an association between transfer decisions and insurance or payment fac- tors [55,56,64]. In fact, uninsured patients in a national sample had twice the odds of transfer as those with private coverage [65], and a second analysis of neonates almost mirrored that effect [66]. This is an especially important consideration for rural populations, as rural areas have higher proportions of Public insurance compared to urban and suburban areas [67].

Interestingly, however, poor insurance status does not reliably pre- dict poorer outcome–just increased probability of transfer. In one study of major trauma patients, those with private insurance were at in- creased risk for admission without transfer to a dedicated trauma cen- ter. This effect was greatest in those who did not appear to require subspecialty care (especially isolated abdominal trauma) and those who presented to a nontrauma teaching hospital [60]. Although the au- thors did not conduct an analysis of clinical outcomes, prior work sug- gests that trauma center admission is associated with 25% decreased mortality, even when accounting for severity of injury [19]. Other inves- tigators propose that improper risk adjustment is responsible for some allegations of inappropriate transfer [68,57,69].

Table 2

Cost-related factors associated with interhospital transfer

Factor Description Examples Factors associated with health care cost savings

Lower resource utilization at regional hospital Avoiding unnecessary testing and therapy because of

more stringent application of Evidence-based guidelines

Decreased overall costs from shorter length of stay Shortening hospital stay through institutional placement

outside the tertiary care hospital and robust home support services

Improved clinical outcomes Improved functional recovery in high-volume clinical centers that care for large numbers of similar patients

Fewer complications Reduction in hospital complications because of strict care plans and standard treatment algorithms

Avoiding extensive diagnostic testing in patients with congestive heart failure, pancreatitis, or minor traumatic brain injury

Early transition from ICU-level care to long-term acute care placement after traumatic brain injury

Improved functional recovery in stroke centers with aggressive care and therapy

Lower rates of ventilator-associated pneumonia in high- volume ICUs from compliance with care guidelines

Factors associated with increased health care costs

Direct hospital costs Direct billing to patient/insurer by the hospital and providers caring for a patient

Transportation costs Direct billing to patient/insurer by ambulance/ helicopter transport agency for transport and care that is provided during transport

Indirect costs All remaining economic expenses borne by patients and families for which no medical bill is generated.

Equipment, supplies, medication, professional fees, facility fees

Ambulance fees

Family lodging near the tertiary center

Transportation of family members over long distances to visit tertiary center

Inability for family members to work because of travel Discontinuity of care because of poor communication with primary physician following patients after discharge Travel and cost of transportation for follow-up care at the tertiary center, especially for patients who re-establish primary care because of complicated medical problems

Hospital revenue factors

The goal of a transfer network is to minimize the costs of infrastruc- ture while maximizing clinical outcomes [70]. Mortality, length of stay, and hospital charges are higher for Transferred patients than nontransferred patients, even after risk adjustment [71]. However, cost-related factors associated with regionalized healthcare are com- plex (Table 2). Transfer to a tertiary center provides both additional cost (cost of transport) and cost-savings (improvED efficiency of high volume), and the net benefits are difficult to predict for an individual patient. This is an example of a heterogeneous treatment effect.

The theory of heterogeneous treatment effect is that patients will de- rive benefits of different magnitude from a treatment as a function of the severity of their disease. Applied to interhospital transfer, some pa- tients benefit dramatically, whereas others benefit little, if at all (eg, the most critically ill patients may derive more benefit from tertiary care than those with less severe illness). Fig. 3 illustrates this relationship. As interhospital transfer becomes more prevalent for a wider variety of diseases, the patients who are transferred on the margin are likely to be less ill than those that were previously being transferred, diluting the beneficial effect of interhospital transfer for the population. Thus, the overall treatment benefit may be smaller with more significant costs than it would have been for a more severely ill population.

Emergency physicians focus primarily on matching the level of care needed to the severity of illness, rather than the financial costs and ben- efits of transfer. Severity may be challenging to estimate early in the treatment course, however, so providers may be conservative in recommending liberal transfer practices. Notably, although rural hospi- tals may benefit from the revenue of caring for additional patients, indi- vidual providers bear nearly all the medicolegal risk in selecting an appropriate level of care. Furthermore, little of the financial incentive to avoid transfer affects individual providers directly. This may lead to an imbalance in the risk and incentive profile that encourages liberal transfer practices. Such financial and legal issues have not been studied rigorously, but they are important aspects of the transfer decision.

Further complicating the landscape of regionalization is the chang- ing rural health care environment. Starting in the late 1990s, small rural hospitals were transitioned to federally designated critical access hospital (CAH) status. Critical access hospitals must be located in a rural area at least 35 miles from any other hospital (15 miles in an

area with mountainous terrain or secondary roads), be state designated as a “necessary provider of health care services,” support a 24-hour ED, and maintain an average inpatient acute length of stay of 96 hours or less. In return for the CAH designation, hospitals are reimbursed for Medicare patients through cost-based accounting, recognizing that maintaining rural health facilities for a smaller patient base is more cost- ly. This fundamentally changed the financial reimbursement structure for small hospitals and in many ways kept them financially viable. In ad- dition, this payment structure also codified the types of patients who are treated in CAHs: patients with anticipated or actual prolonged length of stay over 96 hours will be transferred for ongoing medical care. More recently, value-based purchasing, capitated reimbursement, and the growth of Accountable Care Organizations have begun to again change how health systems seek to provide cost-conscious care [72-74]. These reimbursement mechanisms in many places have also led to consolidated health systems, in which CAHs formalize relation- ships with large hospital networks to provide a referral base for specialty services at tertiary care centers.

Finally, interhospital transfer can be a tool used to shift cost between

health care systems. Transfer shifts expected costs between health care organizations, as hospitals select which patients they are best prepared to serve. Hospitals may choose to offer or not offer specific health ser- vices (eg, dialysis or mechanical ventilation) to change the overall risk profile of the inpatient panel [75]. As health systems expand and rural hospitals become part of larger networks, these factors may change the role of rural health centers in caring for complex patients. Although transfers may currently be a source of revenue for large hospital sys- tems, under accountable care organizations and capitated reimburse- ment, providing care locally without bearing the cost of transfer for a hospital network may become more financially attractive.

Risks of interhospital transfer

Despite the clinical benefits of regional transfer networks for many medical conditions detailed above, the transfer process itself confers risks. Those risks can be divided into 6 groups: delay in time-sensitive care, transportation risks, clinical deterioration, intratransfer resources, poor handoff communication, and neglected patient preferences (Fig. 4).

Fig. 3. Heterogeneous treatment effects. This is a theoretical model of how patient-oriented costs and benefits relate to transfer prevalence and heterogeneous disease severity. When interhospital transfer is a low-prevalence event (point A) performed only for select patients who will maximally benefit from the resources available at the receiving hospital, then the benefit of transfer is higher than the costs. However, as interhospital transfer becomes more prevalent (point B) for many medical conditions, the average overall medical benefit per patient decreases. Meanwhile, the average per-patient costs of transfer (eg, ambulance expenses, indirect costs to patients/families; demonstrated by horizontal line) stays relatively constant.

Time delay

Transfer is often time intensive. Time is required not only for the transfer duration itself but also for initial triage, identifying a destination hospital, obtaining acceptance by a receiving clinician, completing documentation, and waiting for transportation to be arranged. In rural Illinois, a study of STEMI transfers showed that even with a mean inter- hospital distance of 45 miles and use of helicopter transport for 66% of patients, the median Ischemic time (symptom onset to PCI) was 199 minutes. Much of the time delay (N 30 minutes) was spent in the non- PCI hospital after STEMI diagnosis arranging transfer [76]. Other diseases, such as sepsis, have time-sensitive medical interventions

that can be initiated in any hospital yet are sometimes delayed for transfer [77].

Transportation risks

Although ground and Air travel are routine, inherent risks exist in- cluding motor vehicle accidents, helicopter crashes, and Unavoidable delays. The mortality rate for ambulance workers has been estimated at 9.6 fatalities per 100000 emergency medical services (EMS) workers per year due to traffic accidents [78]. The mortality rate for rotor wing ambulance crew has increased in recent years, being estimated at 164 fatalities per 100000 helicopter EMS crew members in 2008 [79].

Clinical deterioration

The risk of a patient’s illness acutely worsening during transport is a dreaded complication of interhospital transfer. However, clinical deteri- oration is relatively rare. Small studies of patients with ARDS [80], myo- cardial infarction [81], and spinal cord injury [82] did not show a significant rate of transfer-associated complications. In 2009, a study es- timated the risk of a critical patient event during air transport to be 5.1%, and a more recent study measured an ambulance interhospital transfer complication rate of 6.5% [83,84]. These complication rates were consis- tent with similarly defined complication rates during anesthesia [85-87]. Transport duration is a significant predictor of complications. Longer transports increase the rate of complications by 15% per 10-minute in- crement of additional transport by ground (95% confidence interval [CI], 12%-18%), although this effect drops to 2% per 10-minute incre- ment by air transport (95% CI, 1%-3%) [83,84]. This difference between transport modality may reflect differences in EMS provider training and experience, technology available in route, or the types of patients transported by the two modes. Clearly, critically ill patients early in their disease are likely to continue evolving, and the risk of clinical

deterioration is high regardless of their location.

Intratransfer resources

One significant factor associated with complications is training of the transport team. Critical care paramedics were associated with a 62.5% reduction (95% CI, 45%-91%) in complications during ground transport as compared to paramedics without critical care designation. Additional training and experience may allow for early intervention before clinical

Fig. 4. Risks associated with interhospital transfer process. Transfer itself confers significant patient-oriented risks, including (1) risks of clinical deterioration, (2) lack of adequate resources available during the transfer period, (3) delays in time-sensitive care, (4) inherent risks of transportation, (5) inadequate handoff communication, and (6) neglected patient preferences.

deterioration, thereby preventing intratransfer complications [84]. The importance of this factor may be magnified during intensive care unit (ICU) transfers [80,88,89].

Poor handoff communication

Inadequate provider communication during the transfer process is common. Both referring and receiving providers cite poor communica- tion as a problem [90-92], and handoff communication is a significant issue addressed in transfer guidelines [93]. Few data exist to associate clinical outcomes with provider communication, but poor communication has been cited as a cause of repeated tests and diagnostic uncertainty at the time of transfer [53,94-96].

Neglected patient preferences

Patients’ perceptions and roles in transfer are a topic that has re- ceived minimal research interest. Data from a series of qualitative inter- views suggest that rural patients transferred to tertiary care centers are concerned about the transfer process and the impersonality of metro- politan hospitals and are confused about the acuity of their illness [97]. Furthermore, anxiety, stress, and fear are common among those being transferred, especially during transitions of care [98]. In a survey, patients and physicians disagreed on the factors most influential in pa- tient transfer decisions. Although referring and receiving physicians had a high level of concordance in reasons for transfer, patient-physician agreement was poor, especially between patients and referring physi- cians [11]. Proximity to home remains an important factor that patients consider in their decision whether they would prefer transfer to a tertiary care center [14]. More research regarding communication and patient preferences during interhospital transfer is necessary.

The future of regionalization in rural areas

Health care delivery in rural hospitals is changing. Rural hospitals struggle to provide medical services, as medicine becomes more technology intensive and specialist directed. Furthermore, the rate of

Table 3

Research agenda for regionalization

Areas of change Research questions

innovation and expansion of medical knowledge challenges rural providers to stay up to date on a vast array of current therapies, especially in centers where specialist-directed innovations are not routinely employed.

In the coming years, a more evidence-based, efficient, and cost- effective system for interhospital transfer should be developed. Chang- ing economic incentives, realignment of hospitals into Hospital systems, increasing standardization of medical records, and a renewed focus on patient-oriented medical care highlight weaknesses in the current sys- tem of regionalization. Table 3 illustrates our proposed research agenda for actively developing rural regionalization. The transfer networks of the future will need to better match patients to appropriate care, use defined criteria for transfer, and share patient information both before and after the transfer so that resources are used effectively.

Telemedicine and telepharmacy

One rapidly advancing strategy for regionalized care in rural areas is telemedicine. Telemedicine first penetrated EDs through the remote in- terpretation of radiographs, especially after-hours and on weekends [99]. Only more recently did telemedicine enter the clinical patient- physician interaction [100]. Telemedicine-enabled critical care deve- loped rapidly in the early 2000s to bring specialty intensivists into the care of ICU patients. Several large networks now provide either primary critical care services or after-hours coverage for ICUs around the nation [101]. More recently, a growing contingent of EDs use telemedicine links to tertiary centers that can provide help with arranging transfers, provide consultation with EPs or other specialists, and assist in the man- agement of complicated or unstable patients [100]. Utilization of these services varies in different regions of the country, but this trend is clear- ly moving the expertise of specialists into rural areas where specialty coverage is scarce.

Telemedicine has also affected pharmacy services. Remote clini- cal pharmacists can support medication review in rural hospitals and can work in collaboration with providers to improve medica- tion prescribing and modify treatment [102]. These relationships are often supported by formulary autosubstitutions to allow a small hospital to expand its formulary to meet its own needs with- out having to completely adapt to the formulary of a larger institu- tion [103]. Creative system-based pharmacy strategies can provide optimal tertiary-level care to rural hospitals as part of a larger initiative to expand availability of clinical pharmacist expertise.

Clinical decision rules

Telemedicine and telepharmacy

How can telehealth impact Patient evaluation, pretransfer triage, and treatment?
Will increased use of telemedicine change the number of patients who are transferred?

How can telemedicine be made cost-effective in regionalized systems of care (eg, cost sharing)?

A possible mechanism for overcoming heterogeneity in transfer practice and creating evidence-based transfer criteria would be the de- velopment of clinical decision rules to guide transfer. Decision rules have been successful in other areas of emergency medicine [104-107]

Back transfer 1. What conditions/interventions are safest for back transfer?

How do clinical outcomes compare for back-transferred patients?
How do patients perceive back transfer?

How will finances be shared between hospitals?

HIEs 1. How do HIE impact repeat testing, admission rates, and clinical outcomes?

2. What is the magnitude of potential cost-savings with integrated HIEs?

3. What are the risks of using HIE for transferred patients?

Clinical decision rules 1. Can decision rules create a more consistent quality of care or decrease in cost for specific conditions?

2. Which emergency conditions will benefit most from creation

of decision rules?

Participatory care 1. How can patients and families be more involved in the transfer decision?

2. What is the impact of participatory decision making on transfer rates and clinical outcomes?

3. Can discussion of end-of-life issues and “heroic” care with patients or families alter the number of transfers?

and have been used to recommend specialty referral [108]. However,

decision models to guide Transfer decision making for seriously ill pa- tients are in their infancy. The best example of a clinical decision rule for transfer is the field triage decision scheme developed by the American College of Surgeons Committee on Trauma in 1976 [109]. This rule outlines a 4-step process for evaluation of a trauma patient by EMS and provides guidance on which patients need transport to a trauma center. In a recent multisite study of more than 122000 injured patients, the FTDS has been shown to have an overall sensitivity of 85.5% and specificity of 63.9% in detecting major trauma but results in an undertriage of more than 10% [110]. The decision of which trauma pa- tients presenting to level III or IV centers need immediate transfer was studied by Newgard et al [111]. This group identified 5 factors (emer- gent airway intervention, initial Glasgow Coma Scale score b 11, need for blood transfusion, initial systolic blood pressure b 100 or N 220 mm Hg, and initial respiratory rate b 10 or N 32 breaths per minute), that cor- related with poorer outcomes and suggested that these could be used as

criteria for immediate transfer [111]. Newgard et al [112] also proposed a decision rule for transfer of blunt force traumatic brain injury patients. To date, neither of these rules has been validated in clinical practice.

One area in which decision rules have been tested is for selection of appropriate transfer personnel. In one example, Australian investigators showed that a 15-minute Education session and reminder posters could train physicians, nurses, and paramedics to categorize patients by illness severity and select an appropriate level of escort during transfer, even 3 months after the training session [113]. Another example is provided in Hong Kong, where an interfacility transport triage guideline and a 20- minute training session improved appropriateness of pretransfer triage from 35% to 55% (P = .11) and transport personnel selection from 73% to 93% (P = .03) [114]. Decision rules have also been developed to iden- tify critically ill patients who are not suitable for transfer, such as those who are likely to clinically deteriorate en route [115].

Research that develops and assesses clinical decision rules for inter- hospital transfer could provide a useful adjunct for providers evaluating patients in rural hospitals.

Health information exchanges

Health information exchanges (HIEs) and shared electronic medical record systems have been heralded as a way to improve continuity of care between providers and to reduce unnecessary repeat testing [116]. Recent studies have shown that accessing an HIE results in less neuroimaging for headache patients [117] and fewer radiographs with- in 30 days of a prior ED visit or admission [118]. In addition, data suggest that use of an HIE results in fewer patients being admitted, especially for short admissions (b 24 hours) and readmissions within 7 days of dis- charge [119]. When ED physicians accessed HIE records, the odds of ad- mission were reduced by up to 30% (odds ratio, 0.70; 95% CI, 0.52-0.95) [120]. Health information exchange access was also associated with an estimated net savings of $1.1 million annually for a 12-hospital network, with reduced admissions accounting for 97% of these savings [121]. More studies should investigate the impact of HIE and shared electronic medical record on repeat testing, admissions, cost-savings, and clinical outcomes for transferred patients.

Back transfer

Patients transferred for specialty services (eg, cardiac catheterization or trauma) often remain at a tertiary center for the remainder of their care. From a health utilization perspective, this can be problematic be- cause centralization is concentrating patients for both specialty care and lower acuity care in beds equipped to care for people with very spe- cialized problems. Some centers have started transferring these patients back to the initial hospital after the specialty care has been provided to open beds to care for more specialized patients [122]. In Europe, back transfer has been studied for hemodynamically stable unstable angina, non-STEMI, and STEMI patients after uncomplicated angiography and PCI at specialty centers [123-126]. Matteau et al suggest that immediate back transfer would be safe for up to two-thirds of STEMI patients after PCI [123]. Back transfer could prove an efficient strategy to lessen the burden of overcrowding in tertiary care centers and provide a steady revenue stream for smaller hospitals. Revenue-sharing strategies to share a single “admission” over several institutions will pose a difficult but surmountable challenge. Back transfer will also be an area in which HIEs and telemedicine could help provide consistent care through sharing of patient information between hospitals.

Conclusions

Regionalization through interhospital transfer has revolutionized medical care, especially in rural areas. Formal networks of care have im- proved outcomes in high-risk diseases, but transfer variability remains significant. Selection of patients who will benefit from transfer,

identification of receiving hospitals, and arranging timely transport all re- main significant challenges. In addition, the rural health care environment is rapidly changing with concurrent processes of regionalization and cen- tralization. Further work is needed to understand the complicated inter- play of resource availability, patient preferences, institutional culture, financial factors, and political realities of interhospital transfer. The future of regionalized emergency care will require new paradigms for evidence- guided transfer decisions, with increased use of telemedicine, shared medical records, and back transfer after medical stabilization.

References

AJEM

Achieving regionalization through rural interhospital transfer

a b s t r a c t

Introduction