Advertisement

Interhospital vascular surgery transfers to a tertiary care hospital

Open ArchivePublished:December 28, 2017DOI:https://doi.org/10.1016/j.jvs.2017.09.044

      Abstract

      Objective

      Interhospital transfers (IHTs) to tertiary care centers are linked to lower operative mortality in vascular surgery patients. However, IHT incurs great health care costs, and some transfers may be unnecessary or futile. In this study, we characterize the patterns of IHT at a tertiary care center to examine appropriateness of transfer for vascular surgery care.

      Methods

      A retrospective review was performed of all IHT requests made to our institution from July 2014 to October 2015. Interhospital physician communication and reasons for not accepting transfers were reviewed. Diagnosis, intervention, referring hospital size, and mortality were examined. Follow-up for all patients was reviewed.

      Results

      We reviewed 235 IHT requests for vascular surgical care involving 210 patients during 15 months; 33% of requested transfers did not occur, most commonly after communication with the physician resulting in reassurance (35%), clinic referral (30%), or further local workup obviating need for transfer (11%); 67% of requests were accepted. Accepted transfers generally carried life- or limb-threatening diagnoses (70%). Next most common transfer reasons were infection or nonhealing wounds (7%) and nonurgent postoperative complications (7%). Of accepted transfers, 72% resulted in operative or endovascular intervention; 20% were performed <8 hours of arrival, 12% <24 hours of arrival, and 68% during hospital admission (average of 3 days); 28% of accepted patients received no intervention. Small hospitals (<100 beds) were more likely than large hospitals (>300 beds) to transfer patients not requiring intervention (47% vs 18%; P = .005) and for infection or nonhealing wounds (30% vs 10%; P = .013). Based on referring hospital size, there was no difference in IHTs requiring emergent, urgent, or nonurgent operations. There was also no difference in transport time, time from consultation to arrival, or death of patients according to hospital size. Overall patient mortality was 12%.

      Conclusions

      Expectedly, most vascular surgery IHTs are for life- or limb-threatening diagnoses, and most of these patients receive an operation. Transfer efficiency and surgical case urgency are similar across hospital sizes. Nonoperative IHTs are sent more often by small hospitals and may represent a resource disparity that would benefit from regionalizing nonurgent vascular care.
      Article Highlights
      • Type of Research: Retrospective single-institution cohort study
      • Take Home Message: Of 235 interhospital transfer (IHT) requests for vascular surgical care of 210 patients during 15 months, 70% of the accepted IHTs were for life- or limb-threatening diagnoses that required surgery, and 28% of transfers were nonoperative and were more often sent by small hospitals.
      • Recommendation: This study suggests that admitting nonoperative IHTs, sent more often by small hospitals, may represent a resource disparity that would benefit from regionalizing nonurgent vascular care.
      Regionalization of specialized vascular care has increased the volume of patients requiring interhospital transfer (IHT) for evaluation and treatment. Patients undergoing IHT in the United States for intact abdominal aortic aneurysm repair have more than doubled.
      • Ilonzo N.
      • Egorova N.N.
      • Nowygrod R.
      Interhospital transfer for intact abdominal aortic aneurysm repair.
      In Canada, IHTs for emergent vascular surgery overall have increased by >500% in the last 10 years.
      • Roche-Nagle G.
      • Bachynski K.
      • Nathens A.B.
      • Angoulvant D.
      • Rubin B.B.
      Regionalization of services improves access to emergency vascular surgical care.
      Although IHT is an important tool for connecting patients with appropriate care, it is a costly resource. Patients transferred to tertiary hospitals have been shown to consume more hospital resources compared with those directly admitted regardless of age, sex, and diagnosis.
      • Bernard A.M.
      • Hayward R.A.
      • Rosevear J.
      • Chun H.
      • McMahon L.F.
      Comparing the hospitalizations of transfer and non-transfer patients in an academic medical center.
      To relieve our increasingly overburdened health care system, hospitals, and medical providers, it is imperative that IHT be used with discretion.
      The first step in effective use of resources is understanding the body of vascular patients undergoing IHT. In this study, our objective was to investigate the pattern of IHT for vascular surgery patients to our institution and to examine the appropriateness of transfer within our system.

      Methods

      The Institutional Review Board at our institution approved this study. As this was a retrospective review that posed minimal risk and would not adversely affect the welfare of the participants, a waiver of consent was granted.

       Recruitment and population

      IHT requests at our institution have been recorded through an emergency department services transfer center since January 2012. The information technology manager of emergency medicine queried the database for IHTs in which the vascular surgery division was the intended recipient service and provided the authors with a list of patient names, medical record numbers, transfer hospital, accepting surgeon, and patient disposition. All patients who were involved in a request for transfer were included in this study. This includes transfer attempts that did not result in transfer.

       Acceptance

      The protocol for discussing referrals requires physician-to-physician communication, and the decision to accept the transfer is performed on a case-by-case basis. There currently is no algorithm for accepting or rejecting referrals.

       Imaging

      Our transfer center facilitates communication between physicians and arranges Digital Imaging and Communications in Medicine imaging transfer when it is available. Our center uses IMPAX (Agfa HealthCare, Carlstadt, NJ). Image availability for transfers was reviewed.

       Chart review

      Chart review using Epic medical records was conducted to obtain information about patients involved in transfer requests. The result of the transfer request was reviewed. Demographic information including age, sex, and insurance type was collected. Diagnosis of completed transfer patients was evaluated, as was type of intervention performed once the patient was transferred. We also collected information on the transfer hospital including the number of beds and the time it took for the transfer to occur. Transfer time is defined as the time elapsed between the patient's leaving the referring hospital and arriving to the destination hospital. This is relevant because delays in transfer time may serve as a proxy for the efficiency of the transfer process. Finally, we examined mortality rates of transferred patients.

      Results

      From July 2014 to October 2015, there were 235 requests for transfer; 78 (33%) requests did not result in transfer. Of the requests that did not result in transfer, 35% stayed at the referring facility after reassurance from the consulted physician, 30% were offered clinic referral, and 11% had need to transfer obviated by further local workup. Of those patients who were offered clinic referral, 60% ultimately followed up in our clinic. Of the remaining patients not transferred, 7% of patients underwent further nonsurgical treatment at the local institution, 6% of patients or patients' families refused to transfer, and 6% were referred to another service. Information was not available for the remaining 5%.
      Of the 233 requested transfers, 102 images were shared digitally with our system from the outside hospital. Of 154 completed transfers, images were shared in 75. Of the remaining transfers, 48 were accepted on the basis of clinical description and no images were specifically requested. In only five instances was there documentation definitively stating that images were requested but not able to be shared. Two of these instances were from the same small regional hospital, and the remaining three were of duplex ultrasound images that were available only on a proprietary imaging system. Of the 25 requested transfers that involved the thoracic or abdominal aorta, 20 transfer requests shared images with our institution. We did not encounter a case in which inability to share imaging resulted in an aborted transfer.
      The mean age of patients transferred was 61 years (standard deviation, 18.02 years), and 63% were men. The average distance of hospitals was 86 miles (standard deviation, 72.73 miles). There was no significant difference in age, gender, referring hospital size, or hospital distance between requests that were accepted and requests that were not accepted (Table).
      TableCharacteristics of patients and hospitals involved in transfer request
      CharacteristicTransferred (n = 154)Not transferred (n = 78)P
      Age, years, mean (SD)61 (18)60 (19).642
      Continuous variables compared with Student t-test.
      Male97 (63)43 (55).248
      Dichotomous variables compared with χ2 test.
      Referring hospital size.531
      Dichotomous variables compared with χ2 test.
       Small (n = 85)56 (65)29 (35)
       Medium (n = 66)49 (74)17 (26)
       Large (n = 65)46 (70)19 (30)
      Hospital distance, miles, mean (SD)85 (73)89 (69).660
      Continuous variables compared with Student t-test.
      Weekend, Saturday-Sunday16 (10)15 (24).062
      Dichotomous variables compared with χ2 test.
      SD, Standard deviation.
      Values are reported as number (%) unless otherwise indicated.
      a Continuous variables compared with Student t-test.
      b Dichotomous variables compared with χ2 test.
      Most requests, 154 (67%), resulted in completed transfer. Diagnoses ranged from ruptured abdominal aortic aneurysm to infected chronic ulcers (Fig). Diagnoses considered life or limb threatening composed the largest share of transfer requests, 70%. The next most common diagnoses were infection or nonhealing wounds and nonurgent postoperative complications, both at 7%.
      Figure thumbnail gr1
      FigDiagnoses of transferred patients. AAA, Abdominal aortic aneurysm; ALI, acute limb ischemia; AMI, acute mesenteric ischemia; CLI, chronic limb ischemia; DVT, deep venous thrombosis; TAAA, thoracoabdominal aortic aneurysm.
      Of the transfer requests completed, 72% resulted in a procedure or surgical intervention. Of these, 20% of interventions occurred within 8 hours, 12% occurred between 8 and 24 hours, and the majority at 68% occurred at some point during admission.
      There was information about specialty of the referring provider for 117 cases. Most referrals came from emergency department physicians at 76%, and 70% of these referrals were accepted. The rest of the providers included 6% primary care, 5% vascular surgery, 4% general surgery, and 3% midlevel providers. Primary care providers had the lowest acceptance rate at 43%. Vascular surgery, general surgery, and midlevel providers had an acceptance rate of 83%, 80%, and 75%, respectively. The number of cases in each category is too low for meaningful statistical comparison.
      Large hospitals (>300 beds) were more likely to transfer patients requiring a procedure or surgical intervention (47% of patients at large hospitals vs 18% of patients at small and medium hospitals; P = .005). Small hospitals (<100 beds) were more likely to transfer patients requiring nonsurgical intervention (32% at small hospitals vs 18% at medium and large hospitals; P = .041) and were more likely to transfer patients for infected or nonhealing wounds (30% of patients sent from small hospitals vs 10% of those sent from medium and large hospitals; P = .013). Types of nonsurgical or nonprocedural intervention included observation and anticoagulation (36%), wound care or intravenous antibiotics (21%), ward-level medical care (18%), specialist consultation other than vascular (9%), further imaging only (9%), and intensive care unit-level medical care (6%).
      At 6-month follow-up, overall mortality rate of transfer patients was 12%. The 30-day mortality rate for transferred patients was 5%. No patients died during transfer or as a direct result of transfer. When patients who died were compared with those who did not, there was no difference in transfer time (58 vs 79 minutes; P = .280) or from first call to arrival (231 vs 303 minutes). Of the 5% with ruptured abdominal aortic aneurysms, all patients (eight in total) were accepted by our institution. Two patients did not complete transfer; for one patient, a vascular surgeon became available at the referring facility, and the patient stayed. Another patient refused transfer; this patient was 95 years old and had orders for do not resuscitate or intubate.
      Although mortality rates trended lower for larger hospitals, there was no significant difference in mortality rate between small, medium, and large hospitals (12%, 25%, and 9%; P = .094).

      Discussion

      We investigated the transfer patterns to our facility both to better understand our practices and to understand the needs in our referral region. We found that 33% of transfer requests did not result in transfer. Patients stayed at referring hospitals usually after reassurance from the consulted physician or after recommendations for further local workup. Most transfers involved life- or limb-threatening diagnoses, and 72% required an operation. The 30-day mortality was 5%, and no patients died during transfer or as a documented direct result of transfer; 28% of patients transferred to our academic medical center from another facility did not require a procedure or surgical intervention. Compared with medium and large hospitals, small hospitals were found to be more likely to send patients who require nonsurgical intervention and were more likely to transfer patients for infected or nonhealing wounds. The most common nonsurgical interventions performed included observation and anticoagulation, wound care or intravenous antibiotics, and ward-level medical care.
      Vascular patients with wounds appear to make up a significant subset of patients transferred from small hospitals to our tertiary facility. The point of access for many patients across Oregon is a small hospital serving a rural community. Oregon is a largely rural state in terms of land; 23.5% of the population of Oregon is considered rural, and this population occupies 80% of the state. There are >60 hospitals in Oregon, and as in most states, the hospitals are clustered in metropolitan areas. In Portland, which is nestled in the northwest corner of the state, there are several major vascular referral centers. Outside of Portland, vascular referral centers exist in other relatively population-dense areas, such as Salem, Eugene, Bend, and Medford. Large swaths of rural areas exist between vascular centers, and small hospitals are instrumental in being the primary access point for rural patients. With this context, two conclusions can be drawn: there is a subset of nonsurgical vascular patients, mostly patients with wounds, who require resource-rich care, and our small hospitals in Oregon are not equipped to take them on. This may speak to an absence of medical infrastructure in place at small hospitals to take care of nonsurgical vascular patients, particularly those with wounds. This is understandable—vascular wound care requires a multidisciplinary team and time-intensive nursing attention to wounds. In addition, the multipronged workup of patients with vascular wounds may be beyond the scope of physicians practicing at small hospitals.
      However, it is possible that some patients may have been transferred without a true need for advanced or specialized care. In a review of acute care surgery transfers to a large academic institution, 20% were deemed not necessary by a consensus of surgeons.
      • Broman K.K.
      • Poulose B.K.
      • Phillips S.E.
      • Ehrenfield J.M.
      • Sharp K.W.
      • Pierce R.A.
      • et al.
      Unnecessary transfers for acute surgical care: who and why?.
      In a review of pediatric emergency IHTs at a single institution, 28.7% were discharged home in 12 hours or less without any intervention,
      • Gattu R.K.
      • De Fee A.S.
      • Lichenstein R.
      • Teshome G.
      Consideration of cost of care in pediatric emergency transfer—an opportunity for improvement.
      as were 22% of pediatric neurosurgery IHTs. An examination of trauma patients transferred to a level 1 trauma center found that 24% were discharged within 24 hours, most requiring observation or minor procedures that could have been performed at the referring institution.
      • Tang A.
      • Hashmi A.
      • Pandit V.
      • Joseph B.
      • Kulvatunyu N.
      • Vercruysse
      • et al.
      A critical analysis of secondary overtriage to a Level I trauma center.
      These findings highlight a problem with correctly identifying patients with complex diagnoses who ultimately require specialist care.
      Our study had several limitations. As a retrospective chart review, it is subject to bias. In addition, this study is a review of a single-center experience in a unique location that may not be generalizable to other regions. Oregon is a geographically diverse state, stretching from rocky coast, over the Cascade Mountains, and through the high desert. Population ranges widely, from dense (1517 persons per square mile in Multnomah County) to sparse (0.8 person per square mile in Harney County).

      U.S. Census Bureau. 2010 Census of Population and Housing. Population and housing unit counts. Washington, D.C.: U.S. Government Printing Office; 2012.

      The austere environment may uniquely affect our transfer process; in a review of emergency department deaths from abdominal aortic aneurysms, the western United States was found to have twice the in-emergency department mortality rate compared with any other region in the United States.
      • Mell M.W.
      • Callcut R.A.
      • Bech F.
      • Delgado M.K.
      • Staudenmayer K.
      • Spain D.A.
      • et al.
      Predictors of emergency department death for patients presenting with ruptured abdominal aortic aneurysms.
      Another variable that detracts from the generalizability of these data is the Oregon health care system. The Oregon Health Plan is the state Medicaid Program, which has been in place since the 1990s. Between the Oregon Health Plan and Medicaid expansion, 95% of Oregonians, as of publication of this manuscript, have insurance coverage. In rural counties, up to 40% of patients are covered under the state health plan. Since the health plan expansion in 2012, hospital readmissions have been reduced by one-third, and emergency department visits have decreased by almost half.
      • Bernard A.M.
      • Hayward R.A.
      • Rosevear J.
      • Chun H.
      • McMahon L.F.
      Comparing the hospitalizations of transfer and non-transfer patients in an academic medical center.
      Nationally, the nonelderly uninsured rate has fallen with the Affordable Care Act from 18% in 2010 to 10.5% in 2015. The uninsured rates have fallen across all states, although they have fallen more rapidly in states like Oregon that have chosen to expand Medicaid.
      • Broman K.K.
      • Poulose B.K.
      • Phillips S.E.
      • Ehrenfield J.M.
      • Sharp K.W.
      • Pierce R.A.
      • et al.
      Unnecessary transfers for acute surgical care: who and why?.
      Health care coverage undeniably plays a role in the actions of providers and of patients and their families, and the nuances of this type of decision-making were not explored in this paper.
      Because of the particularities of Oregon's population, terrain, and health care system, the conclusions drawn about the vascular patients in these data may not be generalizable across the country. That was not our aim; rather, we sought to uncover systemic issues in our IHTs that may aid other institutions in examining their own practices.
      Another significant limitation to our study was the inability to track the patients who were not transferred to our institution. When transfers are not completed, there is no further record of the patient unless the patient is followed up at our institution in another encounter. It would be useful to analyze the outcomes of patients who were not transferred, particularly in the context of why they were not transferred. Our study raises the question of necessity of some transfers to a tertiary care center; the ability to track outcomes of patients who were managed locally, particularly those who may have had worse outcomes, would add an important dimension to this conversation.
      Future studies should include examination of how to improve physician-to-physician communication with referral centers. Currently, communication between our institution and outlying referral centers is performed only over the telephone. This can be improved on; for acute stroke consultations, for example, it has been shown that audiovisual telemedicine is more sensitive and specific for determination of thrombolysis eligibility compared with telephone communication.
      • Capampangan D.J.
      • Wellik K.E.
      • Bobrow B.J.
      • Aguilar M.I.
      • Ingall T.J.
      • Kiernan T.E.
      • et al.
      Telemedicine versus telephone for remote emergency stroke consultations: a critically appraised topic.
      Oregon is taking steps to integrate telemedicine into the state health care system. As of 2015, Oregon law requires coverage of telemedicine by all private insurers. To be covered by Medicaid or Medicare, providers must be enrolled in a program and patients must live in jurisdictions that justify telemedicine use.
      The Medicare Health Parity Act of 2014, currently under committee review, seeks to remove these restrictions.
      • Lamar C.
      The state of telemedicine/telehealth in Oregon.
      Increasing use of telemedicine could potentially help elevate wound care in rural Oregon and lessen the strain on both patients and hospitals by reducing the need for transfer. This may aid in bridging the knowledge gap for nonurgent vascular diagnoses in referral centers, which would take a step toward regionalizing some aspects of vascular care.
      Other areas of future study should include standardization of the transfer process. Despite the growing number of national IHTs, it remains a largely unstandardized process.
      • Mueller S.K.
      • Schnipper J.L.
      Physician perspectives on interhospital transfers.
      For example, ruptured abdominal aortic aneurysms constitute a vascular diagnosis widely recognized as benefiting from specialty vascular care, and yet there are currently no guidelines for transfer criteria.
      • Mell M.W.
      • Schneider P.A.
      • Starnes B.W.
      Variability in transfer criteria for patients with ruptured abdominal aortic aneurysm in the western United States.
      Information handoff for transfers is frequently lacking, with only about 60% of critically ill patients transferring with complete documentation.
      • Usher M.G.
      • Fanning C.
      • Wu D.
      • Muglia C.
      • Balonze K.
      • Kim D.
      • et al.
      Information handoff and outcomes of critically ill patients transferred between hospitals.
      The lack of standardization of transfer criteria may be a contributing factor to patients' being transferred without need for intervention.

      Conclusions

      Our study sought to characterize IHT of vascular surgery patients at our institution. We found that as expected, most transfers were for life- and limb-threatening diagnoses. Despite the geographic challenges presented by the Pacific Northwest, we did not see any transfer-related deaths in this series. An area we found that may merit further investigation involves transfers from small hospitals of patients who do not ultimately require a procedure. These transfers represent patients with nonurgent vascular diagnoses. It would serve our health care system well to understand how to help small hospitals manage these patients without the need for transfer.

      Author contributions

      Conception and design: SH, DW, GL, TL, GM, EM
      Analysis and interpretation: SH, DW, EJ, AA, GM
      Data collection: SH
      Writing the article: SH, EM
      Critical revision of the article: DW, EJ, AA, GL, TL, GM, EM
      Final approval of the article: SH, DW, EJ, AA, GL, TL, GM, EM
      Statistical analysis: Not applicable
      Obtained funding: Not applicable
      Overall responsibility: SH

      References

        • Ilonzo N.
        • Egorova N.N.
        • Nowygrod R.
        Interhospital transfer for intact abdominal aortic aneurysm repair.
        J Vasc Surg. 2016; 63: 859-865.e2
        • Roche-Nagle G.
        • Bachynski K.
        • Nathens A.B.
        • Angoulvant D.
        • Rubin B.B.
        Regionalization of services improves access to emergency vascular surgical care.
        Vascular. 2013; 21: 69-74
        • Bernard A.M.
        • Hayward R.A.
        • Rosevear J.
        • Chun H.
        • McMahon L.F.
        Comparing the hospitalizations of transfer and non-transfer patients in an academic medical center.
        Acad Med. 1996; 71: 262-266
        • Broman K.K.
        • Poulose B.K.
        • Phillips S.E.
        • Ehrenfield J.M.
        • Sharp K.W.
        • Pierce R.A.
        • et al.
        Unnecessary transfers for acute surgical care: who and why?.
        Am Surg. 2016; 82: 672-678
        • Gattu R.K.
        • De Fee A.S.
        • Lichenstein R.
        • Teshome G.
        Consideration of cost of care in pediatric emergency transfer—an opportunity for improvement.
        Pediatr Emerg Care. 2017; 33: 334-338
        • Tang A.
        • Hashmi A.
        • Pandit V.
        • Joseph B.
        • Kulvatunyu N.
        • Vercruysse
        • et al.
        A critical analysis of secondary overtriage to a Level I trauma center.
        J Trauma Acute Care Surg. 2014; 77: 969-973
      1. U.S. Census Bureau. 2010 Census of Population and Housing. Population and housing unit counts. Washington, D.C.: U.S. Government Printing Office; 2012.

        • Mell M.W.
        • Callcut R.A.
        • Bech F.
        • Delgado M.K.
        • Staudenmayer K.
        • Spain D.A.
        • et al.
        Predictors of emergency department death for patients presenting with ruptured abdominal aortic aneurysms.
        J Vasc Surg. 2012; 56: 651-655
        • Capampangan D.J.
        • Wellik K.E.
        • Bobrow B.J.
        • Aguilar M.I.
        • Ingall T.J.
        • Kiernan T.E.
        • et al.
        Telemedicine versus telephone for remote emergency stroke consultations: a critically appraised topic.
        Neurologist. 2009; 15: 163-166
      2. Eligible Providers and Services (relating to reimbursement). Telehealth Alliance of Oregon, 2016
        • Lamar C.
        The state of telemedicine/telehealth in Oregon.
        Oregon Healthcare News, December 7, 2015
        • Mueller S.K.
        • Schnipper J.L.
        Physician perspectives on interhospital transfers.
        J Patient Saf. 2016 Nov 2; ([Epub ahead of print])
        • Mell M.W.
        • Schneider P.A.
        • Starnes B.W.
        Variability in transfer criteria for patients with ruptured abdominal aortic aneurysm in the western United States.
        J Vasc Surg. 2015; 62: 326-330
        • Usher M.G.
        • Fanning C.
        • Wu D.
        • Muglia C.
        • Balonze K.
        • Kim D.
        • et al.
        Information handoff and outcomes of critically ill patients transferred between hospitals.
        J Crit Care. 2016; 36: 240-245