Abstract
Background
Emergency general surgery (EGS) conditions are common, require urgent surgical evaluation, and are associated with high mortality and costs. Although appropriate interhospital transfers are critical to successful EGS care, the performance of EGS transfer systems remains unclear. We aimed to describe EGS transfer patterns and identify factors associated with potentially avoidable transfers (PAT).
Methods
We performed a retrospective cohort study of EGS episodes in eight U.S. states using the 2016 Healthcare Cost and Utilization Project State Inpatient and Emergency Department (ED) Databases and the American Hospital Association Annual Surveys. We identified ED-to- Inpatient and Inpatient-to-Inpatient interhospital EGS transfers. PAT was defined as discharge within 72 hours after transfer without undergoing any procedure or operation at the destination hospital. We examined transfer incidence and characteristics. We performed multilevel regression examining patient-level and hospital-level factors associated with PAT.
Results
Of 514,410 adult EGS episodes, 26,281 (5.1%) involved interhospital transfers (ED-to- Inpatient: 65.0%, Inpatient-to-Inpatient: 35.1%). Over 1 in 4 transfers were potentially avoidable (7,188, 27.4%), with the majority occurring from the ED. Factors associated with increased odds of PAT included self-pay (vs. government insurance, OR: 1.26, 95% CI: 1.09–1.45, p=0.002), level 1 trauma centers (vs. non-trauma centers, OR: 1.24, 95% CI: 1.05–1.47, p=0.01), and critical access hospitals (vs. non-critical access, OR: 1.30, 95% CI: 1.15–1.47, p<0.001).Hospital-level factors (size, trauma center, ownership, critical access, location) accounted for 36.1% of PAT variability.
Conclusions
Over 1 in 4 EGS transfers are potentially avoidable. Understanding factors associated with PAT can guide research, quality improvement, and infrastructure development to optimize EGS care.
Article Summary
Over 1 in 4 emergency general surgery (EGS) transfers are potentially avoidable, with hospital-level factors accounting for a substantial portion of its variability. The importance of these findings is identification of potential inefficiencies in EGS care delivery and targets to guide necessary future research, quality improvement, policy, and infrastructure development to improve EGS transfer systems and patient outcomes.
Introduction
Over 3 million hospital encounters for emergency general surgery (EGS) conditions occur annually in the United States.(1–7) These conditions require urgent surgical evaluation and are associated with high mortality, costs, and resource utilization.(1–7) With gaps in access to EGS care, interhospital transfers for these patients frequently occur, comprising up to 13% of EGS encounters.(2, 5) However, there are risks associated with transfers, including increased mortality and costs, in addition to burden on patients, families, providers, and healthcare systems.(2, 4, 6–10) Debate on regionalization of EGS care is ongoing, balancing the benefits of providing EGS care to communities that lack surgical resources with the risks of placing undue burden on tertiary centers and patients who have to travel far away from home with unclear clinical advantage.(2, 11) Yet standardized guidelines for the EGS transfer system are lacking.(2, 8, 11) Although appropriate and efficient interhospital transfers are critical to successful EGS care, the epidemiology and performance of EGS transfer systems remain unclear. In order to design effective regionalized systems for EGS, we need to first comprehensively understand the existing interhospital transfer network, and then identify inefficiencies and potential causes to target improvement efforts.
There is increasing effort to examine the EGS transfer system, but our current understanding is fragmented and often does not include the full scope of EGS as defined by the American Association for the Surgery of Trauma (AAST).(1–4, 6, 7, 9, 12, 13) EGS encompasses diverse diseases involving multiple organ systems, with variable severity and management, from nonoperative to operative interventions.(1–7) Several prior studies examined subgroups of EGS, such as only operatively-managed cases,(6, 7, 9) even though only 29–34% of EGS encounters are managed operatively,(9, 14, 15) or included only select EGS conditions or procedures that represented the majority, but not the full burden, of EGS on our healthcare system.(3, 7) Some studies were limited in geographic scope, with cohorts limited to a single state(13) or receiving institution.(3, 8) Additionally, referring and destination hospitals, pre- and post-transfer care, or whether transfers originated from the emergency department (ED) or inpatient settings were not always identified.(4, 6, 7, 9, 12)
Unnecessary transfers, previously defined as discharge within 72 hours after transfer without surgical intervention, have been described in EGS and acute surgical care.(3, 8, 11) This represents inefficient utilization of limited resources and possible inefficiency in our interhospital transfer system. Certain patient and transfer characteristics have been associated with unnecessary transfers.(8) However, there is paucity of data on factors, especially referring hospital characteristics, associated with potentially avoidable transfers (PAT) in EGS, which are critical for targeted quality improvements to optimize resource utilization within EGS transfer systems.(8)
To address these gaps, we sought to (1) comprehensively describe interhospital transfer patterns in EGS by examining a large geographic area, using an inclusive EGS definition involving nonoperative and operative management, and identifying referring and destination hospitals for transfers from both ED and inpatient settings. Secondly, our objective was to (2) identify patient-level and hospital-level factors associated with PAT to identify opportunities to improve EGS care delivery.
Methods
Data Source
We performed a retrospective cohort study using all-payer claims data from the 2016 Agency for Healthcare Research and Quality’s Healthcare Cost and Utilization Project (HCUP) State Inpatient and ED Databases from Arkansas, Florida, Maryland, Massachusetts, Nebraska, New York, Vermont, and Wisconsin. These states were included as they provide unique patient identifiers that allowed us to track patients across hospitals and time. The State Inpatient Databases capture all inpatient hospital stays, including those originating from the ED(16) – we defined these encounters as inpatient encounters. The State ED Databases capture all ED visits that do not result in an admission at the same hospital(16) – we defined these encounters as ED- only encounters. Variables in these databases included patient demographics, admission times, ICD-10-CM diagnosis and procedure codes, in-hospital mortality, and hospital identifiers. Since states in HCUP submit varying numbers of procedure codes, the number of codes was standardized to the minimum provided by any of the included states (14 procedure codes).(16–19)
Identifying Emergency General Surgery Episodes of Care
Episodes of care included at least one hospital encounter. Based on previously published methods using similar claims data, we created episodes by linking patients’ temporally adjacent ED and inpatient encounters that occurred at different hospitals.(17–19) Encounters for each patient were considered temporally adjacent if the admission date to the second hospital encounter was on the same or following day of the discharge date to the first hospital encounter (Table 1).(17–19)
Table 1:
Definition of Emergency General Surgery Episode Types
| Episode Type | First Encounter Referral Hospital (Discharge Day: N) | Second Encounter Destination Hospital (Admission Day: N or N+1) |
|---|---|---|
| No Transfer | Inpatient with EGS diagnosis | - |
| ED-to-Inpatient Transfer | ED | Inpatient with EGS diagnosis |
| Inpatient-to-Inpatient Transfer | Inpatient with EGS diagnosis | Inpatient ± EGS diagnosis |
| Inpatient without EGS diagnosis | Inpatient with EGS diagnosis |
Abbreviations: EGS = emergency general surgery, ED = emergency department
Based on prior studies, we defined an EGS episode as any episode in which the primary diagnosis for the first or second inpatient encounter is an EGS diagnosis, as defined by the AAST.(1, 4, 10, 12, 20) Since 2016 HCUP data included ICD-10-CM codes, we used the Centers for Medicare and Medicaid Services’ General Equivalence Mappings (CMS GEMs) to convert the previously defined EGS diagnosis codes from ICD-9-CM to ICD-10-CM.(21)
Definition of Episode Types
We identified three types of EGS episodes: those that 1) did not involve interhospital transfers, 2) involved ED-to-Inpatient interhospital transfers, and 3) involved Inpatient-to- Inpatient interhospital transfers (Table 1). The ED-to-Inpatient transfer episodes included ED-to-ED transfers that resulted in inpatient admission at the destination hospital. Patients who were discharged from the ED without an inpatient admission were not included. We included EGS episodes in which patients were ≥18 years-old at admission for the first inpatient EGS encounter. Both male and female patients were included.
Based on published definitions of unnecessary transfers in EGS and acute surgical care, PAT was defined as transfers in which the patient was discharged alive within 72 hours after transfer without undergoing any procedure or operation at the destination hospital.(3, 11) This definition was originally developed by Broman and colleagues, through consensus among clinicians who collectively have over 60 years of experience requesting and accepting transfer patients.(8) A 72-hour timeframe allows for comprehensive evaluation and therapeutic interventions, if needed, to be performed by the primary and consulting teams at the destination hospital prior to discharge. Discharges to skilled nursing facilities, long-term care facilities, and hospice were included. Nighttime transfers occurred between 6:00PM and 6:00 AM the following day, based on post-transfer admission hour.(22, 23) Weekend transfers occurred on Saturday or Sunday, based on post-transfer admission date.
Definition of Episode Characteristics
We identified operations and procedures based on ICD-9-CM procedure codes defined by the HCUP Surgery Flags Software.(24) Operations were defined by the “Narrow” surgery flag value and included invasive operations such as laparotomy, laparoscopy, and complex tissue repair. Procedures were defined by the “Broad” surgery flag value and included moderately invasive procedures such as endoscopy and percutaneous drainage.(24) We used the CMS GEMs to convert procedure codes from ICD-9-CM to ICD-10-CM.(21) In cases in which both “Broad” and “Narrow” ICD-9-CM codes mapped onto a single ICD-10-CM code, we classified that ICD- 10-CM code as “Narrow”. We identified complications using ICD-9-CM diagnosis codes previously defined in literature (e.g., myocardial infarction, respiratory failure, pulmonary embolus/deep vein thrombosis, pneumonia, postoperative hemorrhage, surgical site infection, renal failure, shock),(25, 26) again using the CMS GEMs to convert ICD-9-CM to ICD-10- CM.(21) We obtained hospital characteristics and location from the American Hospital Association (AHA) Annual Survey.(27) We obtained trauma center designation from the American College of Surgeons trauma centers listing as of December 20, 2019.(28) We calculated transfer distances using AHA hospital latitude and longitude and grouped hospitals into U.S. geographical regions based on the U.S. Census Bureau.(27, 29)
Statistical Analysis
We reported descriptive statistics of transfer incidence, transfer episode characteristics, and patient outcomes by transfer type. To examine patient-level and referring hospital-level factors associated with PAT, we performed multilevel logistic regression. Patient-level factors included in the analysis were age (adult 18–64 years, elderly ≥65 years), sex, race, Hispanic ethnicity, insurance status (government insurance, private insurance, self-pay, other), Elixhauser comorbidities,(30) and EGS diagnosis categories as defined by the AAST.(1) Hospital-level factors of the referring hospitals included in the analysis were hospital size (total number of hospital beds), number of intensive care unit (ICU) beds, trauma center level, teaching status, ruralness, ownership status (non-profit, for-profit, government), critical access hospital, and location (U.S. census geographic regions). Random effects for hospital were included to account for clustering.(31) To calculate the proportion of variation in PAT due to patient-level and referring hospital-level factors, we used variance from the multilevel models with the above covariates and a null model with no covariates.(31) All hospital-level factors refer to the referring hospital from which the transfer was sent, as these would be more valuable in identifying intervention targets.
Exploratory analyses were performed using multilevel logistic regression models examining the interactions between trauma center level and other factors that were associated with PAT (e.g., hospital ownership, critical access hospital, geographical region, Hispanic ethnicity, patient insurance, psychiatric comorbidities) or off-hours transfers (e.g., weekend, nighttime). Similar exploratory analyses were performed to examine interactions between geographical region and the above factors. Again, all hospital-level characteristics refer to the referring hospital from which the transfer was initiated. Finally, we used Empiric Bayes Kriging to examine the geographic distribution of the proportion of PAT and the mean transfer distance among PAT.(32) This method uses the value of these outcomes at the zip code location for each patient to interpolate a continuous surface map of values across each included state, assuming the distance between known outcome values at sample points reflects spatial correlation related to the variation in outcome values in space.(32)
Continuous and categorical variables were presented as median [interquartile range] and frequency (percentage), respectively. Odds ratio and 95% confidence intervals were reported for the multilevel models. A p-value <0.05 indicated statistical significance. Analyses were performed using STATA 15.1 (copyright 2017 StataCorp LLC, College Station, TX, USA) and geospatial analysis performed using ArcGIS v10.5 (ESRI; Redlands, CA).
Results
Interhospital Transfer Incidence, Characteristics, Outcomes
Of 514,410 adult EGS episodes, 26,281 (5.1%) involved interhospital transfers. Among all transfer episodes, 65.0% were ED-to-Inpatient interhospital transfers, while 35.1% were Inpatient-to-Inpatient interhospital transfers (Table 2). Over 1 in 4 transfers were potentially avoidable (7,188, 27.4%). Figure 1 illustrates the geographic distribution of proportions of PAT, with location based on patient zip code. Median [interquartile range] transfer distance for all transfers was 23.4 [9.4–41.7] miles. 11,947 (45.5%) EGS transfer episodes involved at least one operation or procedure. For inpatient-inpatient transfers, median hospital length of stay at the referring hospital prior to transfer was 3 [1–6] days. Cardiothoracic conditions was the most common EGS diagnosis category(1) for all transfers, transfers from level 1 trauma centers, and PAT (Supplementary Data, eTable 1). Destination hospital characteristics for overall EGS transfers are in Supplementary Data, eTable 2.
Table 2:
Characteristics of Transferred Emergency General Surgery Episodes of Care
| Type of Transfer Episode | |||
|---|---|---|---|
| Characteristic | All Transfers (N=26,281) | ED-Inpatient (N=17,070) | Inpatient-Inpatient (N=9,211) |
| Patient Demographics | |||
| Age (years) | 62 [47–75] | 60 [45–74] | 65 [51–76] |
| Female | 13,520 (51.4) | 8,936 (52.4) | 4,584 (49.8) |
| Race* | |||
| White | 19,986 (80.4) | 13,136 (81.4) | 6,850 (78.6) |
| Black | 3,226 (13.0) | 1,998 (12.4) | 1,228 (14.1) |
| Other | 1,642 (6.6) | 1,005 (6.2) | 637 (7.3) |
| Hispanic † | 1,849 (7.4) | 1,171 (7.2) | 678 (7.8) |
| Insurance † | |||
| Government Insurance | 18,020 (68.6) | 11,437 (67.1) | 6,583 (71.5) |
| Private Insurance | 6,392 (24.3) | 4,305 (25.2) | 2,087 (22.7) |
| Self-Pay | 1,023 (3.9) | 735 (4.3) | 288 (3.1) |
| Other | 831 (3.2) | 581 (3.4) | 250 (2.7) |
| Referral Hospital Characteristics | |||
| Hospital size | |||
| <100 beds | 10,675 (40.6) | 8,163 (47.8) | 2,512 (27.3) |
| 100–250 beds | 8,124 (30.9) | 4,723 (27.7) | 3,401 (36.9) |
| >250 beds | 7,474 (28.5) | 4,176 (24.5) | 3,298 (35.8) |
| ICU beds | |||
| 0 to 10 beds | 11,455 (43.6) | 8,650 (50.7) | 2,805 (30.5) |
| 11 to 25 beds | 6,321 (24.1) | 3,697 (21.7) | 2,624 (28.5) |
| >25 beds | 8,497 (32.3) | 4,715 (27.6) | 3,782 (41.1) |
| Teaching hospital | 8,495 (32.3) | 4,767 (27.9) | 3,728 (40.5) |
| Trauma Center Designation | |||
| Non-trauma center | 23,183 (88.2) | 15,203 (89.1) | 7,980 (86.6) |
| Level 3 | 967 (3.7) | 610 (3.6) | 357 (3.9) |
| Level 2 | 812 (3.1) | 424 (2.5) | 388 (4.2) |
| Level 1 | 1,319 (5.0) | 833 (4.9) | 486 (5.3) |
| Hospital ownership | |||
| Government | 2,308 (8.8) | 1,504 (8.8) | 804 (8.7) |
| Non-profit | 18,947 (72.1) | 12,480 (73.2) | 6,467 (70.2) |
| For-profit | 5,018 (19.1) | 3,078 (18.0) | 1,940 (21.1) |
| Rural hospital | 3.561 (13.6) | 2,816 (16.5) | 745 (8.1) |
| Critical access hospital | 4,739 (18.0) | 3,921 (23.0) | 818 (8.9) |
| U.S. Region § | |||
| Northeast | 9,526 (36.3) | 6,340 (37.1) | 3,186 (34.6) |
| Midwest | 4,670 (17.8) | 3,267 (19.1) | 1,403 (15.2) |
| South | 12,085 (46.0) | 7,463 (43.7) | 4,622 (50.2) |
| Transfer Characteristics | |||
| Potentially avoidable transfer ∥ | 7,188 (27.4) | 5,415 (31.7) | 1,773 (19.3) |
| Weekend transfer ¶ | 7,079 (26.9) | 4,940 (28.9) | 2,139 (23.2) |
| Nighttime transfer # | 9,688 (56.0) | 6,281 (57.2) | 3,407 (54.0) |
| Transfer distance (miles) | 23.4 [9.4–41.7] | 24.1 [10.8, 40.9] | 20.3 [8.0–42.7] |
| Operation or Procedure use | |||
| Any operation or procedure during episode | 11,947 (45.5) | 6,941 (40.7) | 5,006 (54.4) |
| Pre-transfer only | 1,212 (4.6) | - | 1,212 (13.2) |
| Post-transfer only | 9,816 (37.4) | 6,941 (40.7) | 2,875 (31.2) |
| Both pre-transfer and post-transfer | 919 (3.5) | - | 919 (10.0) |
| Neither pre-transfer nor post-transfer | 14,334 (54.5) | 10,129 (59.3) | 4,205 (45.7) |
| Outcomes | |||
| Complication ** | 5,327 (20.3) | 2,318 (13.6) | 3,009 (32.7) |
| In-hospital death | 944 (3.6) | 451 (2.6) | 493 (5.4) |
| With surgery or procedure | 529 (4.4) n=11,947 |
227 (3.3) n=6,941 |
302 (6.0) n=5,006 |
| Without surgery or procedure | 415 (2.9) n=14,334 |
224 (2.2) n=10,129 |
191 (4.5) n=4,205 |
| Failure to rescue †† | 480 (9.0) n=5,327 |
192 (8.3) n=2,318 |
288 (9.6) n=3,009 |
All results are median [interquartile range] or frequency (%).
Abbreviations: ED=emergency department, ICU=intensive care unit, U.S.=United States.
Race was available for 94.6% of all transfer episodes.
Hispanic ethnicity was available for 94.7% of all transfer episodes.
Insurance information was available for 99.9% of all transfer episodes.
U.S. regions were based on geographical regions defined by the U.S. Census Bureau.(29)
Potentially avoidable transfers were defined as post-transfer discharge within 72 hours without undergoing any procedure or operation at the destination hospital.
Weekend transfers were defined as transfers that occurred on Saturday or Sunday.
Nighttime transfers were defined as transfers that occurred between 6:00 PM and 6:00 AM the following day. Data were not available for Massachusetts, Maryland, and Wisconsin. Data were available for 65.8% of all transfer episodes.
Figure 1:
Geographical distribution based on patient zip code of the proportion of potentially avoidable transfers in eight U.S. states.
Among PAT, 75.3% occurred from the ED, 27.6% occurred during the weekend, and 54.6% occurred during nighttime. Characteristics of potentially avoidable transfers that occurred during the weekend and nighttime are provided in Table 3. The majority of patients whose transfers were potentially avoidable had government insurance (4,658, 64.9%). 3.1% of PAT had an operation or procedure performed at the referring hospital prior to transfer, but was discharged within 72 hours after transfer without additional operation or procedure at the destination hospital. Among transfers from level 1 trauma centers, 63.1% were from the ED and 36.9% were inpatient transfers. 33.8% of transfers from level 1 trauma centers were potentially avoidable. Figure 2 illustrates the geographic distribution, based on patient zip code, of transfer distances for PAT.
Table 3:
Characteristics of Types of Potentially Avoidable Interhospital Transfers in Emergency General Surgery
| Type of Potentially Avoidable Transfer | ||||
|---|---|---|---|---|
| Characteristic | Weekday * (N=5,206) | Weekend * (N=1,982) | Daytime † (N=2,172) | Nighttime † (N=2,614) |
| Demographics | ||||
| Age (years) | 57 [42–71] | 57 [42–72] | 56 [40–72] | 55 [40–70] |
| Female | 2,670 (51.3) | 1,021 (51.5) | 1,111 (51.2) | 1,296 (49.6) |
| Race ‡ | ||||
| White | 3,879 (78.8) | 1,474 (78.1) | 1,556 (77.8) | 1,867 (75.6) |
| Black | 687 (14.0) | 284 (15.0) | 304 (15.2) | 384 (15.6) |
| Other | 357 (7.3) | 130 (6.9) | 140 (7.0) | 218 (8.8) |
| Hispanic § | 430 (8.7) | 164 (8.7) | 210 (10.5) | 255 (10.3) |
| Insurance ∥ | ||||
| Government Insurance | 3,384 (65.1) | 1,274 (64.4) | 1,362 (62.9) | 1,685 (64.6) |
| Private Insurance | 1,357 (26.1) | 497 (25.1) | 533 (24.6) | 629 (24.1) |
| Self-Pay | 269 (5.2) | 133 (6.7) | 173 (8.0) | 194 (7.4) |
| Other | 188 (3.6) | 75 (3.8) | 99 (4.6) | 102 (3.9) |
| Transfer Characteristics | ||||
| Transfer distance (miles) | 20.3 [8.3–38.8] | 22.2 [10.1–40.2] | 19.7 [7.2–39.6] | 20.9 [7.3–39.6] |
| Pre-transfer operation or procedure use | 168 (3.2) | 54 (2.7) | 78 (3.6) | 86 (3.3) |
| Referral Hospital Characteristics | ||||
| Hospital size | ||||
| <100 beds | 2,058 (39.6) | 826 (41.7) | 862 (39.7) | 980 (37.5) |
| 100–250 beds | 1,447 (27.8) | 565 (28.5) | 520 (24.0) | 686 (26.3) |
| >250 beds | 1,697 (32.6) | 591 (29.8) | 789 (36.3) | 945 (36.2) |
| ICU beds | ||||
| 0 to 10 beds | 2,180 (41.9) | 859 (43.3) | 889 (41.0) | 1,029 (39.4) |
| 11 to 25 beds | 1,153 (22.1) | 460 (23.2) | 421 (19.4) | 530 (20.3) |
| >25 beds | 1,869 (35.9) | 663 (33.5) | 861 (39.7) | 1,052 (40.3) |
| Teaching hospital | 1,820 (35.0) | 625 (31.5) | 776 (35.7) | 958 (36.7) |
| Trauma center level | ||||
| Non-trauma center | 4,537 (87.2) | 1,737 (87.6) | 1,962 (90.3) | 2,329 (89.1) |
| Level 3 | 163 (3.1) | 76 (3.8) | 24 (1.1) | 26 (1.0) |
| Level 2 | 168 (3.2) | 61 (3.1) | 58 (2.7) | 65 (2.5) |
| Level 1 | 338 (6.5) | 108 (5.5) | 128 (5.9) | 194 (7.4) |
| Hospital ownership | ||||
| Non-profit | 3,699 (71.1) | 1,404 (70.8) | 1,328 (61.2) | 1,657 (63.5) |
| For-profit | 1,047 (20.1) | 415 (20.9) | 552 (25.4) | 642 (24.6) |
| Government | 456 (8.8) | 163 (8.2) | 291 (13.4) | 312 (12.0) |
| Rural hospital | 660 (12.7) | 264 (13.3) | 304 (14.0) | 338 (12.9) |
| Critical access hospital | 971 (18.7) | 380 (19.2) | 402 (18.5) | 465 (17.8) |
| U.S. Region ¶ | ||||
| Northeast | 1,817 (34.9) | 681 (34.4) | 669 (30.8) | 819 (31.3) |
| Midwest | 861 (16.5) | 329 (16.6) | 161 (7.4) | 133 (5.1) |
| South | 2,528 (48.6) | 972 (49.0) | 1,342 (61.8) | 1,662 (63.6) |
All results are median [interquartile range] or frequency (%).
Abbreviations: ED=emergency department, ICU=intensive care unit, U.S.=United States
Weekend transfers were defined as transfers that occurred on Saturday or Sunday. Weekday transfers were defined as transfers that occurred on Monday through Friday.
Nighttime transfers were defined as transfers that occurred between 6:00 PM and 6:00 AM the following day. Daytime transfers were defined as transfers that occurred between 6:01 AM and before 5:59 PM the same day. Data were not available for Massachusetts, Maryland, and Wisconsin. Data were available for 66.6% of all PAT episodes.
Race was available for 94.8% of all PAT episodes.
Hispanic ethnicity was available for 95.0% of all PAT episodes.
Insurance information was available for 99.8% of all PAT episodes.
U.S. regions were based on geographical regions defined by the U.S. Census Bureau.(29)
Figure 2:
Geographical distribution based on patient zip code of the interhospital transfer distances (in miles) of potentially avoidable transfers in eight U.S. states.
Factors associated with Potentially Avoidable Transfers
For patient-level factors, self-pay status was associated with 1.26-fold increased odds for PAT (vs. government insurance, Table 4). Among patient comorbidities, a history of psychiatric or substances use disorder was associated with 1.14-fold increased odds of PAT (Table 4). PAT was associated with most EGS diagnosis categories, especially intestinal obstruction, which was associated with 4.73-fold increased odds of PAT (Table 4).
Table 4:
Multilevel Analysis of Patient-level Factors Associated with Potentially Avoidable Transfers
| Multilevel Model | |||||
|---|---|---|---|---|---|
| Patient Characteristics | Potentially Avoidable Transfer (N=7,188) | Non-potentially Avoidable Transfer (N=19,093) | Univariate Model p-value | OR (95% CI) | Multilevel Model p-value |
| Age ≥65, years | 2,639 (36.7) | 9,238 (48.4) | <0.001 | 0.69 (0.64, 0.74) | <0.001 |
| Female | 3,691 (51.4) | 9,829 (51.5) | 0.73 | 0.96 (0.90, 1.02) | 0.17 |
| Race * | |||||
| White | 5,353 (78.6) | 14,633 (81.1) | - | 1 [Reference] | - |
| Black | 971 (14.3) | 2,255 (12.5) | 0.001 | 1.08 (0.99, 1.18) | 0.10 |
| Other | 487 (7.2) | 1,155 (6.4) | 0.04 | 0.99 (0.88, 1.13) | 0.93 |
| Hispanic † | 594 (8.7) | 1,255 (6.9) | <0.001 | 1.16 (1.03, 1.31) | 0.02 |
| Insurance ‡ | |||||
| Government insurance | 4,658 (64.9) | 13,362 (70.0) | - | 1 [Reference] | - |
| Private insurance | 1,854 (25.8) | 4,538 (23.8) | <0.001 | 0.88 (0.82, 0.96) | 0.002 |
| Self-pay | 402 (5.6) | 621 (3.3) | <0.001 | 1.26 (1.09, 1.45) | 0.002 |
| Other | 263 (3.7) | 568 (3.0) | <0.001 | 1.01 (0.86, 1.19) | 0.87 |
| Medical Comorbidities | |||||
| Cardiovascular disease | 3,828 (53.3) | 11,671 (61.1) | <0.001 | 0.98 (0.92, 1.05) | 0.62 |
| Pulmonary disease | 1,335 (18.6) | 4,410 (23.1) | <0.001 | 0.86 (0.80, 0.93) | <0.001 |
| Neurological disease | 562 (7.8) | 2,051 (10.7) | <0.001 | 0.74 (0.66, 0.82) | <0.001 |
| Metabolic disease | 3,390 (47.2) | 11,320 (59.3) | <0.001 | 0.65 (0.61, 0.69) | <0.001 |
| Renal disease | 173 (2.4) | 568 (3.0) | 0.44 | 0.89 (0.74, 1.08) | 0.25 |
| Hepatic disease | 536 (7.5) | 1,456 (7.6) | 0.90 | 0.93 (0.83, 1.05) | 0.23 |
| Malignancy | 377 (5.2) | 1,602 (8.4) | <0.001 | 0.68 (0.60, 0.77) | <0.001 |
| Coagulopathy/Anemia | 1,412 (19.6) | 4,867 (25.5) | <0.001 | 0.76 (0.70, 0.82) | <0.001 |
| Psychiatric/Substance use disorders | 1,062 (14.8) | 2,288 (12.0) | <0.001 | 1.14 (1.04, 1.24) | 0.005 |
| EGS Diagnosis Category § | |||||
| Resuscitation | 153 (2.1) | 932 (4.9) | 0.69 | 0.94 (0.73, 1.20) | 0.61 |
| General abdominal conditions | 1,105 (15.4) | 2,062 (10.8) | <0.001 | 3.36 (2.77, 1.20) | <0.001 |
| Intestinal obstruction | 941 (13.1) | 1,120 (5.9) | <0.001 | 4.73 (3.87, 5.77) | <0.001 |
| Upper GI | 890 (12.4) | 4,290 (22.5) | 0.005 | 1.14 (0.94, 1.39) | 0.17 |
| Hepatopancreaticobiliary | 74 (1.0) | 229 (1.2) | <0.001 | 2.01 (1.45, 2.78) | <0.001 |
| Colorectal | 1,212 (16.9) | 2,972 (15.6) | <0.001 | 2.50 (2.07, 3.02) | <0.001 |
| Hernia | 169 (2.4) | 273 (1.4) | <0.001 | 3.00 (2.29, 3.93) | <0.001 |
| Soft tissue | 163 (2.3) | 695 (3.6) | 0.002 | 1.57 (1.22, 2.03) | <0.001 |
| Vascular | 502 (7.0) | 965 (5.1) | <0.001 | 2.80 (2.27, 3.46) | <0.001 |
| Cardiothoracic | 1,819 (25.3) | 4,539 (23.8) | <0.001 | 2.44 (2.03, 2.94) | <0.001 |
| Other | 160 (2.2) | 1,016 (5.3) | - | - | - |
Abbreviations: AIDS=acquired immunodeficiency syndrome, EGS=emergency general surgery, GI=gastrointestinal
Race was available for 94.6% of all transfer episodes.
Hispanic ethnicity was available for 94.7% of all transfer episodes.
Insurance information was available for 99.9% of all transfer episodes.
EGS diagnosis categories as defined by the American Association for the Surgery of Trauma.(1) “Other” EGS diagnosis category omitted due to collinearity.
For hospital-level factors at the referring hospital, level 1 trauma center designation was associated with 1.24-fold increased odds of PAT (vs. non-trauma centers, Table 5). Increased odds of PAT was also associated with critical access hospitals (vs. non-critical access) and hospitals in the southern region (vs. Northeast region, Table 5).
Table 5:
Multilevel Analysis of Referring Hospital factors associated with Potentially Avoidable Transfers
| Multilevel Model | |||||
|---|---|---|---|---|---|
| Referring Hospital Characteristics | Potentially Avoidable Transfer (N=7,188) | Non-potentially Avoidable Transfer (N=19,093) | Univariate Model p-value | OR (95% CI) | Multilevel Model p-value |
| Hospital size | |||||
| <100 beds | 2,884 (40.1) | 7,791 (40.8) | - | 1 [Reference] | - |
| 100–250 beds | 2,012 (28.0) | 6,112 (32.0) | 0.04 | 0.86 (0.76, 0.98) | 0.02 |
| >250 beds | 2,288 (31.9) | 5,186 (27.2) | <0.001 | 1.04 (0.88, 1.24) | 0.62 |
| ICU beds | |||||
| 0 to 10 beds | 3,039 (42.3) | 8,416 (44.1) | - | 1 [Reference] | - |
| 11 to 25 beds | 1,613 (22.5) | 4,708 (24.7) | 0.39 | 1.04 (0.92, 1.18) | 0.54 |
| >25 beds | 2,532 (35.2) | 5,965 (31.3) | <0.001 | 1.10 (0.94, 1.30) | 0.24 |
| Teaching hospital | 2,445 (34.0) | 6,050 (31.7) | 0.002 | 1.06 (0.96, 1.16) | 0.27 |
| Trauma center level | |||||
| Non-trauma center | 6,274 (87.3) | 16,909 (88.6) | - | 1 [Reference] | - |
| Level 3 | 239 (3.3) | 728 (3.8) | 0.20 | 0.92 (0.76, 1.13) | 0.44 |
| Level 2 | 229 (3.2) | 583 (3.1) | 0.73 | 0.97 (0.80, 1.19) | 0.80 |
| Level 1 | 446 (6.2) | 873 (4.6) | <0.001 | 1.24 (1.05, 1.47) | 0.01 |
| Hospital ownership | |||||
| Non-profit | 5,103 (71.0) | 13,844 (72.5) | - | 1 [Reference] | - |
| For profit | 1,462 (20.4) | 3,556 (18.6) | 0.04 | 1.06 (0.97, 1.17) | 0.20 |
| Government | 619 (8.6) | 1,689 (8.9) | 0.83 | 0.89 (0.79, 1.01) | 0.07 |
| Rural hospital | 924 (12.9) | 2,637 (13.8) | 0.04 | 0.82 (0.72, 0.93) | 0.002 |
| Critical access hospital | 1,351 (18.8) | 3,388 (17.8) | 0.47 | 1.30 (1.15, 1.47) | <0.001 |
| U.S. Region * | |||||
| Northeast | 2,498 (34.8) | 7,028 (36.8) | - | 1 [Reference] | - |
| Midwest | 1,190 (16.6) | 3,480 (18.2) | 0.27 | 0.94 (0.85, 1.05) | 0.27 |
| South | 3,500 (48.7) | 8,585 (45.0) | 0.001 | 1.14 (1.05, 1.24) | 0.002 |
Abbreviations: ICU=intensive care unit, U.S.=United States
U.S. regions were based on geographical regions defined by the U.S. Census Bureau.(29)
Referring hospital-related characteristics accounted for 36.1% of variability in PAT. In contrast, patient-related factors accounted for 22.6% of PAT variability. For transfers during off- hours (weekend, nighttime), referring hospital characteristics accounted for increasing proportions of variation in PAT (weekend: 41.4%, nighttime: 63.6%).
Exploratory analyses
Trauma center level of the referring hospital significantly modified associations between PAT and patient insurance status and nighttime transfer (p<0.05). At level 1 trauma centers, increased odds of PAT was associated with self-pay (vs. government insurance, OR: 1.86, 95% CI: 1.19–2.90, p=0.007) and nighttime transfer (vs. daytime transfer, OR: 1.35, 95% CI: 1.07–1.69, p=0.01). Whereas, at level 3 trauma centers, PAT was not significantly associated with self-pay or daytime transfer. There was no significant interaction between trauma center level and Hispanic ethnicity, psychiatric comorbidities, hospital ownership, or geographical region.
Hospital geographical region significantly modified associations between PAT and patient insurance status and psychiatric comorbidities (p<0.05). In the Southern region, increased odds of PAT was associated with self-pay (vs. government insurance, OR: 1.73, 95% CI: 1.19–2.52, p=0.004) and history of psychiatric disorders (vs. no psychiatric comorbidities, OR: 1.37, 95% CI: 1.16–1.63, p<0.001). Whereas in the Midwest, PAT was not significantly associated with self-pay or psychiatric comorbidities. There was no significant interaction between Hispanic ethnicity, Southern geographical region and hospital ownership, critical access hospital designation, or off-hours transfer. There was also no significant interaction between Hispanic ethnicity and patient insurance status.
Discussion
In this large, population-based cohort including both nonoperatively and operatively managed EGS patients from multiple healthcare networks and geographic regions, we characterized patterns in EGS interhospital transfers and identified factors associated with potentially avoidable transfers. PAT occurred in over 1 in 4 EGS transfers, especially in transfers originating from the ED. Notably, hospital-related, instead of patient-related, factors accounted for more substantial variation in PAT, particularly during off-hours (weekend, nighttime). These findings reveal opportunities to improve EGS outcomes and resource utilization by targeting hospital organization or policies within the interhospital transfer system, through more coordinated and standardized guidelines.
Unlike trauma, there is no well-established comprehensive national registry for patients with EGS conditions despite the increasing incidence, morbidity, and mortality associated with this population. Examining the entire scope of EGS is critical to understanding the disease burden that our healthcare system needs to support.(1, 2, 15) This was our first objective, as our current understanding of the EGS interhospital transfer system is fragmented, often due to limitations in available datasets.(2–4, 6, 7, 9, 12, 13) Consistent with prior literature,(2, 11, 33) we found that management practices are variable, with <50% of EGS episodes involving procedures or operations. We agree with Havens and colleagues(33) on the call to establish a national EGS data registry that includes both operatively- and nonoperatively-managed, transferred and non-transferred patients, through which research and national quality improvement initiatives could be developed with consistency.
Our results show that over 1 in 4 transfers are potentially avoidable, especially for transfers from the ED. This is consistent with previously reported incidence of unnecessary transfers to two tertiary institutions.(3, 11) Hospital-level factors account for substantial variation in PAT, particularly during off-hours. This could be a consequence of staffing limitations and variable on-call models utilized across institutions.(34) Lack of general surgery for emergency care is one of the most commonly cited reasons for transfer in EGS,(3, 8, 11) with 37% of ED directors nationally reporting inadequate general surgery coverage.(35, 36) Inadequate on-call specialist coverage was associated with increased outgoing transfers, particularly at rural hospitals, while hospitals utilizing Acute Care Surgery (ACS) models receive more interhospital transfers for EGS.(34–36) However, the type of EGS coverage model is highly variable across institutions.(34) In a national survey of university-affiliated hospitals, 31% implemented an ACS model with a dedicated EGS team, 52% maintained a general surgeon on-call model, while 15% utilized a hybrid model.(34) On-call and staffing structure could be a modifiable target to improve EGS care delivery and resource utilization.
PAT are common during off-hours (weekend, nighttime). The “off-hour effect” on outcomes has been described, in which increased morbidity and mortality were associated with hospital admissions, trauma presentations, and certain EGS operations performed during the weekend or nighttime.(22, 23, 37, 38) Staffing patterns with more in-house availability of general surgeons were associated with better outcomes.(39) However, the “off-hour effect” of surgical staffing variability on transfer triage is unknown and warrants further investigation.
Critical access hospitals, which have ≤25 acute care beds and are ≥35 miles from another hospital by primary road or ≥15 miles by secondary road,(40) are associated with increased odds of PAT. Critical access hospitals are capable of providing excellent general surgery care for procedures such as appendectomy, cholecystectomy, colorectal resections, and hernia repairs.(41) However, up to 60% of surgical patients are transferred from critical access hospitals.(42) Since a hospital’s temporal variability in specialist coverage, instead of its overall capability of providing excellent surgical care, may contribute to this phenomenon, structured support for non-tertiary hospitals is essential when designing effective EGS regionalization.(2, 43) Tele-medicine has emerged for peri-operative care coordination and critical care.(44, 45) Tele-consultation from surgeons and specialists could assist non-tertiary hospitals that lack surgical coverage with transfer and management decisions for patients with EGS conditions.(2, 8, 11) Our results highlight opportunities to improve the transfer triage process and resource utilization, possibly through multidisciplinary initiatives among ED providers, specialists, and regional call centers.
Unexpectedly, level 1 trauma center designation was associated with PAT in EGS. However, transfers from level 1 trauma centers were uncommon, and a variety of factors could contribute to this finding. These include specialist availability (e.g., cardiothoracic and vascular surgery), insurance agreements, bed capacity limitations, continuity of care, and established inter-institution relationships. Further research examining these potential contributing factors is necessary to determine the reasons underlying this unexpected finding. Although level 1 trauma center designation is a proxy for well-resourced institutions with specialist coverage, it does not guarantee dedicated in-house EGS coverage, specialty service availability (e.g., cardiothoracic or vascular surgery), or an ACS staffing model.(34) Institutional and temporal variability in EGS and specialty service availability could contribute to this finding, as well as insurance agreements and out-of-network reimbursement policies,(46) given that self-pay patients had higher odds of undergoing PAT compared to those with government insurance at level 1 trauma centers.Hospital bed capacity limitations and continuity of care may also influence transfer decisions and destination hospital selection. Additionally, established inter-institution partnerships, instead of individualized efforts to optimize each patient’s care, may influence EGS transfer patterns.(47) Further research examining these potential contributing factors is needed to understand this phenomenon.
Stratification of hospitals based on capability to provide EGS care, EGS volume, and outcomes could help standardize transfer triage and facilitate appropriate destination hospital selection. Similar to the trauma center verification system,(48) we can develop a stratified system for EGS care. Elements to consider include resources that are often required for comprehensive EGS care, such as 24-hour coverage by general surgery, interventional radiology, gastroenterology, anesthesiology, operative or procedure room capabilities, and critical care capacity. Annual EGS volume, other measures of EGS-specific surgical quality such as failure to rescue, training opportunities, and quality improvement programs should also be considered.(2, 25, 26) Development of a standardized system to organize hospital resources, quality, and infrastructure for EGS care could improve transfer decision-making and optimize regionalization.
Patient self-pay status, particularly at level 1 trauma centers and Southern region hospitals, is associated with PAT in EGS. In California, “good payer” patients (self-pay, private and government insurance, workers’ compensation) were more likely to be transferred than “poor payer” patients.(46) However, the transfers examined by Green et al. (2005) were not specific to EGS conditions or PAT.(46) There could be socioeconomic, institutional, political, or regional factors related to this observation, warranting further network- and region-specific research. Aligning reimbursement practices and incentives with optimal patient care may be important to improving EGS care delivery.
Psychiatric disorders and substance abuse, which are associated with PAT, are prevalent, undertreated, and associated with high social burden.(49) Psychiatric disorders have medical and social consequences that may influence clinicians’ evaluation of disease severity and decision to transfer a patient. Further research on disease risk stratification and physician decision-making regarding transfer triage is needed.
Validated risk stratification based on severity assessment of EGS conditions could be helpful in guiding transfer triage decision-making. The AAST proposed a model for a severity grading system for 16 common EGS conditions, but a validated risk stratification tool for all EGS conditions does not currently exist.(33, 50) Although some studies show promising results in utilization of this severity grading system for EGS conditions like diverticulitis and small bowel obstruction,(51, 52) further research and validation of a risk stratification tool that could be used for all EGS conditions are needed, which could guide transfer criteria development.
There are limitations to this study. Administrative data lack clinical information on disease severity at presentation, reasons for transfer, and readmission tracking. As reason for transfer and readmission tracking are not available in administrative claims datasets, we were unable to identify whether an admission or transfer was due to continuity of care and readmission for a recent postoperative complication or due to socioeconomic reasons (e.g., patient request to be closer to home). Although the precise reason for transfer is not available in administrative claims datasets utilized in this study, in other published studies on interhospital transfers to tertiary centers in EGS, over 80% of transfers were not due to “continuity of care” or “patient request”, and these reasons were not among the top three reasons for transfer.(3, 11) Thus, we expect these transfers to be less common overall. Further research could utilize different methodology like qualitative interviews and surveys to examine physician and patient transfer decisions. Despite the limitations inherent to administrative claims data, these data included different states and geographic regions, ED and inpatient encounters, and care at referring and destination hospitals to provide a more comprehensive understanding of interhospital transfers in EGS.
There may also be concerns about misclassification of patients with EGS conditions as the scope of EGS is broad, involving different organ systems and systemic diseases, in addition to variable management practices, from nonoperative to operative. However, we utilized the EGS definition and corresponding ICD-9-CM diagnosis codes developed by the AAST Committee on Severity Assessment and Patient Outcomes through a rigorous iterative Delphi methodology.(1) These conditions and associated presentation and sequelae (e.g., abdominal pain, shock) are within the clinical realm of EGS evaluation and management as defined by the American Board of Surgery, and thus were included in our evaluation of EGS care.(1)
We collected data from multiple sources for more comprehensive hospital characteristics, but lacked data on EGS staffing model, off-hours staffing, and specialist availability at each hospital. Further institution- and region-specific research examining effects of EGS staffing variability on transfer patterns and outcomes is needed. Referring hospital characteristics instead of destination hospital characteristics were examined in our model because our objective was to identify intervention targets prior to transfer at the referring hospital, which would be more actionable. Future work should examine transfer patterns to different types of destination hospitals and examine decision-making processes for appropriate destination hospital selection. Although higher overtriage has been accepted to prevent undertriage in trauma, benchmarks for PAT in EGS is not established and likely a distinct concept reflecting lower-value care, which requires further investigation.(53) Additional research on strategies to design effective transfer systems and targeted quality improvement efforts is needed, with the goal of guiding healthcare policy to optimize EGS care delivery.
Conclusions
We characterized interhospital transfer patterns in EGS using population-based claims data, identifying inefficiencies in EGS care delivery with opportunities for targeted quality improvement. Over 1 in 4 EGS transfers are potentially avoidable, with hospital-level factors accounting for substantial portion of its variability, especially during off-hours. Several patient- level and hospital-level factors are associated with PAT. These factors may be useful in guiding necessary future research, quality improvement, policy, and infrastructure development to improve EGS transfer systems and patient outcomes. Development of a comprehensive national EGS registry to include nonoperative and operative management, hospital categorization based on EGS capabilities and outcomes, validated risk stratification tools for EGS conditions, and policies to align financial incentives with appropriate patient care are needed. Other strategies to optimize the interhospital transfer system to improve EGS outcomes may include guidelines for efficient transfer triage, protocols to facilitate appropriate transfer decisions and communication, multi-institutional collaboration among stakeholders, hospitals, and call centers, and support for non-tertiary centers in the care of EGS patients, such as tele-medicine consultation.
Supplementary Material
Acknowledgements
The authors would like to thank Andrew-Paul Deeb for engaging discussions during the development of this project.
Funding/Financial Support
Cindy Y. Teng was supported by the National Institutes of Health (training grant number T32HL007820). Jeremy M. Kahn was supported by the National Institutes of Health and National Heart, Lung, and Blood Institute (grant number R35HL144804).
Footnotes
Conflict of Interest/Disclosures
The authors have no related conflicts of interest to declare.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Shafi S, Aboutanos MB, Agarwal S Jr., Brown CV, Crandall M, Feliciano DV, et al. Emergency general surgery: definition and estimated burden of disease. J Trauma Acute Care Surg. 2013;74(4):1092–7. [DOI] [PubMed] [Google Scholar]
- 2.Santry H, Kao LS, Shafi S, Lottenberg L, Crandall M. Pro-con debate on regionalization of emergency general surgery: controversy or common sense? Trauma Surg Acute Care Open. 2019;4(1):e000319. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Philip JL, Saucke MC, Schumacher JR, Fernandes-Taylor S, Havlena J, Greenberg CC, et al. Characteristics and Timing of Interhospital Transfers of Emergency General Surgery Patients. J Surg Res. 2019;233:8–19. [DOI] [PubMed] [Google Scholar]
- 4.Yelverton S, Rozario N, Matthews BD, Reinke CE. Interhospital transfer for emergency general surgery: An independent predictor of mortality. Am J Surg. 2018;216(4):787–92. [DOI] [PubMed] [Google Scholar]
- 5.Khubchandani JA, Shen C, Ayturk D, Kiefe CI, Santry HP. Disparities in access to emergency general surgery care in the United States. Surgery. 2018;163(2):243–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Keeven DD, Davenport DL, Bernard AC. Escalation of mortality and resource utilization in emergency general surgery transfer patients. J Trauma Acute Care Surg. 2019;87(1):43–8. [DOI] [PubMed] [Google Scholar]
- 7.Castillo-Angeles M, Uribe-Leitz T, Jarman M, Jin G, Feeney T, Salim A, et al. Transferred Emergency General Surgery Patients Are at Increased Risk of Death: A NSQIP Propensity Score Matched Analysis. J Am Coll Surg. 2019;228(6):871–7. [DOI] [PubMed] [Google Scholar]
- 8.Broman KK, Poulose BK, Phillips SE, Ehrenfeld JM, Sharp KW, Pierce RA, et al. Unnecessary Transfers for Acute Surgical Care: Who and Why? Am Surg. 2016;82(8):672–8. [PubMed] [Google Scholar]
- 9.Lucas DJ, Ejaz A, Haut ER, Spolverato G, Haider AH, Pawlik TM. Interhospital transfer and adverse outcomes after general surgery: implications for pay for performance. J Am Coll Surg. 2014;218(3):393–400. [DOI] [PubMed] [Google Scholar]
- 10.Philip JL, Yang DY, Wang X, Fernandes-Taylor S, Hanlon BM, Schumacher J, et al. Effect of Transfer Status on Outcomes of Emergency General Surgery Patients. Surgery. 2020;168(2):280–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bruenderman EH, Block SB, Kehdy FJ, Benns MV, Miller KR, Motameni A, et al. An evaluation of emergency general surgery transfers and a call for standardization of practices. Surgery. 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ingraham A, Wang X, Havlena J, Hanlon B, Saucke M, Schumacher J, et al. Factors Associated With the Interhospital Transfer of Emergency General Surgery Patients. J Surg Res. 2019;240:191–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lauerman MH, Herrera AV, Albrecht JS, Chen HH, Bruns BR, Tesoriero RB, et al. Interhospital Transfers with Wide Variability in Emergency General Surgery. Am Surg. 2019;85(6):595–600. [PMC free article] [PubMed] [Google Scholar]
- 14.Shah AA, Haider AH, Zogg CK, Schwartz DA, Haut ER, Zafar SN, et al. National estimates of predictors of outcomes for emergency general surgery. J Trauma Acute Care Surg. 2015;78(3):482–90; discussion 90–1. [DOI] [PubMed] [Google Scholar]
- 15.Gale SC, Shafi S, Dombrovskiy VY, Arumugam D, Crystal JS. The public health burden of emergency general surgery in the United States: A 10-year analysis of the Nationwide Inpatient Sample--2001 to 2010. J Trauma Acute Care Surg. 2014;77(2):202–8. [DOI] [PubMed] [Google Scholar]
- 16.Nationwide HCUP Databases: Agency for Healthcare Research and Quality; [Available from: https://www.hcup-us.ahrq.gov/databases.jsp.
- 17.Iwashyna TJ, Christie JD, Moody J, Kahn JM, Asch DA. The structure of critical care transfer networks. Med Care. 2009;47(7):787–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ames SG, Davis BS, Angus DC, Carcillo JA, Kahn JM. Hospital Variation in Risk- Adjusted Pediatric Sepsis Mortality. Pediatr Crit Care Med. 2018;19(5):390–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ames SG, Davis BS, Marin JR, Fink EL, Olson LM, Gausche-Hill M, et al. Emergency Department Pediatric Readiness and Mortality in Critically Ill Children. Pediatrics. 2019;144(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Reinke CE, Thomason M, Paton L, Schiffern L, Rozario N, Matthews BD. Emergency general surgery transfers in the United States: a 10-year analysis. J Surg Res. 2017;219:128–35. [DOI] [PubMed] [Google Scholar]
- 21.ICD-10: Centers for Medicare and Medicaid; [Available from: https://www.cms.gov/Medicare/Coding/ICD10. [PubMed]
- 22.Egol KA, Tolisano AM, Spratt KF, Koval KJ. Mortality rates following trauma: The difference is night and day. J Emerg Trauma Shock. 2011;4(2):178–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Althoff FC, Wachtendorf LJ, Rostin P, Santer P, Schaefer MS, Xu X, et al. Effects of night surgery on postoperative mortality and morbidity: a multicentre cohort study. BMJ Qual Saf. 2020. [DOI] [PubMed] [Google Scholar]
- 24.Surgery Flags Software for ICD-9-CM: Agency for Healthcare Research and Quality; [Available from: https://www.hcup-us.ahrq.gov/toolssoftware/surgflags/surgeryflags.jsp.
- 25.Wakeam E, Hevelone ND, Maine R, Swain J, Lipsitz SA, Finlayson SR, et al. Failure to rescue in safety-net hospitals: availability of hospital resources and differences in performance. JAMA Surg. 2014;149(3):229–35. [DOI] [PubMed] [Google Scholar]
- 26.Ghaferi AA, Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients. Ann Surg. 2009;250(6):1029–34. [DOI] [PubMed] [Google Scholar]
- 27.AHA Annual Survey Database: American Hospital Association; [Available from: https://www.ahadata.com/aha-annual-survey-database. [Google Scholar]
- 28.Searching for Verified Trauma Centers: The Committee on Trauma, American College of Surgeons; [Available from: https://www.facs.org/search/trauma-centers. [Google Scholar]
- 29.Census Regions and Divisions of the United States: US Census Bureau; [Available from: https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf.
- 30.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27. [DOI] [PubMed] [Google Scholar]
- 31.Zhou Q, Rosengart MR, Billiar TR, Peitzman AB, Sperry JL, Brown JB. Factors Associated With Nontransfer in Trauma Patients Meeting American College of Surgeons’Criteria for Transfer at Nontertiary Centers. JAMA Surg. 2017;152(4):369–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Brown JB, Rosengart MR, Billiar TR, Peitzman AB, Sperry JL. Distance matters: Effect of geographic trauma system resource organization on fatal motor vehicle collisions. J Trauma Acute Care Surg. 2017;83(1):111–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Havens JM, Neiman PU, Campbell BL, Croce MA, Spain DA, Napolitano LM. The Future of Emergency General Surgery. Ann Surg. 2019;270(2):221–2. [DOI] [PubMed] [Google Scholar]
- 34.Santry HP, Madore JC, Collins CE, Ayturk MD, Velmahos GC, Britt LD, et al. Variations in the implementation of acute care surgery: results from a national survey of university-affiliated hospitals. J Trauma Acute Care Surg. 2015;78(1):60–7; discussion 7–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Menchine MD, Baraff LJ. On-call specialists and higher level of care transfers in California emergency departments. Acad Emerg Med. 2008;15(4):329–36. [DOI] [PubMed] [Google Scholar]
- 36.Rao MB, Lerro C, Gross CP. The shortage of on-call surgical specialist coverage: a national survey of emergency department directors. Acad Emerg Med. 2010;17(12):1374–82. [DOI] [PubMed] [Google Scholar]
- 37.Pauls LA, Johnson-Paben R, McGready J, Murphy JD, Pronovost PJ, Wu CL. The Weekend Effect in Hospitalized Patients: A Meta-Analysis. J Hosp Med. 2017;12(9):760–6. [DOI] [PubMed] [Google Scholar]
- 38.Hajibandeh S, Hajibandeh S, Satyadas T. Impact of weekend effect on postoperative mortality in patients undergoing emergency General surgery procedures: Meta-analysis of prospectively maintained national databases across the world. Surgeon. 2020;18(4):231–40. [DOI] [PubMed] [Google Scholar]
- 39.Daniel VT, Rushing AP, Ingraham AM, Ricci KB, Paredes AZ, Diaz A, et al. Association between operating room access and mortality for life-threatening general surgery emergencies. J Trauma Acute Care Surg. 2019;87(1):35–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.H.R.2015 - Balanced Budget Act of 1997: 105th Congress (1997–1998); [Available from: https://www.congress.gov/bill/105th-congress/house-bill/2015.
- 41.Ibrahim AM, Hughes TG, Thumma JR, Dimick JB. Association of Hospital Critical Access Status With Surgical Outcomes and Expenditures Among Medicare Beneficiaries. JAMA. 2016;315(19):2095–103. [DOI] [PubMed] [Google Scholar]
- 42.Musgrove KA, Abdelsattar JM, LeMaster SJ, Ballou MC, Kappel DA, Borgstrom DC. Optimal Resources for Rural Surgery. Am Surg. 2020;86(9):1057–61. [DOI] [PubMed] [Google Scholar]
- 43.Teng CY, Ingraham AM, Zuckerbraun BS. Emergency General Surgery-To Regionalize, or Not to Regionalize, That Is the Question. JAMA Surg. 2020;155(9):849–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Huang EY, Knight S, Guetter CR, Davis CH, Moller M, Slama E, et al. Telemedicine and telementoring in the surgical specialties: A narrative review. Am J Surg. 2019;218(4):760–6. [DOI] [PubMed] [Google Scholar]
- 45.Vranas KC, Slatore CG, Kerlin MP. Telemedicine Coverage of Intensive Care Units: A Narrative Review. Ann Am Thorac Soc. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Green A, Showstack J, Rennie D, Goldman L. The relationship of insurance status, hospital ownership, and teaching status with interhospital transfers in California in 2000. Acad Med. 2005;80(8):774–9. [DOI] [PubMed] [Google Scholar]
- 47.Lomi A, Pallotti F. Relational collaboration among spatial multipoint competitors. Social Networks. 2012;34(1):101–11. [Google Scholar]
- 48.Verification, Review, and Consultation (VRC) Program: American College of Surgeons; [Available from: https://www.facs.org/quality-programs/trauma/tqp/center-programs/vrc.
- 49.Vigo D, Thornicroft G, Atun R. Estimating the true global burden of mental illness. The Lancet Psychiatry. 2016;3(2):171–8. [DOI] [PubMed] [Google Scholar]
- 50.Tominaga GT, Staudenmayer KL, Shafi S, Schuster KM, Savage SA, Ross S, et al. The American Association for the Surgery of Trauma grading scale for 16 emergency general surgery conditions: Disease-specific criteria characterizing anatomic severity grading. J Trauma Acute Care Surg. 2016;81(3):593–602. [DOI] [PubMed] [Google Scholar]
- 51.Hernandez MC, Birindelli A, Bruce JL, Buitendag JJP, Kong VY, Beuran M, et al. Application of the AAST EGS Grade for Adhesive Small Bowel Obstruction to a Multi-national Patient Population. World J Surg. 2018;42(11):3581–8. [DOI] [PubMed] [Google Scholar]
- 52.Shafi S, Priest EL, Crandall ML, Klekar CS, Nazim A, Aboutanos M, et al. Multicenter validation of American Association for the Surgery of Trauma grading system for acute colonic diverticulitis and its use for emergency general surgery quality improvement program. J Trauma Acute Care Surg. 2016;80(3):405–10; discussion 10–1. [DOI] [PubMed] [Google Scholar]
- 53.American College of Surgeons. Resources for Optimal Care of the Injured Patient 2014/Resources Repository [Available from: https://www.facs.org/Quality-Programs/Trauma/TQP/center-programs/VRC/resources.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.