Abstract
Objective
Children's HeartLink and the University of California San Francisco (UCSF) have partnered with Vietnam National Children's Hospital (VNCH). Compared with low- and middle-income countries (LMICs) enrolled in the International Quality Improvement Collaborative for Congenital Heart Disease, VNCH is a high performing center. Our goal is to compare outcomes of VNCH with North American centers leveraging data from the Pediatric Cardiac Critical Care Consortium (PC4) in 2023.
Methods
Children's HeartLink established a partnership with VNCH in 2017 with a consistent senior-level medical volunteer team from UCSF (Dr Sano currently at Showa University Hospital) engaged in semiannual training visits and weekly case conferences. In-person visits have evolved to hands-off teaching. Areas of improvement are identified, and outcomes presented at the subsequent visit. International Quality Improvement Collaborative and PC4 perform site specific audits to confirm data accuracy. Data presented are based on primary operations. Conversion from Risk Adjustment for Congenital Heart Surgery (RACHS-1) to STAT is as follows: RACHS-1 1-4 equal STAT 1-4; RACHS-1 5-6 equal STAT 5.
Results
Results represent 8 on-site visits and 109 case discussions. VNCH performed 1,056 operations, and PC4 represents 14,506 operations in 2023. VNCH had comparable outcomes, except higher surgical mortality for infants with prematurity and noncardiac abnormalities, and a higher rate of surgical site infections. VNCH performed fewer operations on patients with noncardiac abnormalities and higher STAT scores.
Conclusions
Partnership of highly resourced North American programs with underresourced LMIC centers can elevate congenital heart surgery outcomes. VNCH is a high-volume center with respectable outcomes. Future focus will be higher STAT level and neonatal surgeries.
Keywords
Introduction
Surgery for patients with congenital heart disease (CHD), particularly those with complex or early congenital heart surgeries, are highly resourced in many countries. Underresourced hospitals from low-and middle-income countries (LMICs) struggle with inadequately sterile facilities and bedspace, lack of medical equipment, and reduced access to pharmaceuticals and insufficient staffing. 1 Fetal diagnosis remains less common in LMICs; therefore, patients often present with late diagnosis in heart failure, with underweight status, and/or with intercurrent infections. In addition, the rigorous medical training and subspecialty training are more commonly replaced with on-the-job training. Outcomes at centers in LMICs are often unexpectedly poorer than those in well-resourced countries. 1
There are several approaches to improving medical care in underresourced countries, most leveraging charity organizations with volunteer time from medical personnel. Specific to cardiac programs, models range from (1) surgical mission trips to developing countries where the visiting team provides all the operative and perioperative care for a very short period of time (operating on 10-20 cases per weekly visit), (2) improving the local infrastructure for open-heart surgeries, catheterization interventions, postoperative care, and training the local team, (3) flying children to developed countries to get surgery, and (4) remote telehealth support.2,3 The specific approach depends upon local resources and experience. For all approaches, language is often a critical barrier.
The International Quality Improvement Collaborative (IQIC) for Congenital Heart Surgery: Improving Care in LMICs was established in 2007 to collect information about CHD surgery in LMICs to identify specific risk factors for mortality of children receiving cardiac care, to evaluate program performance in emergent world programs, and to engage in quality improvement. 4 Currently, over 70 sites in more than 25 different countries are enrolled in the IQIC.
Children's HeartLink (CHL), a United States (US)-based nonprofit organization, fosters training partnership between top North American teaching and research institutions and programs in LMICs utilizing the train-the-trainer model. 5 The University of California San Francisco (UCSF), in partnership with CHL, approached a collaboration with Vietnam National Children's Hospital (VNCH) with a goal of local training to support program development. Since 2017, a dedicated senior-level team from UCSF, sponsored by CHL, has supported VNCH in Hanoi, Vietnam, with in-person and remote education. VNCH has benchmarked their outcomes through their participation in IQIC. Within IQIC, VNCH has become a high performing center with high surgical volume and relatively low mortality and complication rates.
Our specific aim was to compare VNCH outcomes with those of North American congenital heart surgical programs leveraging the Pediatric Cardiac Critical Care Consortium (PC4) data registry. 6 Our secondary aim was to determine if collaboration between a highly resourced academic heart center with formal specialized advanced training programs can facilitate comparable outcomes at an underresourced heart center with on-the-ground training.
Methods
Local Volume
Vietnam National Children′s Hospital was established in 1969 in Hanoi and is the single free-standing children's hospital in Vietnam and a tertiary referral hospital. VNCH serves over 30 million people in Northern Vietnam. 7 VNCH currently manages a daily census of about 2,000 inpatients and 4,000 outpatients. Established in 2016, the Heart Center sees about 120 to 160 outpatients a day and admits 5 to 12 patients into the hospital daily. There are a total of about 200 echocardiograms a day, 50 of which are performed on inpatients. VNCH performs an average number of 22 open-heart surgeries and 12 closed heart surgeries per week. The cardiac intensive care unit (CICU) has 20 beds for patients with open-heart surgeries and the cardiac medical intensive care unit (ICU) has 10 beds for both closed heart surgeries and preoperative patients.
Local Resources
The World Bank designates Vietnam as a lower middle-income country. VNCHl Heart Center has dedicated anesthesiologists, surgeons, intensivists, interventionalists, imaging staff (echocardiography, computer tomography, and magnetic resonance imaging with 3D capability), and dedicated cardiac nurses. There are five independent surgeons, seven cardiac anesthesiologists, and 22 cardiologists. Facilities include operating rooms, preoperative and cardiac medical intensive care unit, postoperative CICU, cardiac stepdown unit, and catheterization laboratory. Deficiencies due to lack of resources include lack of respiratory therapists, nutritionists, heart transplant, extracorporeal life support (ECLS) for neonates, in addition to limited inspired nitric oxide (iNO), and other crucial pharmaceuticals. ICU bedside nurses have low nurse-to-patient ratios. Commonly, an ICU nurse has 3 to 4 patients on mechanical ventilation and is expected to provide all respiratory care, prepare medications, and type and screen blood products.
Comparisons of resources between VNCH, IQIC centers, and recommended comprehensive and essential care centers in the US were done using surveys from centers in LMICs and Congenital Heart Surgeons’ Society (CHSS) recommendations publication.8,9 (Table 1).
Comparisons of Centers in LMICs, United States (CHSS Recommendations), and VNCH.
Abbreviations: CHSS, Congenital Heart Surgeons' Society; CICU, cardiac intensive care unit; CPB, cardiopulmonary bypass; ECMO, extracorporeal membrane oxygenation; HFOV, high-frequency oscillatory ventilation; iNO, inhaled nitric oxide; IQIC, International Quality Improvement Collaborative; LMICs, low-and-middle income countries; NR, not reported; NICU, neonatal intensive care unit; OR, operating room; PGE, prostaglandin; STS, Society of Thoracic Surgeons; TEE, transesophageal echocardiogram; VNCH, Vietnam National Children's Hospital.
IQIC includes 37 centers (54% upper-middle-income, 46% lower-middle-income).
High complexity = STAT 4 and 5, RACHS-1 3 to 6.
Collaboration
Children's HeartLink established a partnership with VNCH in 2017 to improve clinical knowledge, technical skills, quality initiatives, and benchmarked outcomes. Strategies include a consistent senior level medical volunteer team from UCSF (Dr Sano currently at Showa University Hospital) engaging in twice annual 5-to-10-day training visits and weekly remote case conferences. The medical volunteer team includes a surgeon, an anesthesiologist, a cardiologist (performs catheterization and echocardiography), a cardiac intensivist, an advance practice nurse, and a bedside nurse. Language interpretation is provided by congenital heart trained nurses, cardiologists, and intensivists. Areas of improvement are identified, and outcomes are presented at the subsequent visit. Hospital leadership actively participates. In-person visits have evolved from primary procedural and patient care management to hands-off teaching in the operating room, intensive care units, and catheterization laboratory.
Weekly virtual case conference discussions were initiated in January 2021 and continue to the present day. Typically, 2 to 3 patients are presented each week for advice on preoperative, surgical, and/or postoperative management. Additional ad hoc virtual case conferences are held for more urgent patient management situations. In addition, monthly virtual nursing conferences are held with UCSF bedside nurses and advance practice nurse practitioners.
Quality Improvement and Benchmarking
Children's HeartLink requires that VNCH benchmark their outcomes. To date, this has been accomplished leveraging the IQIC registry. Vietnam National Children’s Hospital submits deidentified diagnostic, procedural, and clinical information for all congenital heart surgeries in patients during a calendar-year period into a web-based data entry tool. 4 Boston Children's Hospital (BCH) maintains the database. At least 10% of all cases submitted into IQIC database are audited annually. IQIC audit process adjudicates 10 data fields, which includes type of surgery, surgery date, age at surgery, sex, prematurity, noncardiac anomaly, in-hospital mortality, patient's status (alive vs death) at 30 days postsurgery, surgical site infections (SSIs), and bacterial sepsis. VNCH (HNN) and remote (CPN) abstractors performed the 2023 VNCH audit with 117 cases reviewed, and with 94.9% accuracy determined. Data sharing is covered by site-specific data use agreements, center level, and BCH Institutional Review Board (IRB) approvals.
PC4 is a quality collaborative of congenital heart programs in North America. In 2023, 68 of the approximately 110 pediatric cardiac surgical centers in the US entered data into PC4. Patient-level data are submitted and audited. Center-level audit process includes both adjudication and primary chart review, with an accuracy rate of 99.4%. 10 Data are overseen by the University of Michigan. Data sharing is covered by site-specific data use agreements, center level, and University of Michigan IRB approvals.
Patient-level data submitted from VNCH to IQIC was compared with aggregate data from PC4 for calendar year 2023, based upon surgery type. Comparisons include demographic data, surgical data, and outcomes. There are a few important differences in data definitions. (1) IQIC defines premature infants as those who were born before 37 weeks’ gestation and had surgery at less than one year of age. PC4 prematurity was defined as age at surgery <30 days and born <37 weeks’ gestation. (2) IQIC leverages Risk Adjustment for Congenital Heart Surgery (RACHS-1). 11 PC4 leverages the Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery (STAT) mortality score. 12 STAT mortality scores are version STS v3.41. Comparisons between the two risk stratification methods were as follows: STAT 1-4 equaled RACHS-1 1-4, STAT 5 equaled RACHS-1 5-6. 13 (3) IQIC benchmarked report limits mortalities to patients with assigned RACHS-1 and age less than 18 years and includes unassigned RACHS-1 (n = 62) and all ages in morbidities. Patients with unassigned STAT scores (<1%) are excluded from PC4 benchmarked outcomes. Mortalities and morbidities were benchmarked with matched age groups. Both PC4 and IQIC have risk adjusted metrics; however, those cannot be directly compared between registries.
Statistical Analysis
Comparisons between the PC4 collaborative data and VNCH data were performed using the chi-square2 test. For age group and STAT/RACHS-1 comparisons, each age group or STAT/RACHS-1 category was compared with the rest of the study population. To facilitate comparison with VNCH's IQIC's data, volume, and mortality (Tables 2 and 3), patients over 18 years of age at time of surgery and without assigned RACHS-1 or STAT score were excluded. Outcome measures for the complications (Table 4) included patients over 18 years of age and those without a RACHS-1 or STAT score.
Comparisons Between VNCH and PC4 Aggregate Hospitals in Terms of Demographics and Surgical Data.
Abbreviations: IQIC, International Quality Improvement Collaborative; PC4, Pediatric Cardiac Critical Care Consortium; RACHS, Risk Adjustment for Congenital Heart Surgery; VNCH, Vietnam National Children’s Hospital.
Data are presented as n and %. All patients less than 18 years of age at time of surgery with an assigned STAT/RACHS-1 score.
Chi-square compares each age group to the rest of the study population.
IQIC prematurity was defined as age at surgery <1 year and born <37 weeks’ gestation. PC4 prematurity was defined as age at surgery < 30 days and born < 37 weeks’ gestation.
Chi-square compares each RACHS-1/STAT score to the rest of the study population.
Comparisons Between VNCH and PC4 Aggregate Hospitals in Terms of in-Hospital Mortality.
Abbreviations: IQIC: International Quality Improvement Collaborative; PC4, Pediatric Cardiac Critical Care Consortium; RACHS, Risk Adjustment for Congenital Heart Surgery; VNCH, Vietnam National Children’s Hospital.
Numerator/denominator (%). Includes all patients less than 18 years of age at time of surgery with an assigned RACHS-1/STAT score.
Chi-square compares each age group with the rest of the study population.
IQIC prematurity was defined as age at surgery <1 year and born <37 weeks’ gestation. PC4 prematurity was defined as age at surgery <30 days and born <37 weeks’ gestation.
Chi-square compares each RACHS-1/STAT score with the rest of the study population; PC4 patients with unassigned STAT score excluded (n = 794).
Comparisons Between VNCH and PC4 Aggregate Hospitals in Terms of Other Markers of Outcomes.a
Abbreviations: AKI, acute kidney injury; CRRT, continuous renal replacement therapy; PC4, Pediatric Cardiac Critical Care Consortium; VNCH, Vietnam National Children′s Hospital.
Comparative data includes patients over 18 years of age and those with unassigned RACHS-1/STAT score (62 patients at VNCH).
Results
Vietnam National Children’s Hospital Resources
Resources and infrastructure for surgical centers in LMICs are nicely presented by survey methodology of 37 centers across 17 countries. 8 A recent CHSS article provides recommendations for comprehensive and essential care centers in the US. 9 Comparisons between IQIC centers, CHSS recommendations, and VNCH are displayed in Table 1. VNCH surgical volume qualifies for a comprehensive center but lacks recommended resources. VNCH compares equivalently with IQIC centers.
Collaboration
Since the establishment of the collaboration between VNCH and UCSF through CHL, from Fall 2017 through the end of 2023, there were eight in-person team visits during which approximately 500 VNCH clinicians participated. In-person training was aborted in 2020 due to the COVID pandemic, which was when remote virtual case discussions were initiated. Through the end of 2023, there were 109 remote case discussion sessions, during which a total of 198 cases were discussed, with approximately 20 VNCH clinician participants in each session and the lead VNCH surgeon was always present. There were five nursing virtual sessions through the end of 2023, with approximately 62 VNCH nursing participants. In-person training visits were reinitiated in the Fall of 2022. Additionally, throughout the years of collaboration between VNCH and UCSF through CHL, approximately a total of 10 surgeons and cardiologists from VNCH have been sponsored to attend 2-to-6 months of training at either UCSF or CHL partner hospitals and more than 100 clinicians have participated in virtual training courses and conferences.
Case Mix
VNCH had a surgical volume of 1,163 index operations for calendar year 2023, 1,056 of which were eligible for RACHS-1 assignment and less than 18 years of age (Table 2). Compared with PC4 aggregate congenital heart centers, VNCH performed slightly fewer surgeries on neonates (15.9% [168/1056] vs 19.5% [2832/14,506], p = .004) and more surgeries on infants (50.5% [133/1056] vs 35.7% [5179/14,506], p < .001). That VNCH performed fewer surgeries on patients with genetic syndrome and/or chromosomal abnormalities (4.7% [53/1056] vs 28% [4,056/14,506], p < .001) may reflect that testing is rare. VNCH performed fewer surgeries on patients with major noncardiac anomalies (2.9% [31/1056] vs 22.7% [3296/14,506], < .001) and fewer complex operations, notably RACHS-1 4/STAT 4 (6.4% [72/1056] vs 25.3% [3511/14,506], p < .001) and RACHS-1 5 and 6/STAT 5 (0.9% vs 4%, < .001).
Outcomes
Overall, in-hospital mortality was similar between VNCH and PC4 aggregate hospitals in 2023 (Table 3). There was no difference in mortality for STAT/RACHS-1 cases or neonatal surgeries. However, the mortality was higher at VNCH on surgeries performed on those with prematurity (20% [3/15] vs 11% [37/336], p < .001) and major noncardiac anomaly (12.9% [4/31] vs [139/3296] 4.2%, p < .001). VNCH had a higher rate for SSIs (4.8% [51/1056] vs 2.3% [334/14,506], p < .001), but a lower rate for need for reoperation due to bleeding (1.3% [14/1056] vs 7.8% [1131/14,506], p < .001), diaphragm paresis (0.5% [5/1056] vs 5.8% [841/14,506], p < .001), or acute kidney injury (AKI) requiring continuous renal replacement therapy (CRRT) (0.2% [2/1056] vs 4.9% [710/14,506], p < .001) (Table 4).
IQIC risk adjusted in-hospital mortality from 2017 to 2023 is shown in Figure 1. Overall, the risk adjusted mortality rate at VNCH has decreased over the course of the collaboration, with a slight increase during the COVID pandemic.
Vietnam National Children’s Hospital (VNCH) risk-adjusted hospital mortality over time.
Discussion
Supporting LMIC Centers
Highly resourced cardiac centers have supported LMIC centers in a variety of approaches from sponsored visits to perform surgery, telehealth support, and sponsored visits to training centers. Benchmarking outcomes is imperative for any cardiac center. Unfortunately, most benchmarking registries based in highly resourced countries are cost prohibitive for centers in LMICs. IQIC is an important cardiac surgical registry collaborative which aims to support benchmarking among pediatric cardiac surgical centers in LMICs with improvement in outcomes. 14
Comparative Resources
Recommendations exist for pediatric congenital heart programs in LMICs. 15 We took this one step further by providing comparative resources between VNCH, IQIC centers, 8 and recommended US essential and comprehensive centers based on CHSS recommendations. 9 VNCH qualifies as a US comprehensive center based upon volume but lacks many CHSS-recommended resources including neonatal extracorporeal membrane oxygenation, ventricular assist devices, transplant, sufficient iNO, CICU dedicated support (respiratory therapy, pharmacy, social work), and hospital and heart center wide resources (rapid response team, child life, pulmonary hypertension team, maternal fetal medicine program, outreach programs, etc). Many of these latter resources likely impact patient/family experience, short- and longer-term morbidities, nutritional status, and access to care which may not be reflected in the comparative outcomes in this article.
Collaboration
The primary aim of the collaboration between UCSF and VNCH is to develop VNCH into a high qualiy cardiac program in a LMIC with outcomes similar to North American and European centers. As such, our initial visits focused on practices to improve neonatal surgery, echocardiography and catheterization (interventional and diagnostic), surgical case planning, engaging nursing, and reducing nosocomial infections. More recent visits have been dedicated to a particular diagnosis and/or procedure (eg, pulmonary atresia with ventricular septal defect and multiple aortopulmonary collaterals, severe Ebstein anomaly, corrected transposition of the great arteries and the double switch procedure, and right ventricular overhaul procedure). Patients are discussed virtually and electively scheduled for additional diagnostic evaluation and surgery at the time of the in-person visit. Over the years, teaching has transitioned from hands-on to hands-off teaching in the operating room, catheterization laboratory, and outpatient clinics. In-person visits are wrapped up with a meeting with hospital leadership outlining accomplishments and areas for improvement. Outside of biannual in-person visits, collaboration continues through weekly virtual case conference discussions. Throughout the years of partnership, hospital mortality improved. This concurs with a recent study that showed a 5-fold decrease in in-hospital mortality after atrial septal defect repair in 71 IQIC sites in LMICs over a 10-year period. 14 Another heart center in Vietnam has also benefitted from this model of on-the-ground training for the local team through partnership with a highly resourced center with formalized training. 16 However, we took a further step to benchmark VNCH outcomes reported to IQIC with the North American PC4 outcomes to further address quality improvement efforts.
Neonates and Young Infants
It is the impression of the UCSF team that neonates and young infants presenting for heart surgery to VNCH, compared with those presenting to UCSF, are more likely to have a late diagnosis, to present with intercurrent infection (commonly pneumonia), to be referred from a remote location, and to be underweight. In addition, iNO is often not available, and ECLS is not available for neonates. Therefore, the comparable neonatal survival rates with the PC4 collaborative are remarkable.
Preterm Infants
Compared with PC4 aggregate hospitals, VNCH operated on a similar percentage of premature infants but struggled with a relative increase of in-hospital mortality. In addition to the typical complications seen in term infants, preterm neonates face additional challenges associated with prematurity including intraventricular hemorrhage, necrotizing enterocolitis, anemia, hyperbilirubinemia, and bronchopulmonary dysplasia, all of which may increase hospital mortality.17–19 Increased mortality is associated with lower gestational age and weight at the time of surgery. 20 Therefore, for centers in LMICs with limited resources, it is not surprising that a specific patient population with high perioperative resource needs would have lower survival.
Noncardiac and Chromosomal Abnormalities
Children with CHD often have noncardiac and chromosomal abnormalities. 21 The lower percentage of patients with congenital anomalies who undergo surgery at VNCH may reflect parental preference. The lower percentage of patients with noncardiac and chromosomal abnormalities who undergo surgery at VNCH primarily reflects a lower incidence of testing due to financial constraints. Parents at VNCH have the right to request end of treatment. This decision is multifaceted but includes financial liability, concern for pain associated with medical interventions, and anxiety over the burden of developmental delay. It is the impression of the UCSF team that families of patients at VNCH move to an earlier withdrawal of care, including patients who may likely survive.
Higher Risk Surgeries
Compared with PC4 aggregate hospitals, VNCH performs a lower relative percentage of RACHS-1 1-4 to 6/STAT 4 and 5 surgeries. This is primarily due to referral patterns. Patients arriving to VNCH needing STAT 4 and 5 procedures receive surgery, including those presenting with hemodynamic compromise. However, patients presenting from remote villages or with long travel often opt not to pursue care. Outcomes for STAT 4 and 5 patients at VNCH are comparable with the PC4 aggregate. This is remarkable, as a majority of STAT 5 patients are those who undergo the Norwood operation and homograft or polytetrafluoroethylene (PTFE) graft and these procedures are not available at VNCH. Instead, a surgical technique that involves aorta reconstruction through autologous pulmonary artery trunk anastomoses with the ascending aorta in an end-to-side fashion is employed. Polytetrafluoroethylene grafts without a ring are available and they are sutured directly to the ventriculotomy with a PTFE suture of 5.0.
Morbidities
Although not risk adjusted, VNCH had a higher rate of SSIs compared with PC4 aggregate hospitals in 2023, at 4.8% (51/1056) compared with 2.3% (334/14,506) and compared with the IQIC aggregate at 2.2%. 22 VNCH′s SSI rate demonstrated a decrease from 2017 to 2022 from 4.1% to 2.1% with a marked increase in 2023. 22 This may reflect an effort to focus on higher risk surgeries. Improvement in SSI at a CHL and IQIC supported site has been previously reported. 23 Surgical site infection reduction training at VNCH has focused on preoperative skin preparation, operative sternal closure, postoperative wound care, and differentiating infection from wound complications such as dehiscence. Going forward, training will focus on reduction in SSIs and higher complexity surgeries, particularly in patients with prematurity and major noncardiac anomalies.
Limitations
Our study was limited by its retrospective nature and different definitions across IQIC and PC4, with the most notable examples being surgical complexity and prematurity. Although IQIC and PC4 both provide risk adjusted benchmarked outcomes, as VNCH does not participate in PC4, comparisons cannot be risk adjusted. Comparisons between RACHS-1 and STAT categories are not directly equivalent. Comparisons on other markers of outcomes (SSI, reoperation for bleeding, diaphragm paresis, and AKI requiring CRRT) included patients without an assigned RACHS-1/STAT score (n = 62/1056, 5.9%) at VNCH; hence, introducing some potential but presumptively small bias. In addition, this is a single center comparison with PC4 aggregate hospitals; hence, data from VNCH are not generalizable to other countries in the developing world.
Footnotes
Abbreviations
Acknowledgment
Children's HeartLink team and sponsors.
Authors’ Note
Data Availability: All authors agree to participate in data sharing.
Author Contributions
Chau P. Nguyen contributed to methodology; investigation; data curation; conceptualization; writing—original draft, and writing—review and editing. Huu Nhat Nguyen contributed to methodology; investigation; and writing—review and editing. Sarah Tabbutt contributed to methodology; investigation; data curation; conceptualization; writing—review and editing; and supervision. Shunji Sano contributed to methodology; investigation; and writing—review and editing. Mark Cocalis contributed to methodology; investigation; and writing—review and editing. Hung Nguyen contributed to methodology; investigation; and writing—review and editing. Helen Busch contributed to writing—review and editing. Adriana Dobrzycka contributed to methodology; investigation; and writing—review and editing. Kathy J. Jenkins contributed to conceptualization; methodology; and writing—review and editing. Kimberlee Gauvreau contributed to methodology; investigation; and writing—review and editing. Hanh Luu Nguyen Duc contributed to writing—review and editing. Truong N. Ly Thinh contributed to conceptualization; writing—review and editing; and supervision. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Ethical Considerations
This is an IRB-approved retrospective study, all patient information was de-identified and patient consent was not required. Patient data will not be shared with third parties. Data sharing is covered by site-specific data use agreements, center level, and University of Michigan and Boston Children's Hospital IRB approvals.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for all hospital partnership activities was provided by Chao Family Foundation and Transparent Fish Fund.
