Secondary Logo

Journal Logo

Pediatric Circulatory Support

Discharge and Readmissions After Ventricular Assist Device Placement in the US Pediatric Hospitals: A Collaboration in ACTION

Bearl, David W.*; Feingold, Brian; Lorts, Angela; Rosenthal, David§; Zafar, Farhan; Conway, Jennifer; Elias, Barbara; Tunuguntla, Hari; Thurm, Cary#; Amdani, Shahnawaz**; Jaworski, Nancy*; Godown, Justin*

Author Information
doi: 10.1097/MAT.0000000000001307

Abstract

Ventricular assist device (VAD) use and postimplant outpatient management has lagged in pediatric patients compared with adults, largely because of the need to adapt adult technology for children, different and more heterogenous indications and fewer overall patients. As these hurdles are overcome, use and practice in pediatric end-stage heart disease has been steadily increasing and improving over the last two decades.1 The first large experience in pediatric VAD use was with the Berlin Heart EXCOR (Berlin Heart, Berlin, Germany).2 Although this represented a significant step forward in advanced heart failure management, the pulsatile, extracorporeal Berlin Heart EXCOR VAD has limited portability and thus necessitates continuous hospitalization while on the device. The evolving off-label use of adult continuous-flow devices in children, including HeartWare HVAD (Medtronic, Minneapolis, MN), HeartMate II and HeartMate 3 (Abbott Laboratories, Abbott Park, IL), has introduced the possibility of discharge while on VAD support in the pediatric population. The known advantages of discharge on VAD support are substantial and include cost savings,3,4 the ability to return to school,5 or leisure activities such as snowboarding and bicycling.6 Additionally, extrapolation from adult data shows further potential advantages such as the ability to participate in cardiac rehabilitation,7 return to pre heart failure activities,8 and improved quality of life.9 Therefore, there is a concerted effort to further tailor pediatric practice by increasing discharges of children on VAD support. Optimizing this process and balancing the risk of unplanned readmissions while facilitating timely discharge represents a quality improvement target of Advanced Cardiac Therapies Improving Outcomes Network (ACTION) (https://www.actionlearningnetwork.org/). This project aimed to describe discharge and readmission frequencies in children on VAD support from ACTION sites using a large multicenter administrative database.

Methods

This study used data from the Pediatric Health Information System (PHIS) administrative and billing database (Children’s Hospital Association, Lenexa, KS). The PHIS database collects clinical and daily resource utilization data for hospital encounters from >50 tertiary children’s hospitals. This includes data from inpatient hospitalizations, observation, ambulatory surgery, and emergency department visits. This database records diagnosis and procedural International Classification of Diseases (ICD)-9 and ICD-10 codes, payer information, along with encounter-level detailed hospital billing data.

All patients who underwent VAD implantation at an ACTION center between 2009 and 2018 were identified from the PHIS database for inclusion. Patients were identified by the presence of ICD-9 or ICD-10 procedure codes for insertion of an implantable heart assist system (Appendix Table S1, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A566), but did not differentiate from left, right, or biventricular support. Because data pertaining to specific VAD types are not available within the PHIS database, the cohort was limited to patients 10–21 years of age to avoid inclusion of younger pediatric patients who are typically supported with extracorporeal devices that would not be eligible for discharge and to include the upper limit of patients cared for in children’s hospitals. Patients were excluded if they died or underwent VAD explant without transplant before hospital discharge. However, patients with codes indicating device explant in conjunction with a second VAD implant procedure were included as this was taken to represent device exchange or revision. Patients who remained in the hospital on VAD support for more than 30 days after VAD placement were included as a comparison cohort of “potentially dischargeable” patients.

The frequency of discharge on VAD support was described across centers and over time. Characteristics of those discharged and “potentially dischargeable” were compared using the χ2 test and Wilcoxon rank sum test, as appropriate. Freedom from first readmission for patients who were discharged on VAD support was assessed using the Kaplan–Meier method. Patients were censored at the time of readmission if heart transplantation occurred within 2 days of the admission date as this was assumed to represent readmission for the purpose of heart transplantation. Patients were also censored if no PHIS encounters occurred after the date of initial discharge (i.e., no follow-up data were available). For each hospital encounter, PHIS captures up to 41 unique ICD codes with the first diagnosis representing the most important in relation to the admission encounter. To determine the etiology for readmission, the primary diagnosis for each readmission was categorized as infection, heart failure, hematologic, neurologic, arrhythmia, fluids/electrolytes/nutrition, or other (Appendix Table S2, Supplemental Digital Content 1, https://links.lww.com/ASAIO/A566). Readmissions were also categorized based on the need for intensive care unit (ICU) level care, extracorporeal membrane oxygenation (ECMO) support, mechanical ventilation, cardiac reoperation including VAD-related surgeries, cardiac catheterization, documented infection, and in-hospital mortality.

All statistical analyses were performed using STATA version 15 (StataCorp LLC, College Station, TX) with two-sided P < 0.05 considered statistically significant. This project was approved by the Vanderbilt University Institutional Review Board and PHIS.

Results

There were 766 patients identified across 25 different ACTION centers with ICD procedure codes indicating VAD placement. A total of 468 patients were excluded because of age (<10 or >21 years old) or death or device explant before hospital discharge (Figure 1). Of the remaining 298 patients, 163 (54.7%) were discharged on VAD support. Of the 135 patients not discharged, 71 (52.6%) underwent early transplant (≤30 days between VAD placement and transplantation) and 64 (47.4%) were identified as “potentially dischargeable” (>30 days on VAD support).

F1
Figure 1.:
Flowsheet of included and excluded patients.

Twenty-two of the 25 included ACTION centers (88%) were identified as discharging at least one patient on VAD support. Among centers with patients discharged on VAD support, the rate of discharge ranged from 8% to 100% with a median of 52% (Figure 2). The frequency of discharge increased over time (36.9% [2009–2012] vs. 59.7% [2013–2018], p = 0.001) (Figure 3).

F2
Figure 2.:
Discharge rates across ACTION centers based on VAD volume. ACTION, Advanced Cardiac Therapies Improving Outcomes Network; VAD, ventricular assist device.
F3
Figure 3.:
Rate of discharge on VAD support over time. VAD, ventricular assist device.

Demographics of discharged and potentially dischargeable patients are shown in Table 1. Patients discharged on VAD support were more likely to be older (median age 14 vs. 13 years, p = 0.003), have commercial insurance (60% vs. 39.7%, p = 0.008), and have a higher median household income ($44,196 vs. $39,194, p = 0.033). There were 20 patients 18 years old or greater (i.e., could only be listed under adult UNOS criteria), and only 14 (70%) were discharged. The total length of hospitalization was greater for patients who were not discharged on VAD support (92 vs. 46 days, p < 0.001).

Table 1. - Demographics of Patients on VAD Support >30 Days Based on Discharge
Total (n = 227) Discharged p
Yes (n = 163) No (n = 64)
Age (years) 14 (12–16) 14 (12–16) 13 (11–15) 0.003
Race
 Caucasian 97 (42.7) 72 (44.2) 25 (39.1) 0.297
 African American 63 (27.8) 39 (23.9) 24 (37.5)
 Hispanic 45 (19.8) 35 (21.5) 10 (15.6)
 Asian 6 (2.6) 4 (2.5) 2 (3.1)
 Other 16 (7) 13 (8) 3 (4.7)
Male sex 152 (67) 109 (66.9) 42 (67.2) 0.964
Insurance
 Commercial 121 (54.3) 96 (60) 25 (39.7) 0.008
 Medicaid/government 82 (36.8) 48 (30) 34 (54)
 CHIP 14 (6.3) 12 (7.5) 2 (3.2)
 Other 6 (2.7) 4 (2.5) 2 (3.2)
Patient distance from transplant center (miles) 42 (17–148) 44 (17–136) 39 (17–166) 0.735
Total length of stay (days) 59 (37–82) 46 (34–71) 92 (65–126) <0.001
Median household Income $42,259 ($31,649–$61,908) $44,196 ($33,761–$63,551) $39,194 ($29,856–$50,369) 0.033
*Data are presented as n (%) for categorical variables and median (interquartile range) for continuous variables.
A p value calculated by χ2 test for categorical variables or Wilcoxon rank sum test for continuous variables comparing each characteristic between discharge and dischargeable patients.
CHIP, Children’s Health Insurance Program; VAD, ventricular assist device.

Among 163 patients discharged on VAD support, 19 (11.7%) had no subsequent follow-up encounters in the PHIS database and were excluded from the readmission analysis. The entire follow-up time was 5,358 days for the remainder of the discharged patients. Ninety-six (66.7%) patients were readmitted for reasons other than transplantation, accounting for 283 total readmission encounters (Figure 4). The median time to first readmission was 45 days (interquartile range [IQR] 18–79 days) with a median of 2 readmissions per patient (IQR 1–4). The 30-day all-cause readmission rate was 34.6%. Overall, there was a rate of 0.54 admissions per patient-month. The etiology for readmission was unable to be readily identified in 27 (9.5%) encounters using the primary ICD diagnosis code. For the remaining 256 encounters, the most common etiologies for readmission were heart failure (71 encounters in 42 patients, 27.7% of readmissions), infection (66 encounters in 40 patients, 25.8%), and hematologic concerns (43 encounters in 29 patients, 16.8%) (Figure 5). Overall events occurring during readmission encounters are shown in Table 2. Intensive care unit care was provided in 33 (46.5%) heart failure readmissions, 32 (48.5%) infection readmissions, and 15 (34.9%) hematologic readmissions. Both heart failure and infection readmissions had higher mechanical ventilation (23.9% and 16.7%, respectively) and ECMO need (2.8% and 4.5%) compared with hematologic readmissions (4.7% mechanical ventilation and 0% ECMO). There were two in-hospital mortalities with heart failure and infection encounters, whereas none associated with a hematologic encounter. Median readmission length of stay for heart failure, infection, and hematologic was 4 days (IQR 1–19), 8 days (3–18), and 5 (2–11), respectively. There were 44 (15.5%) readmissions that involved a cardiac-related operative procedure during the hospitalization, although 23 (52.3%) of those were for heart transplantation >2 days after readmission. Overall mortality was uncommon with 5 out of 283 (1.8%) readmission encounters resulting in a patient death.

Table 2. - Events During Readmission Encounters
Readmissions (n = 283)
Required ICU care 126 (44.5%)
Required ECMO support 6 (2.1%)
In-hospital mortality 5 (1.8%)
Required mechanical ventilation 52 (18.4%)
Cardiac reoperative procedure 44 (15.5%)
VAD repair or replacement 7 (2.5%)
Explant 3 (1.1%)
Heart transplant >2 days after admission 23 (8.1%)
Other cardiac operation 11 (3.9%)
Infection documented 105 (37.1%)
Underwent cardiac catheterization 40 (14.1%)
ECMO, extracorporeal membrane oxygenation; ICU, intensive care unit; VAD, ventricular assist device.

F4
Figure 4.:
Freedom from readmission for patients discharged on VAD support. VAD, ventricular assist device.
F5
Figure 5.:
Etiology of readmission encounters based on primary ICD diagnosis codes. CHF, congestive heart failure; FEN, fluids, electrolytes and nutrition; Neuro, neurologic.

Discussion

We found that just more than half of patients ages 10–21 years implanted with VADs across 25 pediatric hospitals who were discharge eligible (i.e., did not undergo heart transplantation within 30 days of VAD placement) were discharged on VAD support. Initial discharge experiences with HVAD in select children showed promise.5,10 Evaluation of more recent discharge trends from 2012 to 2015 in the Pediatric Interagency Registry for Mechanical Circulatory Support (Pedimacs) database showed that 45% of eligible patients were discharged.11 Our analysis showed an increasing discharge rate over the last decade. An important factor in the pediatric VAD discharge rate in the United States is the current heart transplant waitlist criteria whereby all patients <18 years of age on VAD support (regardless of device type or hospitalization status) qualify for priority listing (1A) in the United States. Furthermore, the eligibility for 1A status does not have a time limit. Therefore, the discharge rates are likely skewed because pediatric centers may not delay listing, as suggested by the nearly 24% of eligible patients in our study who underwent heart transplant within 30 days of VAD implantation. This is in stark contrast to the current adult practice that is largely defined by a significant portion of VAD implants as destination therapy and the Organ Procurement and Transplantation Network adult heart status assignments where patients on dischargeable VAD do not qualify for priority status (currently status 4, with up to 30 days at status 3). In Europe, where organ scarcity is worse, they had a higher discharge rate of 72% in intracorporeal VADs.12 In adults in the United States, 89% were discharged in 2014 (with all nondischarges comprised mortalities).13 According to the most recent Intermacs data, discharge rates post-VAD were not specifically included, but 30 day survival was 95%,14 and the median length of stay was 20 days.15 Interestingly, the discharge rate among adults from pediatric centers in our study was still only 70%. Although a discussion about pretransplant VAD management in children is beyond the scope of this article, it is notable that recent data shows better posttransplant outcomes for patients transplanted >2 months after VAD implant compared with those transplanted earlier.16 Additionally, the single-center experience at Texas Children’s Hospital, where it is standard practice to delay listing for 3 months post-VAD implantation for uncomplicated patients on dischargeable devices, shows excellent waitlist and posttransplant outcomes along with a 95% discharge rate.17

Among patients discharged in our study, approximately 2/3 were readmitted at least once. This is in line with the 61% of patients readmitted in the Pedimacs cohort,11 as well as several other studies that have shown even higher readmission numbers of 83% and 84%, respectively.5,18 A more representative event rate per patient-month, however, shows possible discrepancies between the study groups. Schweiger et al.5 demonstrated a very low readmission rate of 0.02 per patient-month, whereas Hollander et al.18 was higher at 0.3 per patient-month. The larger Pedimacs cohort and adult Intermacs cohort were in the middle at 0.15 and 0.18 per patient-month, respectively.11 Although pediatric studies thus far have not reported all-cause 30 day readmissions, adult 30 day readmission rates vary between 10 and 28%.13,19 Our data showed the readmission rates on the higher side at almost 35% 30 day readmissions and 0.54 per patient-month. This could reflect the relative comfort of getting patients to discharge (i.e., earlier studies may have been more selective in who was discharged), but an otherwise conservative approach to readmission. Alternatively, pediatric centers may be discharging higher risk patients compared with earlier studies. Last, it could reflect the small numbers and follow-up time for the prior studies, and overall readmission rates are not significantly different between them.

More than 70% of the readmissions in our study were classified in three categories: heart failure, infection, and hematologic complications. Hematologic complications included both bleeding events and thromboembolic events, as well subtherapeutic or supratherapeutic anticoagulation, but excluded neurologic or intracranial events. Hematologic events in 17% of children in our study falls within the observed rates of hematologic-related readmissions or late hematologic complications that ranged from 17 to 30%.5,18 Interestingly, adult data show similar overall hematologic-related readmissions; however, the majority of are due to gastrointestinal bleeding,20,21 whereas we only noted two readmissions related to gastrointestinal bleeding. The rate of infection-related readmissions was similarly comparable to previously published data.11,22 Although the infectious and bleeding complications were common in other studies, heart failure, and especially the prevalence of, was unique to this study. Out of 46 readmissions, Hollander et al.18 showed only four were caused by heart failure exacerbation (9%) and the earlier study by Schweiger et al.5 had none of the 30 readmissions. Conflicting adult data also exist on heart failure readmission after VAD with ranges reported from 11 to 33%.20,21 Although the right heart failure specifically has been identified as a risk factor readmission in adults,23,24 there are no studies that specifically address this problem in pediatrics. Among the discharged cohort in our study, there were no patients who required late right-sided VAD placement.

Despite the significant number of readmissions while on VAD support, we saw low overall mortality (1.8%) for readmissions. This is likely consistent with other pediatric reports; however, interpretation may be difficult in cases of destination therapy. Because of this confounder, comparison with adult studies, where destination therapy is significantly more common, is difficult. However, the increasing number of readmissions for adults with a VAD within the first 90 days postimplant was strongly associated with worse long-term outcomes.25 The acuity of readmissions remained high in our study including the need for ICU level care, mechanical ventilation, and cardiac operations. Of the 44 operations, seven were related to VAD repair, whereas three were VAD replacement presumably for device thrombosis. This rate of 0.039 per patient-month of VAD-related surgery was similar to the Stanford experience of 0.042 per patient-month.18

This study of pediatric VAD discharges and readmissions has limitations. Only 25 pediatric centers affiliated with both ACTION and PHIS (all within the United States) were represented, which may not represent practice across the entire pediatric VAD community. The age range chosen was to avoid patients with paracorporeal devices who would not be eligible for discharge because of the device type but may have also excluded younger discharged or dischargeable patients with intracorporeal devices. Furthermore, it is not possible to distinguish whether the differences between centers regarding discharge and readmission are a function of a patient population or institutional-specific practices. Some patients had no follow-up data because PHIS does not track outpatient events or encounters, and readmissions at non-PHIS hospitals would also not be captured. Defining “potentially dischargeable” patients as those who remain hospitalized >30 days post-VAD implantation is arbitrary but is an easily identifiable cutoff. A major limitation to the data, and reason for age cutoffs, was the PHIS data lacks granularity regarding device type. It is therefore unknown through this data if patients received a paracorporeal or intracorporeal VAD, which brand of VAD, or which side (left, right, or biventricular support). Another limitation of the study is the use of ICD codes to identify etiologies for readmission. In many instances, readmission diagnoses could be grouped into a large category such as heart failure, but the data source limits further classification. Although most readmissions had a clear primary diagnosis, there were a small minority (9%) that could not be categorized further limiting the sample size.

Despite the limitations, this study provides foundational data from which improvement in VAD management, including postoperative care, discharge planning, and outpatient services can lead to better outcomes for children with the end-stage heart disease. Focusing on ACTION sites allowed for convenient sampling to establish a baseline for discharge rates and readmissions, such that collaboration in ACTION can be used to address many of the data granularity issues found in this study and begin to apply quality improvement methodology similarly to other VAD-related challenges it has focused on.26 As part of improving care for this group, a recent survey of ACTION centers showed that most already have a standardized approach to discharge preparation and teaching.27 Overall, this analysis demonstrates the current performance of a group of active pediatric VAD centers in the United States. The next steps to improve pediatric VAD outcomes include optimizing inpatient post-VAD care to facilitate safe discharge and shorten length of stay, research into timely and effective outpatient care that minimizes readmissions, and confirm that improvements in VAD management ultimately lead to better waitlist and posttransplant outcomes.

Conclusion

Pediatric patients discharged with a VAD from ACTION centers has increased over time, with high variability across centers, however, still lag behind adult and European rates. Readmissions are common with diverse indications that present both similar and unique challenges compared with adults; however, the risk of mortality during a readmission encounter is low. We have established a new baseline by which interventions to increase discharge rates, shorten length of stay, and optimize outpatient management, can be evaluated for this vulnerable population.

References

1. Dipchand AI, Kirk R, Naftel DC, et al.; Pediatric Heart Transplant Study Investigators. Ventricular assist device support as a bridge to transplantation in pediatric patients. J Am Coll Cardiol. 72: 402–415, 2018.
2. Almond CS, Morales DL, Blackstone EH, et al. Berlin Heart EXCOR pediatric ventricular assist device for bridge to heart transplantation in US children. Circulation. 127: 1702–1711, 2013.
3. Rossano JW, Cantor RS, Dai D, et al. Resource utilization in pediatric patients supported with ventricular assist devices in the United States: A multicenter study from the Pediatric Interagency Registry for Mechanically Assisted Circulatory Support and the Pediatric Health Information System. J Am Heart Assoc. 7: e008380, 2018.
4. Godown J, Smith AH, Thurm C, et al. Mechanical circulatory support costs in children bridged to heart transplantation - Analysis of a linked database. Am Heart J. 201: 77–85, 2018.
5. Schweiger M, Vanderpluym C, Jeewa A, et al. Outpatient management of intra-corporeal left ventricular assist device system in children: A multi-center experience. Am J Transplant. 15: 453–460, 2015.
6. Helman DN, Addonizio LJ, Morales DL, et al. Implantable left ventricular assist devices can successfully bridge adolescent patients to transplant. J Heart Lung Transplant. 19: 121–126, 2000.
7. Grosman-Rimon L, Lalonde SD, Sieh N, et al. Exercise rehabilitation in ventricular assist device recipients: A meta-analysis of effects on physiological and clinical outcomes. Heart Fail Rev. 24: 55–67, 2019.
8. Morales DL, Argenziano M, Oz MC. Outpatient left ventricular assist device support: A safe and economical therapeutic option for heart failure. Prog Cardiovasc Dis. 43: 55–66, 2000.
9. Grady KL, Meyer PM, Mattea A, et al. Change in quality of life from before to after discharge following left ventricular assist device implantation. J Heart Lung Transplant. 22: 322–333, 2003.
10. Miera O, Kirk R, Buchholz H, et al. A multicenter study of the HeartWare ventricular assist device in small children. J Heart Lung Transplant. 35: 679–681, 2016.
11. Rossano JW, Lorts A, VanderPluym CJ, et al. Outcomes of pediatric patients supported with continuous-flow ventricular assist devices: A report from the Pediatric Interagency Registry for Mechanical Circulatory Support (PediMACS). J Heart Lung Transplant. 35: 585–590, 2016.
12. Schweiger M, Miera O, de By TMMH, et al.; EUROMACS members. Cerebral strokes in children on intracorporeal ventricular assist devices: Analysis of the EUROMACS Registry. Eur J Cardiothorac Surg. 53: 416–421, 2018.
13. Setareh-Shenas S, Thomas F, Cole RM, et al. Evaluation of 30 day readmissions after index ventricular assist device implantation in the United States. ASAIO J. 65: 601–604, 2019.
14. Teuteberg JJ, Cleveland JC Jr, Cowger J, et al. The Society of Thoracic Surgeons Intermacs 2019 annual report: The changing landscape of devices and indications. Ann Thorac Surg. 109: 649–660, 2020.
15. Cotts WG, McGee EC Jr, Myers SL, et al. Predictors of hospital length of stay after implantation of a left ventricular assist device: An analysis of the INTERMACS registry. J Heart Lung Transplant. 33: 682–688, 2014.
16. Riggs KW, Zafar F, Lorts A, Villa CR, Bryant R 3rd, Morales DLS. Optimizing postcardiac transplantation outcomes in children with ventricular assist devices: How long should the bridge be? ASAIO J. 66: 787–795, 2020
17. Adachi I, Zea-Vera R, Tunuguntla H, et al. Centrifugal-flow ventricular assist device support in children: A single-center experience. J Thorac Cardiovasc Surg. 157: 1609.e2–1617.e2, 2019.
18. Hollander SA, Chen S, Murray JM, et al. Rehospitalization patterns in pediatric outpatients with continuous-flow VADs. ASAIO J. 63: 476–481, 2017.
19. Kormos RL, Cowger J, Pagani FD, et al. The Society of Thoracic Surgeons Intermacs database annual report: Evolving indications, outcomes, and scientific partnerships. J Heart Lung Transplant. 38: 114–126, 2019.
20. Tsiouris A, Paone G, Nemeh HW, Brewer RJ, Morgan JA. Factors determining post-operative readmissions after left ventricular assist device implantation. J Heart Lung Transplant. 33: 1041–1047, 2014.
21. Forest SJ, Bello R, Friedmann P, et al. Readmissions after ventricular assist device: Etiologies, patterns, and days out of hospital. Ann Thorac Surg. 95: 1276–1281, 2013.
22. Agrawal S, Garg L, Shah M, et al. Thirty-day readmissions after left ventricular assist device implantation in the United States: Insights from the Nationwide Readmissions Database. Circ Heart Fail. 11: e004628, 2018.
23. Rich JD, Gosev I, Patel CB, et al.; Evolving Mechanical Support Research Group (EMERG) Investigators. The incidence, risk factors, and outcomes associated with late right-sided heart failure in patients supported with an axial-flow left ventricular assist device. J Heart Lung Transplant. 36: 50–58, 2017.
24. Takeda K, Takayama H, Colombo PC, et al. Incidence and clinical significance of late right heart failure during continuous-flow left ventricular assist device support. J Heart Lung Transplant. 34: 1024–1032, 2015.
25. Vivo RP, Krim SR, Estep JD, et al. How do readmissions impact survival among patients with continuous-flow left ventricular assist devices? Findings from INTERMACS. J Heart Lung Transplant. 33: S87, 2014.
26. Peng DM, Rosenthal DN, Zafar F, Smyth L, VanderPluym CJ, Lorts A. Collaboration and new data in ACTION: A learning health care system to improve pediatric heart failure and ventricular assist device outcomes. Transl Pediatr. 8: 349–355, 2019.
27. Elias B, Tunuguntla HP, Smyth L, et al. Preparing for discharge in pedaitric ventricular assist device supported patients. J Heart Lung Transplant. 39: S220, 2020.
Keywords:

ventricular assist device; pediatrics; discharge; readmissions

Supplemental Digital Content

Copyright © ASAIO 2020