Abstract
Objective:
Prolonged critical illness (PCI) after congenital heart surgery (CHS) disproportionately harms patients and the healthcare system, yet much remains unknown. We aimed to define PCI, delineate between non-modifiable and potentially preventable predictors of PCI and PCI-mortality, and understand the inter-hospital variation in PCI.
Design:
Observational analysis
Setting:
Pediatric Cardiac Critical Care Consortium (PC4) clinical registry.
Patients:
All patients, stratified into neonates (≤28 days) and non-neonates (29 days-18 years), admitted to the pediatric cardiac ICU after CHS at PC4 hospitals.
Interventions:
None.
Measurements and Main Results:
There were 2,419 neonates and 10,687 non-neonates from 22 hospitals. The PCI-cutoff (90th percentile length of stay) was ≥35 and ≥10 days for neonates and non-neonates, respectively. Cardiac ICU PCI-mortality was 24% in neonates and 8% in non-neonates (vs. 5% and 0.4%, respectively, in non-PCI patients). Multivariable logistic regression identified 10 neonatal and 19 non-neonatal PCI predictors within strata, and 8 predictors of mortality. Only mechanical ventilation days and acute renal failure requiring renal replacement therapy predicted PCI and PCI-mortality in both strata. Approximately 40% of the PCI predictors were non-modifiable (pre-operative/patient and operative factors), while only 1/8 PCI-mortality predictors was non-modifiable. The remainder were potentially preventable (post-operative critical care delivery variables and complications). Case-mix adjusted PCI rates were compared across hospitals; 6 hospitals each had lower- and higher-than-expected PCI-incidence.
Conclusions:
While many PCI predictors are non-modifiable, we identified several predictors to target for improvement. Moreover, we observed that complications and prolonged critical care therapy drive PCI-mortality. Wide variation of PCI-incidence suggests that identifying practices at hospitals with lower-than-expected PCI could lead to broader quality improvement initiatives.
MeSH Keywords: congenital cardiac surgery, pediatric critical illness, length of stay, newborn, intubation, reoperation
Introduction
Patients who experience prolonged critical illness (PCI) after congenital heart surgery (CHS) present important challenges in cardiac intensive care units (CICU) [1–4]. Prior literature from pediatric intensive care units (PICU) demonstrates PCI patients account for a small percentage of admissions (5%), but disproportionately high bed-days (36%) [5]. These investigators also observed wide variation in incidence (2–8%) and prevalence (15–58%) of PCI among hospitals, and that PCI patients had higher morbidity, mortality, and hospital cost. Further, PCI significantly impacts parents’ stress [6].
Existing studies examining PCI after CHS are few and limited. Single-center [1, 4, 7] and single lesion studies [3] are not generalizable. The definition of PCI varies widely among studies [8]. Prior analyses of PCI occurs almost exclusively in high-risk neonates obscuring insights into non-neonates. The current literature does not adequately address potentially preventable predictors of PCI, such as critical care practice or quality variables. Lastly, almost no data exist describing inter-hospital variation in PCI after CHS – variation that may provide insight into best practice.
In order to better understand PCI after CHS, we performed an analysis of the Pediatric Cardiac Critical Care Consortium (PC4) clinical registry. We aimed to define PCI within distinct age-groups and to identify unique predictors of PCI and PCI-mortality within these strata. Finally, we analyzed variation in case-mix-adjusted PCI across hospitals to determine whether opportunities for quality improvement (QI) exist for hospitals with higher-than-expected rates of PCI.
Materials and Methods
Data Source
The Pediatric Cardiac Critical Care Consortium (PC4) is a quality improvement collaborative that collects data on all patients with primary cardiac disease admitted to the CICU service of participating hospitals [9]. PC4 maintains a clinical registry to support research and quality improvement initiatives. At the time of this analysis, 22 hospitals were submitting cases to the PC4 registry.
Each participating hospital has a trained data manager who has completed a certification exam. The data managers collect and enter data in accordance with the standardized PC4 Data Definitions Manual. The PC4 registry shares common terminology and definitions with applicable data points from the International Pediatric and Congenital Cardiac Code (IPCCC), Society of Thoracic Surgeons (STS) Congenital Heart Surgery Database, and American College of Cardiology Improving Pediatric and Adult Congenital Treatment (IMPACT) Registry, as previously described [9]. Participating hospitals are audited on a regular schedule and audit results suggest complete, accurate and timely submission of data across hospitals, with the most recent published results demonstrating a major discrepancy rate of 0.6% across 29,476 fields [10]. The University of Michigan Institutional Review Board provides oversight for the PC4 Data Coordinating Center; this study was reviewed and approved with waiver of informed consent.
Patient selection
We analyzed all index surgical CICU encounters 8/2014–12/2016. These encounters included the index operation for the hospitalization as defined by the Society of Thoracic Surgeons Congenital Heart Surgery Database [11]. The sample was stratified into neonatal (≤28 days) and non-neonatal (29 days-18 years) strata for analysis. Twenty-two PC4 hospitals were included.
Defining PCI
The postoperative critical care duration (CCD) was aggregated across all hospitals for each stratum, beginning with CICU admission after the index operation. To define its end, we compared patients’ “CICU discharge date” (physically discharged from CICU) and “critical care end time” (no longer required CICU therapies/monitoring). The latter allows for single-bed care delivery systems (no physical CICU discharge) or when non-patient related factors delay CICU discharge (e.g. lack of general care bed). We found no difference (data not shown) therefore “critical care end time” was utilized to include all hospitals. CCD was capped at 180 days. CICU readmissions were not included. The 90th percentile of CCD within each strata was selected as the PCI-cutoff (35 and 10 days for neonates and non-neonates respectively).
Statistical analysis
Most recorded clinical events are time stamped allowing for 3 analyses: 1) identifying clinical predictors of PCI, 2) describing outcomes after PCI, and 3) identifying clinical predictors of PCI-mortality. Each was performed within the neonate and non-neonate strata.
Candidate variables predicting PCI and PCI-mortality included A) pre-operative patient demographic factors and comorbidities, B) operative factors, C) post-operative critical care delivery, and D) complications. Each candidate variable was assessed for association with PCI or PCI-mortality using the appropriate parametric or non-parametric test based on the distribution of data. Due to the high degree of multicollinearity between variables, we used the variance inflation factor test to determine those variables that should be dropped before multivariable analysis. The investigative team also excluded some collinear variables based on clinical relevance. We included each remaining predictor associated with the outcome at p<0.1 in a multivariable logistic regression model accounting for clustering within hospitals. We used backward stepwise selection to derive the final set of predictors independently associated with the outcome at p<0.05.
In order to assess variation in PCI across hospitals, case-mix adjusted PCI was calculated for each hospital. A model of quality-independent patient factors was created to calculate expected PCI-incidence based on the results of the multivariable logistic regression described above for analysis 1. Post-operative care and complications factors associated with PCI were excluded because they may be heavily influenced by a hospital’s quality of care, and therefore inappropriate for the purpose of case-mix adjustment and quality measurement. Using this more parsimonious model, we calculated observed-to-expected ratios (O:E) by dividing observed PCI-incidence by what was expected from the model. We then empirically derived 95% confidence intervals (CI) around the O:E ratio using bootstrap resampling (1000 samples). We measured hospital- adjusted PCI within both neonate and non-neonate strata. As neonates comprise a relatively small portion of the overall case volume, we report these data in the entire population at each hospital after calculating a weighted overall PCI-incidence rate (O:E ratio) based on the frequency of neonates and non-neonates. Analyses were performed using SAS Version 9.4 (SAS Institute, Cary, NC) or STATA Version 14 (Stata Corp, College Station, TX), with statistical significance at a p-value of less than 0.05.
Results
Defining PCI
The cohort included 2,419 neonates and 10,687 non-neonates. The PCI-cutoff was 35 and 10 days for neonates and non-neonates, respectively. One-hundred seven neonates (4%) and 37 non-neonates (0.4%) died prior to the PCI-cutoff and were excluded from subsequent analyses, leaving final analytic cohorts of 2312 neonates and 10,650 non-neonates. Table 1 shows the characteristics of each stratum.
Table 1:
Study population characteristics
| Characteristic | Neonates n=2312 |
Non-neonates n=10650 |
|---|---|---|
| Age at surgery | 7 days (4–11) | 24 months (5–93) |
| Male | 1387 (60%) | 5824 (55%) |
| Hispanic | 354 (15%) | 1622 (15%) |
| Any chromosomal anomaly, syndrome, or extracardiac anomaly | 591 (26%) | 3338 (31%) |
| Any STS pre-operative risk factor | 823 (36%) | 3418 (32%) |
| CICU | 3 (0–6) | 0 (0–0) |
| Unassigne | 7 (0.3%) | 162 (2%) |
| 2 or more | 3 (0.1%) | 2429 (23%) |
Data reported as n (%) or Median (Intra-quartile range) as appropriate unless noted otherwise
STS = The Society of Thoracic Surgeons, CICU = cardiac intensive care unit, STAT = The Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery
Identifying predictors of PCI
Univariate analysis identified 70/84 neonatal and 79/86 non-neonatal candidate predictors of PCI with p<0.1 (See Supplementary Table 1) that occurred prior to PCI-cutoff (when time stamp data was available). Table 2 shows the independent predictors of PCI from multivariable logistic regression after removing collinear factors.
Table 2:
Independent predictors of both prolonged critical illness and mortality after prolonged critical illness in neonates and non-neonates
| Covariate | PCI Predictors Odds Ratio (95% CI) |
PCI-Mortality Predictors Odds Ratio (95% CI) |
||
|---|---|---|---|---|
| Neonates | Non-neonates | Neonates | Non-neonates | |
| Patient and pre-operative factors | ||||
| Any extracardiac abnormality | 2.1 (1.4–3.0) | NS | NS | NS |
| Any chromosomal anomaly or syndrome | NS | 1.6 (1.3–1.9) | NS | NS |
| Patient intubated pre-operatively | NS | 2.1 (1.6–2.8) | NS | NS |
| Preterma | 2.6 (1.7–3.9) | N/A | NS | NS |
| Pre-operative CICU length of stay, days(d)a | 1.1 (1.0–1.1) | NS | NS | |
| Age at surgery <6 monthsb | N/A | 1.6 (1.3–2.0) | NS | NS |
| Pre-op hospitalization >1 day (Ref=≤1)b | 1.3 (1.0–1.7) | NS | NS | |
| ≥2 previous cardiac surgeries (Ref =0) | 2.1 (1.6–2.9) | NS | NS | |
| Any pre-operative STS risk factor | NS | NS | NS | 2.3 (1.2–4.2) |
| Operative factors | ||||
| STAT Score | 1.4 (1.2–1.7) | 1.2 (1.0–1.3) | NS | NS |
| Circulatory arrest required | NS | 1.8 (1.0–3.1) | NS | NS |
| Post-operative care delivery | ||||
| Post-operative mechanical ventilation, d | 1.2 (1.2–1.3) | 1.9 (1.7–2.0) | 1.01 (1.00–1.02) | 1.02 (1.01–1.03) |
| Time to post-operative enteral feed initiation, d | 1.1 (1.0–1.1) | 1.3 (1.3–1.4) | NS | NS |
| Inhaled nitric oxide use | NS | 1.8 (1.1–2.8) | NS | 1.8 (1.1–2.8) |
| Vasoactive support, d | NS | 2.0 (1.3–3.1) | 1.01 (1.00–1.02) | NS |
| Complications | ||||
| Deep surgical site infection | 20.0 (7.0–57.6) | 12.0 (4.2–34.6) | NS | NS |
| Any infection, non-deep surgical site | 3.4 (1.3–8.8) | 2.9 (1.7–4.7) | NS | NS |
| Renal replacement therapy for ARF | 5.4 (2.0–14.4) | 16.2 (1.6–160.6) | 3.4 (1.3–8.5) | 8.3 (4.1–16.9) |
| Treatment for chylothorax/pleural effusion | 2.2 (1.3–3.8) | 2.8 (1.9–4.0) | NS | NS |
| Cardiac arrest | NS | 1.6 (1.1–2.5) | 2.6 (1.2–5.6) | 5.9 (3.3–10.6) |
| Paralyzed diaphragm | NS | 9.1 (2.4–34.7) | 0.1 (0.01–0.7) | NS |
| Stroke or intracranial hemorrhages | NS | 2.5 (1.0–5.9) | NS | NS |
| Unplanned reoperation | NS | NS | 3.1 (1.4–6.8) | 3.1 (1.7–5.5) |
| Mechanical circulatory support | NS | NS | NS | 3.9 (2.2–7.1) |
PCI = prolonged critical illness, CI = confidence interval, NS = not significant, N/A = not applicable, CICU = cardiac intensive care unit, STS = The Society of Thoracic Surgeons, STAT = The Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery, ARF = acute renal failure
Factor unique to neonates
Factor unique to non-neonates
Outcomes after PCI
A greater proportion of PCI neonates died during the index postoperative CICU encounter (57 of 242 (24%)) compared to PCI non-neonates (92 of 1184 (8%)). Overall hospital mortality was 67 of 242 for (28%) and 110 of 1184 (9%) for PCI neonates and non-neonates respectively. The majority of overall deaths was in PCI patients (176/323) with PCI neonates and non-neonates accounting for 39% (67/174) and 75% (110/147) of in-stratum mortality, respectivley.
Median length of stay (LOS) for PCI neonates and non-neonates was 58 and 17 days, respectively. Survivors from both strata had shorter LOS than those who died (55 vs 83 days in neonates, p=0.0027; 16 vs 28.5 days in non-neonates, p<0.0001).
PCI patients were commonly readmitted to the CICU after discharge from their index CICU encounter. Of the 185 neonates who survived to initial CICU discharge, 48 required readmission to the CICU (26%) and 7 of those died prior to hospital discharge (15%). Of the 1092 non-neonates who survived to initial CICU discharge, 172 required readmission to the CICU (16%) and 9 of those died prior to hospital discharge (6%). Additional outcomes are shown in Supplementary Table 2.
Identifying predictors of mortality in PCI patients
Univariate analysis identified 32/81 neonatal and 47/79 non-neonatal candidate predictors of PCI-mortality with p<0.1 (See Supplementary Table 3). These occurred through the entirety of the index encounter. Surviving PCI neonates suffered a median 3 complications (IQR 2–5) compared to 6 (IQR 5–7) in non-survivors (p<0.0001). Surviving PCI non-neonates experienced a median 2 complications (IQR 1–3) compared to 6 (IQR 4–7) in non-survivors (p<0.0001).
Table 2 shows the independent predictors of PCI-mortality. Of note, total CICU LOS was associated with mortality after PCI in both strata but was highly collinear with duration of mechanical ventilation (MV) and duration of vasoactive support so LOS was removed from the model. Figure 1 shows survival after the PCI-cutoff over time.
Figure 1:
Survival after PCI. This Kaplan-Meier curve shows survival over time with PCI-cutoff serving as day 0 (post-operative day 35 and for neonates and non-neonates day 10 respectively).
Figure 2 summarizes the predictors of both PCI and PCI-mortality. Of note, paralyzed diaphragm independently protected against PCI-mortality in neonates and was excluded from the figure.
Figure 2:
Predictors of PCI and PCI- mortality. This Venn diagram includes predictors of PCI (large circle), PCI-mortality (small circle), and both (overlap of circles). Predictors unique to neonatal stratum are in the upper section, predictors unique to non-neonatal strata are in the lower section, and predictors for both strata are in the middle section. Italics indicates factor appears in diagram twice.
STS = The Society of Thoracic Surgeons, STAT = The Society of Thoracic Surgeons-European Association for Cardio-Thoracic Surgery
a, b, and c Denote similar factors that are measured uniquely in each stratum where a is preterm in neonates and age <6 months for non-neonates, b is as any extracardiac abnormality in neonates and chromosomal anomaly or syndrome in non-neonates, and c is pre-operative CICU length of stay (in days) in neonates and pre-operative hospitalization >1 day in non-neonates
Variation in PCI across hospitals
The observed incidence of PCI for each of the 22 hospitals ranged from 5–21% overall, 1–30% for neonates, and 6–20% for non-neonates. Figure 3 shows O:E ratios of PCI-incidence by hospital. Overall, given their case-mix, PCI-incidence was lower-than-expected in 6 hospitals, higher-than-expected in 6 hospitals, and as expected in 10 hospitals. One had lower- and 4 had higher-than-expected PCI-incidence in both neonates and non-neonates. One hospital had lower-than-expected incidence of PCI for its neonates, but higher-than-expected for its non-neonates.
Figure 3:
Variation in PCI-incidence by hospital. These graphs show the observed-to expected ratio (O:E) and 95% confidence interval for the overall cohort (A), neonatal stratum (B), and non-neonatal stratum (C) for 22 hospitals. Hospitals with lower-than-expected (*) and higher-than-expected (^) incidence of PCI are denoted.
Discussion
In this analysis of PCI after CHS from a multi-institutional CICU clinical registry, we identified important predictors of PCI and PCI-mortality and important variation in PCI-incidence across hospitals. This study reveals critical insights on a population that accounts for the majority of overall mortality. We noted different factors predicting PCI and PCI-mortality in the two study strata, which may have clinical implications.
In defining PCI, our cutoffs of 35 days in neonates and 10 days in non-neonates are remarkably similar to those proposed in a review by Shapiro et al. [12] (>28 days post-term corrected age and >14 days respectively).
Our analysis identified over 20 independent predictors of PCI in what we believe is the largest study of this population to-date. In many previous investigations, age drives morbidity and mortality [1]. Stratifying allowed identification of unique predictors, which can inform PCI-prevention. For example, focusing on improving management of paralyzed diaphragm would more likely benefit a hospital with higher-than-expected PCI-incidence in non-neonates but not in neonates.
We identified 8 predictors of PCI-mortality specific to CHS patients, many of which are unique compared to previous literature because of our stratified approach. Of note, we grouped predictors into four categories: pre-operative/patient factors, operative factors, post-operative care delivery, and complications. We cannot determine if the factors associated with PCI and PCI-mortality are cause or effect; some may be surrogates for other clinical entities (e.g. residual cardiac lesions after the index operation). Nevertheless, in general, we consider the pre-operative/patient factors and operative factors to be non-modifiable, while need for advanced and/or prolonged critical care therapies and complications are potentially preventable. In contrast to PCI-incidence predictors where there was a mix of preventable and non-modifiable factors, 7 of 8 of the mortality predictors are potentially preventable.
We categorize persistent dependence on critical care therapies as “potentially preventable”. We note that this is may be a marker for underlying illness and prevention must be aimed at the underlying causes. However, hospitals can derive important knowledge by understanding the epidemiology of their potentially preventable predictors of PCI, focusing efforts on the highest leverage drivers of PCI and PCI-mortality. At the very least, better and earlier understanding of when patients are at risk for PCI and PCI-mortality can be used for prognostication, counseling, and informing when to change course in a patient displaying these characteristics. Since we don’t yet know the optimal way to act on these data for at-risk patients to prevent PCI and its sequelae, dedicated investigation is necessary to discover the practices at high performing centers with low case-mix adjusted PCI and PCI-mortality rates.
Two such predictors predicting PCI and PCI-mortality in both strata are duration of post-operative MV and acute renal failure requiring RRT. Respiratory failure is multifactorial, however this raises important questions about approaches to chronic respiratory failure. Specifically, should patients requiring long-term MV after CHS receive earlier tracheostomy than is common in current practice? MV via a tracheostomy can allow for weaning of sedatives, better nutrition, and other reductions in critical care therapies. Though current pediatric literature does not suggest improved mortality with tracheostomy, this may be because it is offered only after PCI, and therefore after the sequelae of PCI. Similar questions may apply to the ideal timing of RRT initiation [13], a well-established predictor of PCI and PCI-mortality [1, 4]. Finally, while it is true that for individual patients these therapies may be life-saving and appropriate care, hospitals with high rates of respiratory and renal failure leading to tracheostomy and RRT would benefit from developing strategies that reduce organ failure in an effort to prevent PCI and PCI-mortality.
Cardiac arrest (CA) and need for mechanical circulatory support (MCS) predicted PCI-mortality in both strata, but more weakly predicted PCI-incidence (CA predicted in non-neonates only, while MCS was not a predictor). This may be influenced by the exclusion of those who died prior to the PCI-cutoff. Surprisingly, unplanned reoperation similarly predicted PCI-mortality in both strata but PCI in neither. We utilized various regression techniques and reoperation continued to show no association with PCI-incidence, indicating our first model was appropriate. Similarly, this could be influenced by excluding those who died prior to PCI-cutoff. Inadequate repair has been shown to correlate with both early mortality (<30 days) and prolonged post-operative LOS [14]. Perhaps early recognition and prompt attention to residual lesions can mitigate these effects. This could also reflect an analysis limitation – PCI predictors had to have occurred prior to the PCI-cutoff. Residual heart disease that was either missed or recognized but not acted upon is not adequately captured.
This study highlights the importance of preventing PCI based on the outcomes these patients experience. Mortality is much worse in PCI patients; 7- and 23-fold higher, respectively, in neonates and non-neonates than their non-PCI counterparts. Further, remaining in the CICU over time is associated with gradually increasing mortality rate as seen in Figure 1. Remaining in the CICU 6 months after the PCI cutoff portends a 60% morality risk. Further, we showed PCI patients who leave the CICU temporarily and require readmission still have important discharge mortality rates (15% and 6% in readmitted neonates and non-neonates, respectively). Our results compare similarly to those published in a recent systematic review [8] in which PCI subjects had both higher discharge and after hospital discharge mortality. Further, they found more deficits in long-term quality of life and function, which is a similar finding among CHS patients who experience impaired neurodevelopment after prolonged ICU LOS [15, 16]. The impact of PCI has been long recognized to correlate tightly with costs as well [17].
We found significant variation in case-mix-adjusted PCI-incidence across hospitals. These metrics do not indicate that one hospital is better or worse than another (a hospital with a lower-than-expected O:E ratio may not perform the same way with another hospital’s case-mix, and vice versa). However, most hospitals in this cohort have a similar case-mix of neonatal and high-complexity surgery. Therefore, differences in PCI suggest that ideal clinical practices to prevent PCI may exist. More research is necessary to understand if and how high-performing hospitals achieve these results. Regardless, if an individual hospital has a higher-than-expected O:E ratio for PCI-incidence, providers can look at their hospital’s rate of PCI-predictors to begin identifying potential drivers of PCI and areas for QI.
The limitations of this study are similar to any observational analysis of a clinical registry, namely the potential for unmeasured confounders of PCI[18]. Further, this study sought to characterize PCI and mortality after PCI, thus excluding those dying prior to PCI-cutoff; possibly, centers with higher pre-PCI-mortality could have lower PCI-incidence. Further analysis of patient data from the 107 neonates and 37 non-neonates who died prior to the PCI-cutoff, along with center-level variation in pre- and post-PCI-mortality will be instrumental in helping understand this critical population. Another limitation is that we studied only the index CICU encounter, though we did include certain outcomes from the entire hospitalization. With ongoing integrative efforts between PC4 and the Pediatric Acute Care Cardiology Collaborative [19] within Cardiac Networks United [20], perhaps a more complete understanding of PCI will be possible in future analyses.
Conclusions
We found potentially preventable predictors of PCI and mortality after PCI for patients undergoing CHS. The wide variation in incidence of PCI across our cohort of hospitals suggests there are opportunities to improve outcomes for children undergoing CHS by reducing the burden of PCI. Improvement begins with hospitals examining post-operative critical care delivery systems and mechanisms of preventing and responding to complications that drive PCI and PCI-mortality. This analysis could promote both discussions between hospitals with high- and low-PCI-incidence to help discover best practices and collaborative QI initiatives to reduce the burden of illness from PCI after CHS.
Supplementary Material
Acknowledgements
We acknowledge the data collection teams of the participating hospitals.
Funding Sources
This study was supported in part by the University of Michigan Congenital Heart Center, CHAMPS for Mott, and the Michigan Institute for Clinical & Health Research (NIH/NCATS UL1TR002240). Dr. Gaies was supported in part by funding from the National Institutes of Health/National Heart, Lung, and Blood Institute (K08HL116639).
Copyright form disclosure: Dr. DeWitt disclosed that this study was supported in part by funding from the University of Michigan Congenital Heart Center, Champs for Mott, and the Michigan Institute for Clinical & Health Research (NIH/NCATS UL1TR002240). Dr. Rossano received funding from Amgen, Bayer, CSL Behring, and Abbott. Dr. Moga’s institution received funding from the Canadian Institutes of Health Research. Dr. Gaies’ institution received funding from the National Institutes of Health (NIH). The remaining authors have disclosed that they do not have any potential conflicts of interest.
Footnotes
Disclosures: G.E. Owens is a consultant for HistoSonics. The remaining authors have disclosed they do not have any potential conflicts of interest.
References
- 1.Namachivayam SP, d’Udekem Y, Millar J, et al. : Survival status and functional outcome of children who required prolonged intensive care after cardiac surgery. J Thorac Cardiovasc Surg 2016, 152(4):1104–1112 e1103. [DOI] [PubMed] [Google Scholar]
- 2.Baker-Smith CM, Wilhelm CM, Neish SR, et al. : Predictors of prolonged length of intensive care unit stay after stage I palliation: a report from the National Pediatric Cardiology Quality Improvement Collaborative. Pediatr Cardiol 2014, 35(3):431–440. [DOI] [PubMed] [Google Scholar]
- 3.Schwartz SM, Lu M, Ohye RG, et al. : Risk factors for prolonged length of stay after the stage 2 procedure in the single-ventricle reconstruction trial. J Thorac Cardiovasc Surg 2014, 147(6):1791–1798, 1798 e1791–1794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mori M, McCracken C, Maher K, et al. : Outcomes of neonates requiring prolonged stay in the intensive care unit after surgical repair of congenital heart disease. J Thorac Cardiovasc Surg 2016, 152(3):720–727 e721. [DOI] [PubMed] [Google Scholar]
- 5.Marcin JP, Slonim AD, Pollack MM, et al. : Long-stay patients in the pediatric intensive care unit. Critical care medicine 2001, 29(3):652–657. [DOI] [PubMed] [Google Scholar]
- 6.Franck LS, McQuillan A, Wray J, et al. : Parent stress levels during children’s hospital recovery after congenital heart surgery. Pediatr Cardiol 2010, 31(7):961–968. [DOI] [PubMed] [Google Scholar]
- 7.Brennan A, Gauvreau K, Connor J, et al. : A Method to Account for Variation in Congenital Heart Surgery Length of Stay. Pediatr Crit Care Med 2017, 18(6):550–560. [DOI] [PubMed] [Google Scholar]
- 8.Trivedi V, Bleeker H, Kantor N, et al. : Survival, Quality of Life, and Functional Status Following Prolonged ICU Stay in Cardiac Surgical Patients: A Systematic Review. Critical care medicine 2019, 47(1):e52–e63. [DOI] [PubMed] [Google Scholar]
- 9.Gaies M, Cooper DS, Tabbutt S, et al. : Collaborative quality improvement in the cardiac intensive care unit: development of the Paediatric Cardiac Critical Care Consortium (PC4). Cardiol Young 2015, 25(5):951–957. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gaies M, Donohue JE, Willis GM, et al. : Data integrity of the Pediatric Cardiac Critical Care Consortium (PC4) clinical registry. Cardiology in the young 2015:1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.STS Congenital Heart Surgery Database Data Specifications. Version 3.22 [http://www.sts.org/sites/default/files/documents/CongenitalDataSpecsV3_22.pdf]
- 12.Shapiro MC, Henderson CM, Hutton N, et al. : Defining Pediatric Chronic Critical Illness for Clinical Care, Research, and Policy. Hosp Pediatr 2017, 7(4):236–244. [DOI] [PubMed] [Google Scholar]
- 13.Namachivayam SP, Butt W, Millar J, et al. : Early Peritoneal Dialysis and Major Adverse Events After Pediatric Cardiac Surgery: A Propensity Score Analysis. Pediatr Crit Care Med 2019, 20(2):158–165. [DOI] [PubMed] [Google Scholar]
- 14.Nathan M, Karamichalis J, Liu H, et al. : Technical Performance Scores are strongly associated with early mortality, postoperative adverse events, and intensive care unit length of stay-analysis of consecutive discharges for 2 years. J Thorac Cardiovasc Surg 2014, 147(1):389–394, 396.e381–396.e383. [DOI] [PubMed] [Google Scholar]
- 15.Fuller S, Nord AS, Gerdes M, et al. : Predictors of impaired neurodevelopmental outcomes at one year of age after infant cardiac surgery. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery 2009, 36(1):40–47. [DOI] [PubMed] [Google Scholar]
- 16.Limperopoulos C, Majnemer A, Shevell MI, et al. : Functional limitations in young children with congenital heart defects after cardiac surgery. Pediatrics 2001, 108(6):1325–1331. [DOI] [PubMed] [Google Scholar]
- 17.Silberbach M, Shumaker D, Menashe V, et al. : Predicting hospital charge and length of stay for congenital heart disease surgery. The American journal of cardiology 1993, 72(12):958–963. [DOI] [PubMed] [Google Scholar]
- 18.Pasquali SK, Gaies M, Banerjee M, et al. : The quest for precision medicine: Unmeasured patient factors and mortality after congenital heart surgery. Ann Thorac Surg 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kipps AK, Cassidy SC, Strohacker CM, et al. : Collective quality improvement in the paediatric cardiology acute care unit: establishment of the Pediatric Acute Care Cardiology Collaborative (PAC3). Cardiology in the young 2018, 28(8):1019–1023. [DOI] [PubMed] [Google Scholar]
- 20.Gaies M, Anderson J, Kipps A, et al. : Cardiac Networks United: an integrated paediatric and congenital cardiovascular research and improvement network. Cardiology in the young 2018:1–8. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.