Predicting long-term neurocognitive outcome after pediatric intensive care unit admission for bronchiolitis-preliminary exploration of the potential of machine learning.

PURPOSE: For successful prevention and intervention, it is important to unravel the complex constellation of factors that affect neurocognitive functioning after pediatric intensive care unit (PICU) admission. This study aims (1) to elucidate the potential relevance of patient and PICU-related characteristics for long-term adverse neurocognitive outcome after PICU admission for bronchiolitis, and (2) to perform a preliminary exploration of the potential of machine learning as compared to linear regression to improve neurocognitive outcome prediction in a relatively small sample of children after PICU admission. METHODS: This cross-sectional observational study investigated 65 children aged 6-12 years with previous PICU admission for bronchiolitis (age ≤ 1 year). They were compared to demographically comparable healthy peers (n = 76) on neurocognitive functioning. Patient and PICU-related characteristics used for the prediction models were as follows: demographic characteristics, perinatal and disease parameters, laboratory results, and intervention characteristics, including hourly validated mechanical ventilation parameters. Neurocognitive outcome was measured by intelligence and computerized neurocognitive testing. Prediction models were developed for each of the neurocognitive outcomes using Regression Trees, k-Nearest Neighbors, and conventional linear regression analysis. RESULTS: The patient group had lower intelligence than the control group (p < .001, d = -0.59) and poorer performance in neurocognitive functions, i.e., speed and attention (p = .03, d = -0.41) and verbal memory (p < .001, d = -0.60). Lower intelligence was predicted by lower birth weight and lower socioeconomic status (R2 = 25.9%). Poorer performance on the speed and attention domain was predicted by younger age at follow-up (R2 = 53.5%). Poorer verbal memory was predicted by lower birth weight, younger age at follow-up, and greater exposure to acidotic events (R2 = 50.6%). The machine learning models did not reveal added value in terms of model performance as compared to linear regression. CONCLUSION: The findings of this study suggest that in children with previous PICU admission for bronchiolitis, (1) lower birth weight, younger age at follow-up, and lower socioeconomic status are associated with poorer neurocognitive outcome; and (2) greater exposure to acidotic events during PICU admission is associated with poorer verbal memory outcome. The findings of this study provide no evidence for the added value of machine learning models as compared to linear regression analysis in the prediction of long-term neurocognitive outcome in a relatively small sample of children. WHAT IS KNOWN: • Adverse neurocognitive outcomes are described in PICU survivors, which are known to interfere with development in other major domains of functioning, such as mental health, academic achievement, and socioeconomic success, highlighting neurocognition as an important outcome after PICU admission. • Machine learning is a rapidly growing field of artificial intelligence that is increasingly applied in health care settings, with great potential to capture the complexity of outcome prediction. WHAT IS NEW: • This study shows that lower birth weight, lower socioeconomic status, and greater exposure to acidotic events during PICU admission for bronchiolitis are associated with poorer long-term neurocognitive outcome after PICU admission. Results provide no evidence for the added value of machine learning models in a relatively small sample of children. • As bronchiolitis seldom manifests neurologically, the relation between acidotic events and neurocognitive outcome may reflect either potentially harmful effects of acidosis itself or related processes such as hypercapnia or hypoxic and/or ischemic events during PICU admission. This study further highlights the importance of structured follow-up to monitor long-term outcome of children after PICU admission.

Long-Term Follow-Up of Daily Life Functioning After Pediatric Intensive Care Unit Admission.

OBJECTIVE: To investigate the long-term impact of pediatric intensive care unit (PICU) admission on daily life functioning while exploring the potential mediating role of neurocognitive outcome. STUDY DESIGN: This cross-sectional observational study compared children aged 6-12 years with previous PICU admission (age ≤1 year) for bronchiolitis requiring mechanical ventilation ("patient group," n = 65) to demographically comparable healthy peers ("control group," n = 76). The patient group was selected because bronchiolitis is not expected to affect neurocognitive functioning in itself. Assessed daily life outcome domains were behavioral and emotional functioning, academic performance, and health-related quality of life (QoL). The role of neurocognitive outcomes in the relationship between PICU admission and daily life functioning was assessed by mediation analysis. RESULTS: The patient group did not differ from the control group regarding behavioral and emotional functioning but performed poorer on academic performance and school-related QoL (Ps ≤ .04, d = -0.48 to -0.26). Within the patient group, lower full-scale IQ (FSIQ) was associated with poorer academic performance and school-related QoL (Ps ≤ .02). Poorer verbal memory was associated with poorer spelling performance (P = .002). FSIQ mediated the observed effects of PICU admission on reading comprehension and arithmetic performance. CONCLUSIONS: Children admitted to the PICU are at risk for long-term adverse daily life outcomes in terms of academic performance and school-related QoL. Findings suggest that lower intelligence may contribute to academic difficulties after PICU admission. Findings underline the importance of monitoring daily life and neurocognitive functioning after PICU admission.

Structured Multidisciplinary Follow-Up After Pediatric Intensive Care: A Model for Continuous Data-Driven Health Care Innovation.

OBJECTIVES: Morbidity after PICU admission for critical illness is a growing concern. Sequelae may occur in various domains of functioning and can only appropriately be determined through structured follow-up. Here, we describe the process of designing and implementing a structured multidisciplinary follow-up program for patients and their parents after PICU admission and show the first results illustrating the significance of our program. DESIGN: Prospective observational cohort study. SETTING: Outpatient PICU follow-up clinic. PATIENTS: Patients 0-18 years old admitted to our PICU. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: In our structured multidisciplinary follow-up program, follow-up care is provided by a pediatric intensivist and psychologist and in addition, depending on patient's critical illness and received PICU treatment(s), by a pediatric pulmonologist, cardiologist, neurologist, and/or neuropsychologist. All consultations are scheduled consecutively. Collected data are stored in a hospital-wide data warehouse and used for yearly health care evaluation sessions as well as scientific research. Challenges in organizing this follow-up program include technological challenges, providing time-efficient care, participation rate, and completeness of questionnaires. In our experience, a dedicated team is essential to tackle these challenges. Our first results, obtained in 307 of 388 referred patients (79.1%), showed the diversity of problems arising after PICU discharge, including physical, neurocognitive, and psychosocial sequelae. In addition, our data also reflected the risk of psychosocial problems among parents. Within the limited operation time of our follow-up program, the program has evolved based on our experiences and the data collected. CONCLUSIONS: We successfully developed and implemented a structured multidisciplinary follow-up program for patients and their parents after PICU admission. This program may help to timely initiate appropriate interventions, improve the standard of care during and after PICU admission, and facilitate scientific research on outcome and prognosis after PICU admission.

The Local and Systemic Exposure to Oxygen in Children With Severe Bronchiolitis on Invasive Mechanical Ventilation: A Retrospective Cohort Study.

OBJECTIVES: Oxygen supplementation is a cornerstone treatment in critically ill children with bronchiolitis in the PICU. However, potential deleterious effects of high-dose oxygen are well-known. In this study, we aim to describe the pulmonary (local) and arterial (systemic) oxygen exposure over the duration of invasive mechanical ventilation (IMV) in children with severe bronchiolitis. Our secondary aim was to estimate potentially avoidable exposure to high-dose oxygen in these patients. DESIGN: Retrospective cohort study. SETTING: Single-center, tertiary-care PICU. PATIENTS: Children younger than 2 years old admitted to the PICU for severe bronchiolitis receiving IMV. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Hourly measurements of Fio2 and peripheral oxygen saturation (Spo2), and arterial blood gas data were collected up to day 10 of IMV. A total of 24,451 hours of IMV were observed in 176 patients (median age of 1.0 mo [interquartile range (IQR), 1.0-2.3 mo]). The pulmonary exposure to oxygen was highest during the first day of IMV (median time-weighted average [TWA]-Fio2 0.46 [IQR, 0.39-0.53]), which significantly decreased over subsequent days. The systemic exposure to oxygen was relatively low, as severe hyperoxemia (TWA-Pao2 > 248 Torr [> 33 kPa]) was not observed. However, overuse of oxygen was common with 52.3% of patients (n = 92) having at least 1 day of possible excessive oxygen exposure and 14.8% (n = 26) with severe exposure. Furthermore, higher oxygen dosages correlated with increasing overuse of oxygen (rrepeated measures, 0.59; 95% CI, 0.54-0.63). Additionally, caregivers were likely to keep Fio2 greater than or equal to 0.50 when Spo2 greater than or equal to 97%. CONCLUSIONS: Moderate to high-dose pulmonary oxygen exposure and potential overuse of oxygen were common in this cohort of severe bronchiolitis patients requiring IMV; however, this was not accompanied by a high systemic oxygen burden. Further studies are needed to determine optimal oxygenation targets to prevent overzealous use of oxygen in this vulnerable population.

Long-term neurocognitive outcomes after pediatric intensive care: exploring the role of drug exposure.

BACKGROUND: Concerns exist regarding the impact of widely used clinical drugs on brain development. This study investigates long-term neurocognitive functioning in relation to frequently used drug exposure at the Pediatric Intensive Care Unit (PICU). METHODS: This study compared children aged 6-12 years with previous PICU admission (age ≤1 year) for bronchiolitis requiring mechanical ventilation (patient group, n = 65) to a demographically comparable control group (n = 76) on a broad range of neurocognitive outcomes. The patient group was selected because bronchiolitis seldom manifests neurologically and is therefore not expected to affect neurocognitive functioning in itself. The relation between exposure to sedatives, analgesics and anesthetics and neurocognitive outcomes was assessed by regression analyses. RESULTS: The patient group had lower intelligence than the control group (p < 0.001, d = -0.59) and poorer performance in neurocognitive functions; i.e., speed and attention (p = 0.03, d = -0.41) and verbal memory (p < 0.001, d = -0.60). Exposure to sedatives, analgesics and anesthetics was not related to neurocognitive outcomes. CONCLUSIONS: Children with PICU admission for bronchiolitis requiring mechanical ventilation are at risk of adverse neurocognitive outcomes. This study found no evidence for a role of exposure to sedatives, analgesics or anesthetics. Findings underline the importance of long-term follow-up after PICU admission, even in the absence of disease with neurological manifestation. IMPACT: Animal studies have indicated that exposing the maturing brain to clinical drugs may cause neurodegeneration. Clinical studies show mixed evidence regarding the association between clinical drugs and neurocognitive outcomes. This study provides evidence for considerably lower neurocognitive functioning among children with a history of PICU admission for bronchiolitis compared to healthy peers. Bronchiolitis seldom manifests neurologically and is therefore not expected to affect neurocognitive functioning in itself. We found no evidence supporting a relation between drug exposure (i.e., sedatives, analgesics and anesthetics) and long-term neurocognitive outcomes. Findings underline the importance of structured follow-up after PICU admission.

Neurocognitive, Psychosocial, and Quality of Life Outcomes After Multisystem Inflammatory Syndrome in Children Admitted to the PICU.

OBJECTIVES: To investigate neurocognitive, psychosocial, and quality of life (QoL) outcomes in children with Multisystem Inflammatory Syndrome in Children (MIS-C) seen 3-6 months after PICU admission. DESIGN: National prospective cohort study March 2020 to November 2021. SETTING: Seven PICUs in the Netherlands. PATIENTS: Children with MIS-C (0-17 yr) admitted to a PICU. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Children and/or parents were seen median (interquartile range [IQR] 4 mo [3-5 mo]) after PICU admission. Testing included assessment of neurocognitive, psychosocial, and QoL outcomes with reference to Dutch pre-COVID-19 general population norms. Effect sizes (Hedges' g ) were used to indicate the strengths and clinical relevance of differences: 0.2 small, 0.5 medium, and 0.8 and above large. Of 69 children with MIS-C, 49 (median age 11.6 yr [IQR 9.3-15.6 yr]) attended follow-up. General intelligence and verbal memory scores were normal compared with population norms. Twenty-nine of the 49 followed-up (59%) underwent extensive testing with worse function in domains such as visual memory, g = 1.0 (95% CI, 0.6-1.4), sustained attention, g = 2.0 (95% CI 1.4-2.4), and planning, g = 0.5 (95% CI, 0.1-0.9). The children also had more emotional and behavioral problems, g = 0.4 (95% CI 0.1-0.7), and had lower QoL scores in domains such as physical functioning g = 1.3 (95% CI 0.9-1.6), school functioning g = 1.1 (95% CI 0.7-1.4), and increased fatigue g = 0.5 (95% CI 0.1-0.9) compared with population norms. Elevated risk for posttraumatic stress disorder (PTSD) was seen in 10 of 30 children (33%) with MIS-C. Last, in the 32 parents, no elevated risk for PTSD was found. CONCLUSIONS: Children with MIS-C requiring PICU admission had normal overall intelligence 4 months after PICU discharge. Nevertheless, these children reported more emotional and behavioral problems, more PTSD, and worse QoL compared with general population norms. In a subset undergoing more extensive testing, we also identified irregularities in neurocognitive functions. Whether these impairments are caused by the viral or inflammatory response, the PICU admission, or COVID-19 restrictions remains to be investigated.

Long-Term Pulmonary Outcomes in Children Mechanically Ventilated for Severe Bronchiolitis.

OBJECTIVES: Bronchiolitis is a common indication for mechanical ventilation in the PICU. Both bronchiolitis and invasive mechanical ventilation may cause adverse long-term pulmonary outcomes. This study investigates children with a history of invasive mechanical ventilation for bronchiolitis, addressing: 1) the extent, 2) potential explanatory factors, and 3) possible impact on daily life activities of adverse long-term pulmonary outcomes. DESIGN: Single-center cohort study. SETTING: Outpatient PICU follow-up clinic. PATIENTS: Children 6-12 years old with a history of invasive mechanical ventilation for bronchiolitis (age < 2 yr). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Long-term pulmonary outcomes were assessed by a standardized questionnaire and by spirometry. Nineteen out of 74 included children (26%) had adverse long-term pulmonary outcomes, of whom the majority had asthma (14/74, 19%). By logistic regression analysis, we assessed whether background characteristics and PICU-related variables were associated with long-term pulmonary outcomes. In general, we failed to identify any explanatory factors associated with adverse long-term pulmonary outcomes. Nonetheless, atopic disease in family and longer duration of invasive mechanical ventilation (days) were associated with greater odds of having asthma at follow-up (odds ratio, 6.4 [95% CI, 1.2-36.0] and 1.3 [95% CI, 1.0-1.7], respectively). Adverse pulmonary outcome at follow-up was associated with more frequent use of pulmonary medication after PICU discharge. In comparison with those without adverse pulmonary outcomes, we did not identify any difference in frequency of sports performance or school absenteeism. CONCLUSIONS: In this single-center cohort, one-quarter of the children attending follow-up with a history of invasive mechanical ventilation for bronchiolitis had adverse, mostly previously undetected, long-term pulmonary outcomes at 6-12 years. Atopic disease in family and longer duration of invasive mechanical ventilation were associated with presence of asthma. The presence of adverse pulmonary outcomes was associated with more frequent use of pulmonary medication after PICU discharge.

Intelligence outcome of pediatric intensive care unit survivors: a systematic meta-analysis and meta-regression.

BACKGROUND: Long-term morbidity after pediatric intensive care unit (PICU) admission is a growing concern. Both critical illness and accompanying PICU treatments may impact neurocognitive development as assessed by its gold standard measure; intelligence. This meta-analysis and meta-regression quantifies intelligence outcome after PICU admission and explores risk factors for poor intelligence outcome. METHODS: PubMed, Embase, CINAHL and PsycINFO were searched for relevant studies, published from database inception until September 7, 2021. Using random-effects meta-analysis, we calculated the standardized mean difference in full-scale intelligence quotient (FSIQ) between PICU survivors and controls across all included studies and additionally distinguishing between PICU subgroups based on indications for admission. Relation between demographic and clinical risk factors and study's FSIQ effect sizes was investigated using random-effects meta-regression analysis. RESULTS: A total of 123 articles was included, published between 1973 and 2021, including 8,119 PICU survivors and 1,757 controls. We found 0.47 SD (7.1 IQ-points) lower FSIQ scores in PICU survivors compared to controls (95%CI -0.55 to -0.40, p < .001). All studied PICU subgroups had lower FSIQ compared to controls (range 0.38-0.88 SD). Later year of PICU admission (range 1972-2016) and longer PICU stay were related to greater FSIQ impairment (R2 = 21%, 95%CI -0.021 to -0.007, p < .001 and R2 = 2%, 95%CI -0.027 to -0.002, p = .03, respectively), whereas male sex and higher rate of survivors were related to smaller FSIQ impairment (R2 = 5%, 95%CI 0.001 to 0.014, p = .03 and R2 = 11%, 95%CI 0.006 to 0.022, p < .001, respectively). Meta-regression in PICU subgroups showed that later year of PICU admission was related to greater FSIQ impairment in children admitted after cardiac surgery and heart- or heart-lung transplantation. Male sex was related to smaller FSIQ impairment in children admitted after cardiac surgery. Older age at PICU admission and older age at follow-up were related to smaller FSIQ impairment in children admitted after heart- or heart-lung transplantation. CONCLUSIONS: PICU survivors, distinguished in a wide range of subgroups, are at risk of intelligence impairment. Length of PICU stay, female sex and lower rate of survivors were related to greater intelligence impairment. Intelligence outcome has worsened over the years, potentially reflecting the increasing percentage of children surviving PICU admission.

Accuracy of the Diagnosis of Bronchopulmonary Dysplasia in a Referral-Based Health Care System.

OBJECTIVE: To evaluate the accuracy of the diagnosis of bronchopulmonary dysplasia (BPD) in a national database of a referral-based health care system, where preterm infants are often transferred back to regional hospitals before 36 weeks postmenstrual age (PMA). STUDY DESIGN: We evaluated preterm infants <32 weeks, born between 2004 and 2008 in the Academic Medical Center in Amsterdam with a high-risk profile for BPD. In addition to patient characteristics and outcomes, we collected data on respiratory support at 36 weeks PMA. True incidence of BPD, defined as needing supplemental oxygen and/or positive pressure support at 36 weeks PMA, was compared with the diagnosis registered in the National Perinatal Registry. Two imputation algorithms for patients transferred before 36 weeks PMA were validated. RESULTS: We identified 243 preterm infants with a high-risk BPD profile. Sixty-seven percent of these infants had a correct BPD diagnosis recorded in the National Perinatal Registry, 2% had a false positive, and 31% a false negative diagnosis. Infants with a false negative diagnosis of BPD were twice as often transferred to a regional hospital before 36 weeks PMA compared with a true positive diagnosis. Imputation algorithms did not improve the accuracy of BPD registration. CONCLUSIONS: Registration of the diagnosis BPD in a national database in countries with a referral-based health care system may not be accurate. Optimizing data collection and monitoring data entry is necessary to improve BPD registration before data can be used for national and international benchmarking.