Designing a bed-side system for predicting length of stay in a neonatal intensive care unit

Development and validation of a novel risk classification tool for predicting long length of stay in NICU blood transfusion infants

Newborns are as the primary recipients of blood transfusions. There is a possibility of an association between blood transfusion and unfavorable outcomes. Such complications not only imperil the lives of newborns but also cause long hospitalization. Our objective is to explore the predictor variables that may lead to extended hospital stays in neonatal intensive care unit (NICU) patients who have undergone blood transfusions and develop a predictive nomogram. A retrospective review of 539 neonates who underwent blood transfusion was conducted using median and interquartile ranges to describe their length of stay (LOS). Neonates with LOS above the 75th percentile (P75) were categorized as having a long LOS. The Least Absolute Shrinkage and Selection Operator (LASSO) regression method was employed to screen variables and construct a risk model for long LOS. A multiple logistic regression prediction model was then constructed using the selected variables from the LASSO regression model. The significance of the prediction model was evaluated by calculating the area under the ROC curve (AUC) and assessing the confidence interval around the AUC. The calibration curve is used to further validate the model's calibration and predictability. The model's clinical effectiveness was assessed through decision curve analysis. To evaluate the generalizability of the model, fivefold cross-validation was employed. Internal validation of the models was performed using bootstrap validation. Among the 539 infants who received blood transfusions, 398 infants (P75) had a length of stay (LOS) within the normal range of 34 days, according to the interquartile range. However, 141 infants (P75) experienced long LOS beyond the normal range. The predictive model included six variables: gestational age (GA) (< 28 weeks), birth weight (BW) (< 1000 g), type of respiratory support, umbilical venous catheter (UVC), sepsis, and resuscitation frequency. The area under the receiver operating characteristic (ROC) curve (AUC) for the training set was 0.851 (95% CI 0.805-0.891), and for the validation set, it was 0.859 (95% CI 0.789-0.920). Fivefold cross-validation indicates that the model has good generalization ability. The calibration curve demonstrated a strong correlation between the predicted risk and the observed actual risk, indicating good consistency. When the intervention threshold was set at 2%, the decision curve analysis indicated that the model had greater clinical utility. The results of our study have led to the development of a novel nomogram that can assist clinicians in predicting the probability of long hospitalization in blood transfused infants with reasonable accuracy. Our findings indicate that GA (< 28 weeks), BW(< 1000 g), type of respiratory support, UVC, sepsis, and resuscitation frequency are associated with a higher likelihood of extended hospital stays among newborns who have received blood transfusions.

Extracellular Vesicle-microRNAs as Diagnostic Biomarkers in Preterm Neonates

Neonates born prematurely (<37 weeks of gestation) are at a significantly increased risk of developing inflammatory conditions associated with high mortality rates, including necrotizing enterocolitis, bronchopulmonary dysplasia, and hypoxic-ischemic brain damage. Recently, research has focused on characterizing the content of extracellular vesicles (EVs), particularly microRNAs (miRNAs), for diagnostic use. Here, we describe the most recent work on EVs-miRNAs biomarkers discovery for conditions that commonly affect premature neonates.

Risk factors for length of NICU stay of newborns: A systematic review

Background

The improvement in survival of preterm infants is accompanied by an increase in neonatal intensive care unit (NICU) admissions. Prolonged length of stay in the NICU (LOS-NICU) increases the incidence of neonatal complications and even mortality and places a significant economic burden on families and strain on healthcare systems. This review aims to identify risk factors influencing LOS-NICU of newborns and to provide a basis for interventions to shorten LOS-NICU and avoid prolonged LOS-NICU.

Methods

A systematic literature search was conducted in PubMed, Web of Science, Embase, and Cochrane library for studies that were published in English from January 1994 to October 2022. The PRISMA guidelines were followed in all phases of this systematic review. The Quality in Prognostic Studies (QUIPS) tool was used to assess methodological quality.

Results

Twenty-three studies were included, 5 of which were of high quality and 18 of moderate quality, with no low-quality literature. The studies reported 58 possible risk factors in six broad categories (inherent factors; antenatal treatment and maternal factors; diseases and adverse conditions of the newborn; treatment of the newborn; clinical scores and laboratory indicators; organizational factors).

Conclusions

We identified several of the most critical risk factors affecting LOS-NICU, including birth weight, gestational age, sepsis, necrotizing enterocolitis, bronchopulmonary dysplasia, and retinopathy of prematurity. As only a few high-quality studies are available at present, well-designed and more extensive prospective studies investigating the risk factors affecting LOS-NICU are still needed in the future.

Predicting clinical outcomes using artificial intelligence and machine learning in neonatal intensive care units: a systematic review

BACKGROUND

Advances in technology, data availability, and analytics have helped improve quality of care in the neonatal intensive care unit.

OBJECTIVE

To provide an in-depth review of artificial intelligence (AI) and machine learning techniques being utilized to predict neonatal outcomes.

METHODS

The PRISMA protocol was followed that considered articles from established digital repositories. Included articles were categorized based on predictions of: (a) major neonatal morbidities such as sepsis, bronchopulmonary dysplasia, intraventricular hemorrhage, necrotizing enterocolitis, and retinopathy of prematurity; (b) mortality; and (c) length of stay.

RESULTS

A total of 366 studies were considered; 68 studies were eligible for inclusion in the review. The current set of predictor models are primarily built on supervised learning and mostly used regression models built on retrospective data.

CONCLUSION

With the availability of EMR data and data-sharing of NICU outcomes across neonatal research networks, machine learning algorithms have shown breakthrough performance in predicting neonatal disease.