Derivation and internal validation of a data-driven prediction model to guide frontline health workers in triaging children under-five in Nairobi, Kenya

Abstract

Background: Many hospitalized children in developing countries die from infectious diseases. Early recognition of those who are critically ill coupled with timely treatment can prevent many deaths. A data-driven, electronic triage system to assist frontline health workers in categorizing illness severity is lacking. This study aimed to develop a data-driven parsimonious triage algorithm for children under five years of age. Methods: This was a prospective observational study of children under-five years of age presenting to the outpatient department of Mbagathi Hospital in Nairobi, Kenya between January and June 2018. A study nurse examined participants and recorded history and clinical signs and symptoms using a mobile device with an attached low-cost pulse oximeter sensor. The need for hospital admission was determined independently by the facility clinician and used as the primary outcome in a logistic predictive model. We focused on the selection of variables that could be quickly and easily assessed by low skilled health workers. Results: The admission rate (for more than 24 hours) was 12% (N=138/1,132). We identified an eight-predictor logistic regression model including continuous variables of weight, mid-upper arm circumference, temperature, pulse rate, and transformed oxygen saturation, combined with dichotomous signs of difficulty breathing, lethargy, and inability to drink or breastfeed. This model predicts overnight hospital admission with an area under the receiver operating characteristic curve of 0.88 (95% CI 0.82 to 0.94). Low- and high-risk thresholds of 5% and 25%, respectively were selected to categorize participants into three triage groups for implementation. Conclusion: A logistic regression model comprised of eight easily understood variables may be useful for triage of children under the age of five based on the probability of need for admission. This model could be used by frontline workers with limited skills in assessing children. External validation is needed before adoption in clinical practice.

Prevalence, outcome and factors associated with dysglycemia among critically ill children presenting to Fort Portal Regional Referral Hospital: A cross sectional study

“…However, there are very few studies that have used this emergency sign at triage. A prospective observational study by Mawji et al in Kenya 2018 found obstructed breathing was not statistically significant in children aged two to 60 months seeking treatment for an acute illness (46). Nevertheless, they recruited critically ill children below the age of five years unlike in this study where study participants were recruited every day and included critically ill children from one month to twelve years.…”

Using self-supervised feature learning to improve the use of pulse oximeter signals to predict paediatric hospitalization

“…The study nurse followed up on patients who were not admitted via telephone calls approximately two weeks after the hospital visit to determine if the patient recovered from the ailment without required admission later or in a different facility. More details about study procedures are described elsewhere ( Mawji et al , 2021 ).…”

“…Features extracted using SSL were used to predict hospitalization using logistic regression. We also evaluated the utility of extracted features in improving logistic regression models trained using eight clinical features identified from previous analysis ( Mawji et al , 2021 ). The eight features include weight, restlessness, mid-upper arm circumference (MUAC), ability to drink/breastfeed, SpO 2 , temperature, difficulty breathing, and heart rate.…”

Using self-supervised feature learning to improve the use of pulse oximeter signals to predict paediatric hospitalization

“…We compared end-to-end deep learning models initialized randomly with those initialized using weights learned by contrastive learning models. The models were trained on soft labels (label smoothing) recommended for noisy targets, and the minority class was oversampled to address class imbalance (Müller et al, 2019;Szegedy et al, 2015). We tuned hyperparameters for regularization (dropout, weight decay, data augmentation), batch size, learning rate, and label smoothing parameter using the ASHA algorithm.…”