Application of Fine–Gray Competing Risk Modeling to Predict Pediatric Care Escalation Pathways in Southeast Ethiopia
This study utilized Fine–Gray competing risk modeling on 407 pediatric patients in Southeast Ethiopia to identify baseline malnutrition as the strongest predictor of care escalation and developed a validated integer-based risk score to facilitate early identification of high-risk children in low-resource settings.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the high-stakes environment of a hospital emergency room, time is the most precious resource, and the path a patient takes is rarely a straight line. When a child arrives in critical condition, doctors must make rapid decisions about whether to treat them in the emergency area, move them to a regular ward, or transfer them to intensive care. However, in many parts of the world, these decisions are complicated by a harsh reality: some children die before they can be moved to the next level of care. In statistical terms, this creates a "competing risk." If a child dies in the emergency room, they can no longer be escalated to a ward, meaning the event of death physically blocks the event of care escalation. Traditional ways of analyzing hospital data often treat death as if the patient simply left the study, which can hide the true reasons why other children get worse. Understanding how these different outcomes compete against one another is essential for building a system that saves lives, particularly in regions where resources are scarce and every minute counts.
Researchers in Southeast Ethiopia recently applied this advanced way of thinking to the pediatric emergency department at Asella Referral and Teaching Hospital. They looked back at the records of 407 children who arrived with acute medical conditions between 2022 and 2025. Their goal was to map out the specific journey of these children, distinguishing between those who were sent home, those who were moved to a ward or intensive care unit, and those who passed away before such a move was possible. By using a specialized statistical method that accounts for the fact that death prevents escalation, the team uncovered a clear picture of what drives a child's condition to worsen so rapidly that they need immediate, higher-level care.
The study revealed that the single strongest predictor of a child needing urgent escalation was not their age or where they lived, but their nutritional status. Children who arrived already suffering from malnutrition were more than eleven times more likely to require immediate transfer to a ward or intensive care compared to well-nourished children. This finding highlights that for many of these patients, the crisis began long before they reached the hospital door. Other significant factors included the child arriving in a state of true medical emergency, having a pre-existing health condition, showing signs of altered mental status, or having waited more than three days before seeking help. The data showed that the window for intervention is incredibly narrow; the median time from arrival to death was just 12 hours, with two-thirds of the deaths occurring within the first day. Similarly, one-third of the children who needed escalation were moved within the first 24 hours.
To turn these complex findings into a tool that doctors can use at the bedside, the researchers created a simple scoring system. They assigned points to the risk factors they identified, such as adding five points for severe malnutrition and four points for a true emergency presentation. A child's total score, which could range from zero to twenty-two, places them into one of three risk categories. A score of ten or higher indicates a high-risk child who needs immediate resuscitation and senior medical attention. The researchers tested this score and found it was highly accurate, with an AUC of 0.85 for identifying high-risk patients and an AUC of 0.80 for identifying low-risk patients. This tool allows medical staff to quickly sort patients, ensuring that the most vulnerable children receive the most intensive monitoring right from the moment they walk in.
The implications of this work extend beyond the walls of a single hospital. The study suggests that the "third delay"—the time between arriving at the hospital and receiving the right level of care—is heavily influenced by underlying structural issues like malnutrition and long travel times to the facility. Because the children who die do so so quickly, often within hours of arrival, the emergency room cannot function as a passive waiting area. Instead, it must act as an active filter that immediately identifies those with the deepest vulnerabilities. By integrating simple checks for malnutrition and using a straightforward risk score, hospitals in resource-limited settings can better allocate their limited staff and beds, potentially saving lives that might otherwise be lost in the critical first hours of admission.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.