A team led by Dr. Emamaullee, Research Director in the Division of Abdominal Organ Transplantation, has developed the Children's Hospital Los Angeles Acute Liver Failure (CHALF) Score, a free, web-based application that is downloadable by smartphone. The CHALF Score predicts if a child experiencing acute liver failure will recover or should be referred to a transplant center. The study, of which Dr. Emamaullee is senior author, was published in the journal Transplantation.
Dr. Emamaullee’s group developed the CHALF score by constructing a machine learning-based model and training it on results from common tests that 147 pediatric acute liver failure patients treated at CHLA received in the emergency room or at hospital admission. The researchers sorted patients by demographics, diagnosis and laboratory results over the course of hospitalization and used statistical methods to arrive at the clinical tests and values that best predicted either the probability of the child surviving with their own liver or needing a transplant to prevent death from liver failure.
The team then validated their predictive model in a larger group of 492 similar patients in the multi-center National Institutes of Health (NIH) Pediatric Acute Liver Failure Study Group (PALFSG). The model was able to predict patient outcomes with high accuracy (0.83), outperforming the other two pediatric decision support tools (PELD and aLIU).
Using the output of the CHALF model, the team built an app they dubbed the CHALF score, which assesses liver failure risk on a scale of 5-60. A score above 30 predicts worse outcomes and should prompt urgent referral to a transplant center. A score under 30 indicates the probability that patients are likely to survive with their own liver.
The team has since used the CHALF score to assess five children at CHLA.