ABSTRACT
Artificial intelligence (AI)-based methods have emerged as powerful tools to transform medical care. Although machine learning classifiers (MLCs) have already demonstrated strong performance in image-based diagnoses, analysis of diverse and massive electronic health record (EHR) data remains challenging. Here, we show that MLCs can query EHRs in a manner similar to the hypothetico-deductive reasoning used by physicians and unearth associations that previous statistical methods have not found. Our model applies an automated natural language processing system using deep learning techniques to extract clinically relevant information from EHRs. In total, 101.6 million data points from 1,362,559 pediatric patient visits presenting to a major referral center were analyzed to train and validate the framework. Our model demonstrates high diagnostic accuracy across multiple organ systems and is comparable to experienced pediatricians in diagnosing common childhood diseases. Our study provides a proof of concept for implementing an AI-based system as a means to aid physicians in tackling large amounts of data, augmenting diagnostic evaluations, and to provide clinical decision support in cases of diagnostic uncertainty or complexity. Although this impact may be most evident in areas where healthcare providers are in relative shortage, the benefits of such an AI system are likely to be universal.
Subject(s)
Deep Learning , Diagnosis, Computer-Assisted , Electronic Health Records , Natural Language Processing , Pediatrics , Adolescent , Artificial Intelligence , Child , Child, Preschool , China , Female , Humans , Infant , Infant, Newborn , Machine Learning , Male , Proof of Concept Study , Reproducibility of Results , Retrospective StudiesABSTRACT
OBJECTIVE: In China, with the development of public health and medical treatment, accident became the first cause of death of children aged form 1 to 14 years, and poisoning became one of the main causes. The present study was conducted to investigate the efficacy of hemoperfusion (HP) on poisoning, and the pharmacokinetics of the poison during and after HP. The study was also to observe the effect of HP on blood cell and blood biochemistry, blood flow and dosage of heparin during HP in children. METHODS: Thirty-five children with acute poisoning (including 26 boys and 9 girls aged from 10 months to 13 years, mean 3.35 +/- 2.50 years) were treated with HP for one to three times. Among them 12 children were treated with HP for two times and 4 children for three times. Two ml blood samples of 6 children with Fluoroacetamide (FAM) poisoning and 10 children with Tetramine (TET) poisoning were collected. The concentration of poison was measured by gas chromatography (GC). RESULTS: The poisoning symptoms of all cases were relieved or alleviated obviously. In the end, 27 (77%) cases recovered and 6 (17%) cases improved, while 2 (6%) cases died of multi-organ failure (MOF). Clinical symptom happened again 6 - 24 hours after HP in 1 case with FAM poisoning and 3 cases with TET whose clinical symptoms were relieved during HP. The PLT, RBC counts and Hb decreased significantly after HP compared with pre-HP (P < 0.05), while WBC, alanine aminotransferase (ALT), aspartate aminotransferase (AST), reatine kinase (CK), blood urea nitrogen (BUN), creatinine (CRE), Creatine kinase isoenzymes (CK-MB), total protein (TP), albumin (ALB) and globulim (GLO) after HP did not significantly change (P > 0.05). The FAM concentration was significantly reduced (P < 0.030). The concentration of TET in the poisoned children also significantly decreased with the treatment (P = 0.001). The cleaning efficacy of HP was higher during the first hour than that during the second hour of HP. The concentration of poison rose again 2 - 6 hours after HP in 1 case with FAM poisoning and 3 cases with TET poisoning, but the level was lower compared with pre-HP level. The t(1/2) of FAM and TET with and without HP was (2.40 +/- 0.66) h, (15.60 +/- 8.22) h, (4.10 +/- 1.66) h and (67.01 +/- 48.42) h, respectively. The first dose of heparin was (0.54 +/- 0.15) mg/kg; then (0.20 +/- 0.06) mg/kg was added for every 30 minutes. The velocity of blood flow was (4.39 +/- 0.99) ml/min. CONCLUSIONS: The t(1/2) of the poison was shortened, and the poison clearing was accelerated by HP. The HP is a safe and effective therapy in children. The concentration of poison in some patient may rise again 2 to 6 hours after HP temporarily. The charcoal HP cannot remove the poison that conjugated with plasma albumin and globulin. The charcoal HP can cause temporary reduction of platelet and erythrocyte. The dosage of heparin used in children was lower than that in adult.