Pre-Consultation System Based on the Artificial Intelligence Has a Better Diagnostic Performance Than the Physicians in the Outpatient Department of Pediatrics

Abstract

Artificial intelligence (AI) has been deeply applied in the medical field and has shown broad application prospects. Pre-consultation system is an important supplement to the traditional face-to-face consultation. The combination of the AI and the pre-consultation system can help to raise the efficiency of the clinical work. However, it is still challenging for the AI to analyze and process the complicated electronic health record (EHR) data. Our pre-consultation system uses an automated natural language processing (NLP) system to communicate with the patients through the mobile terminals, applying the deep learning (DL) techniques to extract the symptomatic information, and finally outputs the structured electronic medical records. From November 2019 to May 2020, a total of 2,648 pediatric patients used our model to provide their medical history and get the primary diagnosis before visiting the physicians in the outpatient department of the Shanghai Children's Medical Center. Our task is to evaluate the ability of the AI and doctors to obtain the primary diagnosis and to analyze the effect of the consistency between the medical history described by our model and the physicians on the diagnostic performance. The results showed that if we do not consider whether the medical history recorded by the AI and doctors was consistent or not, our model performed worse compared to the physicians and had a lower average F1 score (0.825 vs. 0.912). However, when the chief complaint or the history of present illness described by the AI and doctors was consistent, our model had a higher average F1 score and was closer to the doctors. Finally, when the AI had the same diagnostic conditions with doctors, our model achieved a higher average F1 score (0.931) compared to the physicians (0.92). This study demonstrated that our model could obtain a more structured medical history and had a good diagnostic logic, which would help to improve the diagnostic accuracy of the outpatient doctors and reduce the misdiagnosis and missed diagnosis. But, our model still needs a good deal of training to obtain more accurate symptomatic information.

Personalized anesthesia and precision medicine: a comprehensive review of genetic factors, artificial intelligence, and patient-specific factors

“…AI is utilized for evidence-based clinical decision support ( 112 ), detecting adverse events, and using electronic health record (EHR) data to predict patients at risk of readmission ( 113 ). By accessing EHR data, AI has demonstrated potential to surpass physicians in diagnostic accuracy ( 114 – 117 ). Algorithms that combine imaging and EHR data with relevant medical records can predict malignancy on biopsy and differentiate between normal and abnormal screening results, significantly reducing missed breast cancer diagnoses ( 118 ).…”

Comparative survey among paediatricians, nurses and health information technicians on ethics implementation knowledge of and attitude towards social experiments based on medical artificial intelligence at children’s hospitals in Shanghai: a cross-sectional study

“…8 9 Robots can also be employed for pre-consultation, triage and referral services for children, further expanding the scope of AI implementation in paediatric healthcare. 10 Therefore, implementing AI in paediatric healthcare in China is indeed a pressing need.…”

Using generative AI to investigate medical imagery models and datasets

“…Many machine learning models have performed as well as highly trained physicians, 1 , 2 , 3 while others have reportedly achieved higher performance than their human counterparts. 4 , 5 , 6 While many of these prediction tasks are based on reproducing tasks that human experts are capable of, several AI models surprisingly can identify diseases and patient characteristics far beyond what the imaging modality was known to reveal. 7 , 8 , 9 In some cases this is due to the model picking up on confounders, 10 selection/information bias in dataset selection 11 , 12 or structural, cultural, and historic biases reflected in data 13 , 14 ; whereas in others the model is picking up on previously unknown manifestations of a condition.…”