Artificial intelligence (AI) is used to make everyday life easier. Here’s how AI systems can help elevate your medical game.
AI in medicine
Speech recognition / dictation software is an example of AI currently used in pediatric practice. Today, Nuance Communication’s most popular medical speech recognition software, Dragon Medical One, has 300,000 words of vocabulary and integrates vocabularies from 90 medical specialties. By integrating deep learning (DL), the software covers the nuances of the user’s speech patterns and improves over time with an accuracy of 99%.
Over the past few years, Nuance has developed and refined a powerful AI virtual writing system called Dragon Ambient experience (DAX). Using a smartphone alone or with a wall-mounted camera and microphone system, the AI system records the entire patient visit and uploads the information to the cloud. The information is analyzed and a full note is created which is reviewed by Nuance technicians before going for review and signature is inserted into the electronic patient record (EHR) by the doctor. The entire process takes 4 hours or less. DAX is available to specialists and general practitioners, including pediatricians, in outpatient departments. It will soon be rolled out to hospitals, and the system will be used by nurses and others who care for patients. Microsoft helped Nuance develop this system and recently acquired the company, apparently to expand DAX for other purposes in patient care and adapt health and beyond.
I had my first experience with AI in pediatrics in 2010 when I wrote about digital stethoscopes and the Cardioscan program from what was then Zargis Medical Corporation. The program uses DL to analyze recorded heart sounds to identify sounds that should be examined with an echocardiogram (ECG). This was known as computer-assisted auscultation and the Cardioscan was much better than pediatricians at detecting potentially pathological sounds. Now digital stethoscopes, including those sold by Eko, can record and interpret ECGs and sounds. Certain AI-powered systems can also screen adults for diabetic retinopathy, screening premature infants for premature retinopathy, examining images of pigmented nevi for elements indicative of melanoma, and determining a child’s bone age and helping doctors review and interpret imaging studies and EEGs.
AI and COVID-19
AI played a huge role in the COVID19 pandemic; for example, various AI systems accurately predicted the SARSCoV2 outbreak before the World Health Organization did. In addition, the AI analyzed the data from the outbreak to predict where and when the next spike in COVID19 cases would occur. These types of systems can review chest CT images to diagnose COVID19 and accurately predict which patients are at risk for serious illness, and provide management and early hospitalization opinions. Furthermore, AI-powered systems have played an important role in the development of COVID-19 vaccines and therapeutics that can be effective against SARSCoV2. Some states even use AI-based algorithms to prioritize patients for vaccines.
Clinical decision support
Several years ago, American pediatricians worked with pediatricians in China to extract information from the EHR of the Guangzhou Women’s and Children’s Medical Center to develop a clinical decision support system (CDSS) tool. A total of 101.6 million data points were extracted from 1,362,559 EHRs from free text EHR notes using natural language processing algorithms. In testing, the study’s CDSS tool outperformed younger but non-senior pediatricians in diagnosing asthma, encephalitis, gastroenteritis, pneumonia, Sinusitis, upper respiratory infections, and psychiatric illnesses. Overall, it was able to make an accurate diagnosis in 90% of the cases and no worse than in 79% of the cases. There is not yet a CDSS released for general pediatric use, but when available, these tools help doctors determine the most likely diagnosis, order the cheapest tests, and prescribe the cheapest antibiotics while reducing medical errors.
Conclusion
Cognitive computing has the potential to improve pediatric practice and computers are unlikely to become “confident” and start competing with pediatricians for patients. Pediatricians need to be open to adopting AI technologies that will improve care and reduce the discomfort associated with pediatric practice.
