Hospitals across the nation are using artificial intelligence (AI) tools to improve health outcomes and patient experiences. While AI offers important insights for medical professionals, these applications have unique implications for pediatrics.
Opportunities for pediatric care
AI has the potential to help providers produce better health outcomes and provide more efficient care delivery for pediatric patients. Various AI tools can streamline administrative tasks, improve diagnostic accuracy, support personalized treatments, and more. Examples of AI use for children include:
- Early diagnosis and predictive analytics: AI quickly analyzes large datasets from electronic health records to identify patterns and predict the onset of disease. This supports clinicians in making earlier, more accurate diagnosis. Highly sensitive machine learning is a great tool for screening and early disease detection in children, which can help prevent disease progression later on in life.
- Imaging and diagnostics: AI tools can improve diagnostic accuracy in pediatric imaging, such as reading X-rays and MRIs. It can also help identify fractures, tumors, and pneumonia, reducing the need for repeated imagery and minimizing young patients’ exposure to radiation.
- Ambient listening: AI can allow health care providers to spend more time focused on patient-facing care with children and families and reduce “pajama time” tasks, or work providers complete at home. Ambient listening tools that record patient visits and create notes for providers is an example of how this can be achieved.
- Administrative tasks: AI streamlines administrative tasks by automating medical coding and billing, scheduling and appointment management, processing claims, managing medical records, and processing administrative documents. This makes administrative processes at children’s hospitals more efficient, improving patient health care delivery.
- Virtual health assistants: AI chatbots and virtual health assistants provide patients with personalized health information, answer families’ medical questions, and offer guidance on managing chronic conditions. These tools enhance patient engagement and enable pediatric providers to deliver more accessible and efficient care.
- Medical education and training: AI-powered simulators and virtual reality platforms provide pediatric health care professionals with realistic training scenarios and feedback, enhancing their clinical skills and knowledge in a safe environment.
Unique considerations for children
AI in health care poses unique considerations for children. These span across data availability, data use consent, data complexity, bias, and others.
Lack of children’s data to inform AI tools: AI models must be trained with existing information and large data sets to produce accurate outputs. Some pediatric conditions have few samples or data points, making this data difficult to obtain. Data experts and researchers at children’s hospitals are working to make data sets interpretable for computerized learning.
Complexity in child medical data: Pediatric medical data has more input than adult medical data. Most adult medical files include notes from the patient and the clinician. However, children’s medical data includes various inputs that can include providers, parents or caregivers, numerous subspecialists, teachers, and more. Merging these data sources increases the difficulty of obtaining accurate predictions.
Challenges in receiving consent: Consent to work with pediatric data is usually obtained first via parental agreement. Approval for use of a child’s health data is reobtained when children reach the age of consent. This requires strict monitoring over consent status and can limit the use of this data to inform AI tools.
Innovative solutions at children’s hospitals
AI technology and tools created for health care are generally tailored to adult care. To combat these challenges, children’s hospitals have leveraged their expertise and invested in efforts to ensure that AI can improve patient care for children.
Reducing epilepsy surgery referral times: Cincinnati Children’s is using AI and machine learning to reduce neurosurgery epilepsy surgery referral times. With the goal of decreasing the current time of six years from diagnosis to surgical referral, Cincinnati Children’s has trained AI to capture electronic health record data and alert physicians when a patient is eligible to be reviewed by the surgical committee.
Predicting neurodevelopmental deficits in infants: Comer Children’s Hospital at the University of Chicago Medical Center has utilized a virtual model of an infant’s gut microbiome so providers can track how microbiome development affects development deficits. This has helped providers predict infants at higher risk of neurodevelopmental deficits and facilitate earlier treatment, leading to improved health outcomes.
Detecting patient deterioration: Akron Children’s Hospital uses an AI-powered tool to produce a deterioration index score indicating a patient’s risk factor for deterioration and illustrating the patient’s trendline over time. Care providers can see developing trends and use the index scores to validate their own judgments when they feel something is not right with a patient.
Improving pediatric heart transplants: Pediatric heart transplants are one of the most challenging surgeries to coordinate. A team from Cincinnati Children’s is using AI to more effectively match heart donors with recipients. A deep-learning model can examine cardiac images in seconds, accomplishing what takes a cardiologist 30 minutes.
Learn more about how AI is being used in pediatric care.
