Close

New Workspace Modernization Research from CDW

See how IT leaders are tackling workspace modernization opportunities and challenges.

Nov 06 2025
Artificial Intelligence

AI in Pediatric Care Brings Challenges and New Efficiencies

Pediatric hospitals are using artificial intelligence to document patient encounters to reduce burnout and develop models for radiology.

From ambient scribes improving documentation to the automation of coding, billing and revenue cycle operations, pediatric hospitals are using artificial intelligence for numerous clinical and nonclinical uses.

AI helps predict patient deterioration and enhances situational awareness in clinical areas such as the intensive care unit, says Dr. Bimal Desai, vice president and chief health informatics officer at Children’s Hospital of Philadelphia (CHOP).

For pediatrics, AI presents challenges in making clinical predictions and diagnoses as kids grow and change over time, Desai says.

“It requires a larger, more diverse set of training data to faithfully represent all of the ages and stages of child development,” he says. 

He notes that fractures in young children would appear different from those of older adolescents or adults. Therefore, training AI to identify these “subtle” fractures presents a challenge.

“Similarly, the clinical patterns that describe deterioration in an adult patient look different than in a child,” Desai says.

Click the banner below to read the new CDW Artificial Intelligence Research Report.

 

Risks and Challenges of Using AI in Pediatrics

Large language models are prone to “sycophancy,” or the desire to satisfy the user prompting it, as well as hallucinations, which involve fabricated findings, Desai explains.

He notes that using AI in pediatrics also presents challenges because many children’s diseases are rare and therefore more difficult for algorithms to predict.

“This is a pure biostatistical limitation: The performance of an algorithm in real life depends heavily on the prevalence of the condition. The more common the condition, the more accurate the algorithm,” Desai explains. “Conversely, the rarer the condition, the worse the algorithm will perform.”

Although scientists use machine learning algorithms to identify rare pediatric conditions such as primary immune deficiencies, they still generate a large amount of “fake positive” results, according to Desai. Children then participate in lots of unnecessary and expensive testing, he says.

READ MORE: AI is being used in healthcare in several different ways today.

“That makes these algorithms very difficult to implement at the point of care,” Desai says.

In addition, training biases bring risks in pediatrics.

“Some groups of patients were perhaps underrepresented in the training data, so the model’s predictions about those patients are inaccurate,” Desai says.

To manage the risks of AI, Texas Children’s has established an AI governance and steering committee, says Teresa Tonthat, the hospital’s vice president and associate CIO.

The governance committee ensures all AI model outcomes require a “human in the middle” to verify information before making patient decisions, Tonthat says. The committee also addresses concerns around regulatory requirements, bias and AI hallucinations, she adds. 

Because the health system is working with data about children, Texas Children’s provides education to care teams about signoffs through Epic’s MyChart and works with vendors such as Microsoft to discuss how they leverage and protect patient data, Tonthat says.

“Our risk tolerance is very low when it comes to information about our pediatric patients,” Tonthat says.

Despite the risks, AI brings opportunities for solving clinical issues and improving efficiency in workflows.

AI for Radiology and Diagnostics in Pediatrics

Texas Children’s has been using AI for more than decade and has turned to predictive modeling, automation, deep learning and machine learning to solve clinical problems, Tonthat says. The hospital has been exploring generative AI for the past two years, particularly with models that improve care team workflows. Interest took time to build among the clinical teams at Texas Children’s, but now the staff is ready to scale up, she says.

Texas Children’s created an AI model for radiologists to predict bone age. The model was trained on millions of X-rays of pediatric hands.

“Because we have millions of X-rays of hand images, we have trained the model to let them know within seconds what the age of that child’s hand is from a bone-density perspective,” Tonthat says.

Using its bone age AI-based prediction model, Texas Children’s was able to improve turnaround time by 50% by integrating AI into radiologists’ clinical workflows, according to Tonthat. The AI model is a collaboration between the hospital’s radiology and information services departments and, its AI governance and guidance committee.

CHOP is also using AI in radiology, according to Desai.

“In research, investigators at CHOP are looking at AI for improved radiology diagnostics, intercepting lab errors, accelerating diagnosis from pathology slides and more,” he says.

If a doctor is seeing a patient with poorly controlled asthma, an ambient AI tool could automatically retrieve the patient’s asthma history, summarize past asthma-related visits and notify a doctor if the patient is at higher risk for influenza, Desai says. It then would check to see which medication the patient’s insurance plan covers and could start composing the order for the asthma controller, Desai says.

Dr. Bimal Desai
Imagine AI systems that can anticipate my information needs in real time, surface the information I need to complete a task automatically, summarize and synthesize complex health data, and so on.”

Dr. Bimal Desai Vice President and Chief Health Informatics Officer, Children’s Hospital of Philadelphia

Ambient Listening Can Reduce Clinician Burnout at Pediatric Hospitals

Because physicians spend close to half their time handling tasks in electronic health records (EHRs)  or performing desk work in an ambulatory setting, they take work home with them at night, says Desai.

Instead, ambient listening allows clinicians to capture notes in real time during patient interactions and could eliminate note-taking, Desai suggests. The notes are completed within minutes of a patient encounter.

“While results suggest not all physicians benefit equally, the national data suggests that these tools can reduce physician burnout, improve turnaround time for note completion and chart closure, and improve patient experience,” Desai says.

A 2024 American Medical Association survey found rates of burnout between practitioners of pediatrics and internal medicine were “similar,” at about 42%.

“Just giving providers an hour or two back in their day would make a massive difference,” Desai says. “Imagine AI systems that can anticipate my information needs in real time, surface the information I need to complete a task automatically, summarize and synthesize complex health data, and so on. That’s where we’ll see clinical AI have its first and most significant impact after the ambient use cases.”

DISCOVER: Minimum viable data governance enables smarter healthcare.

Special Considerations for AI in Pediatric Care

When using AI tools on pediatric data, providers must secure consent, Tonthat notes. She says that most kids are not old enough to provide consent on the use of generative AI to help with their diagnosis or treatment, so their parents or guardians would provide this consent.

Pediatric hospitals such as CHOP require fine-tuning AI tools because the data is different from other organizations.

“Tools might not work out of the box for pediatrics,” Desai says. “As an example, the idea of an AI that can draft portal message replies is great, but it must also honor state adolescent privacy laws in the process.”

AI tools should not automatically let parents get a refill of their child’s medications, according to Desai.

“If a parent asks, ‘Can I get a refill for my child’s medications,’ the AI should not include any mention of medications the teen obtained during a confidential visit,” Desai says. “The only way to identify and address these pediatric-specific issues is to have pediatric institutions participate in testing, continue to evaluate and share our lessons learned, and give feedback and clear guidance to AI developers.”

Texas Children’s is piloting an AI project in which parents and patients email and chat with physicians through MyChart with simple questions, and an AI model reviews the EHR to draft responses for physicians and nurses, Tonthat says.

“That helps tremendously with the amount of time our care teams take after or between clinics to go review emails and respond,” Tonthat says. “It even integrates empathy levels and looks at chart information.”

Click the banner below to sign up for HealthTech’s weekly newsletter.

 

Preparing for AI in Pediatrics

Health systems should train physicians and senior leaders on what AI can realistically do and collaborate on what problems they could solve using it, Tonthat says.

Texas Children’s also provides education to staff on data literacy and readiness.

“We want to make sure that we identify our problem and then create a model to address that problem with clean, high-quality, high-efficacy data,” Tonthat says. 

Texas Children’s adopted the 80/20 rule with AI. That means 80% of the initiative involved reviewing problem statements and data cleansing, and 20% involved building and testing the model and performing change management, Tonthat adds.

Regarding children, she says, “They may not be equipped to protect their own data, and we have full access to very critical data sets for them, not only protected health information but with personal identifiable information as well.”

She adds, “If that gets in the hands of bad actors or unauthorized users, when that child turns 18, they can be in a world of chaos if we aren’t doing everything we can to safeguard that data on their behalf.”

Pekic/Getty Images