AI Personalizes Pediatric Care and Improves Visits for Rural Families at Valley Children's

Valley Children's uses AI to ease visits, cut paperwork, and connect to rare-disease experts while protecting privacy. Genomic-guided dosing and smarter intake boost safer, faster care.

Categorized in: AI News Healthcare
Published on: Sep 23, 2025
AI Personalizes Pediatric Care and Improves Visits for Rural Families at Valley Children's

How Valley Children's Hospital is using AI to improve pediatric care and access

Valley Children's Hospital serves 1.3 million kids across California's Central Valley. The largely rural service area makes access a constant challenge. Under the leadership of Dr. Michael Scahill, medical director of physician informatics and digital health, the organization has been deploying AI to streamline visits, support complex decisions, and personalize treatment.

The result: a better office experience for families, lighter documentation load for clinicians, and faster access to expertise for rare conditions-without compromising patient privacy.

Ambient documentation resets the visit

Ambient scribe technology runs in the background during appointments, so clinicians can focus on patients instead of keyboards. Physicians report they "absolutely love it," and families notice the difference. As Dr. Scahill puts it, "Nobody likes seeing a doctor whose back is to them while they're typing."

Early pilots show strong physician uptake. The hospital is also piloting ambient tools for nursing documentation, starting with vitals, to reduce manual workload and improve throughput.

Data-scarce pediatrics: collaborate to move faster

Pediatrics represents about 5% of U.S. hospitalizations, which limits training data and commercial investment for AI. That slows clinical AI development. It also pushes teams to collaborate across institutions to find the best answers for rare disease care.

Finding rare-disease expertise with Epic Cosmos

Valley Children's uses Epic Cosmos to locate providers who see specific rare cases-without exposing patient identities. Example: identifying clinicians in other states who routinely treat phenylketonuria (PKU) to discuss rare complications and treatment approaches.

This shortens the path to the right consultation. It turns nationwide EHR data into a practical directory of experience.

Pharmacogenomics at the point of care

With thousands of medications and countless gene-drug interactions, manual review isn't realistic. The team analyzes genomic data to guide therapy selection and dose. Automation routes whole-genome sequencing results into decision support that alerts prescribers: "Use a higher dose here, lower there," or avoid a drug due to intolerance.

Clinical teams align these alerts with recognized standards such as the CPIC pharmacogenomics guidelines. The aim is simple: safer, more effective medications from the start.

Smarter intake, less friction

The hospital is working with Epic on AI-driven patient histories. Instead of families checking boxes on developmental forms, AI handles structured intake questions and populates the record. That reduces busywork and improves data quality before the clinician walks in.

Access: the Central Valley mandate

Recruiting physicians remains tough in a region without dense academic hubs. Dr. Scahill's goal is parity: "Kids here deserve the same care that kids in Palo Alto get." AI is one of the tools helping stretch clinical capacity and expertise across a wide geography.

What healthcare leaders can do now

  • Run a focused ambient scribe pilot (4-6 weeks). Track note time, visit length, physician satisfaction, and patient feedback. Establish clear consent scripts and fallback workflows.
  • Use federated EHR analytics to connect clinicians on rare cases (e.g., Cosmos). Create a standard process for outreach and documentation of external input.
  • Operationalize pharmacogenomics. Route lab results into clinical decision support, surface gene-drug alerts at order entry, and align with CPIC guidance.
  • Automate the routine. Start with nursing vitals and patient history intake. Measure accuracy, time saved per shift, and impact on throughput.
  • Strengthen governance. Keep humans in the loop, validate models locally, and monitor bias, privacy, and drift. Audit alerts for signal-to-noise.
  • Upskill clinicians on AI literacy and workflow design. For role-based options, see AI courses by job.

The takeaway: pick use cases that reduce friction, protect time, and connect your clinicians to expertise they don't have locally. Valley Children's shows that with targeted AI, rural pediatric care can be more personal, more precise, and more accessible.