How City of Hope Is Bringing Data Science Into the Clinical Workflow

Summary

City of Hope's Applied AI and Data Science team uses Posit Connect as the central hub for deploying tools across clinical and business operations, from comorbidity documentation to a tool that accelerates clinical trial feasibility by identifying eligible patient cohorts and assessing site readiness using a single, AI‑enabled front end. Now they're working with Posit to embed those tools directly inside Epic, so clinicians can access data science output without ever leaving the workflow where they actually work.

Spotlight PDF

About:

City of Hope is one of the largest and most advanced cancer research and treatment organizations in the United States, and one of the leading research centers for diabetes and other life-threatening illnesses. Its mission is to make hope a reality for all touched by cancer and diabetes. City of Hope’s leading centers and outpatient sites are in the major metropolitan areas in Southern California, Arizona, Illinois, and Georgia, making City of Hope accessible to more than 86 million Americans.es.

Industry:

Healthcare

Technology used:

Posit Connect

The challenge

Production Data Science in a Clinical Environment

Deploying data science in healthcare means more than building a good model. The tools that Matt Maloney, Director of Applied AI and Data Science, and his team build at City of Hope end up in clinical workflows, billing systems, and executive dashboards, and getting something deployed, maintained, and used where the real work happens.

Maloney's team sits within the Enterprise Technology Group (ETG) and owns more of the production stack than most data science groups do. As the scope of their work expanded from statistics and ML to DevOps, LLM frameworks, and AI engineering, the team needed a deployment platform that could keep pace, without every new project becoming a handoff to software engineering and without data scientists working outside their domain expertise.

The traditional handoff between teams can slow progress, add coordination complexity, and increase the risk that important context is lost during transitions. The data scientist who built the model understands its nuance in ways that don't always translate cleanly once the work moves into separate deployment workflows. When issues arise, such as drift or upstream changes, it can be harder to diagnose and resolve them quickly if visibility into how the model is deployed is split across teams. What Maloney needed was a way for his team to maintain clear, end-to-end visibility from development through deployment, while staying focused on the work data scientists do best. Additionally, deploying production-level data science work required coordinating across multiple teams (server provisioning, networking, storage, and security) each operating on its own timelines and processes. As a result, the turnaround time from a ready data science product to production deployment could stretch into weeks, even when the underlying work was complete.

The solution

Posit Connect as the Hub

Posit Connect gave the City of Hope data science team a place to iterate and deploy. Shiny applications, scheduled scripts, API endpoints, and automated reports all live on Connect, accessible to internal stakeholders across clinical and business functions without a handoff in between.

The use cases span a wide range:

A tool for clinical coders helps identify gaps in comorbidity documentation, improving documentation completeness and the accuracy of information.
Automated email notifications cross-reference lab evidence with clinical notes to flag potential documentation issues before they become billing problems.
A clinical trial feasibility assessment tool, built in collaboration with the research side of the organization, has evaluated over 200 trials across City of Hope’s sites in nine months.
On the business side, Posit Connect hosts applications supporting revenue cycle work, philanthropy, and financial forecasting, including charge capture for account review and enterprise cash forecasting. These tools cumulatively review nearly 1,000 notes for over 100 patient accounts daily.
Streamline governance through City of Hope’s AI Governance Committee that reviews every use case for HIPAA compliance, privacy, and ethical considerations before anything goes live.
Other teams around City of Hope, including Clinical Finance, Biostatistics, and others, also use the platform for analytics workflows.

The next frontier

Embedding Posit Inside Epic

The most ambitious integration City of Hope is working with Posit on right now is to embed Shiny applications hosted on Posit Connect inside Epic, their electronic health record system, so that clinicians never have to leave the environment where they spend their day.

The goal is straightforward: when a physician opens a patient record, the tool they need is already there, pre-populated with that patient's context. No separate window, no manual entry, no context switch. In a clinical setting where time in any patient interaction is already constrained, an extra step can mean a tool simply doesn't get used.

The integration will work by passing patient context, like a medical record number, directly from Epic into the Shiny app. For the City of Hope team, that means:

Tools built by data scientists can reach clinicians inside the workflow where decisions get made
Patient context loads automatically, removing manual steps and number of clicks for already time-pressed users
The data science team can continue owning and iterating on the tools with fewer moving technical pieces to manage.