healthsystemCIO – Strategies for Hospital IT Leaders

Jim Bilsky, SVP and CTO, and Ryan Hooley, VP of Enterprise Architecture and AI, at Baptist Health in Jacksonville, Fla., discuss how they brought Epic, Microsoft, IGEL, ControlUp and Imprivata into one room to speed up virtual desktop sign-in. They explain the log-level evidence their architects gathered, why physicians told the vendors what slow logins cost, and how sign-in fell from about 90 seconds to 10 to 15.

Chris Carmody, senior vice president and CIO at UPMC, explains how the health system weighs AI solutions and agents against their return and is prepared to turn off those that fall short. Carmody discusses token costs, the policy and review group behind AI adoption, and the pilot of 1,000 physicians per vendor that chose Abridge over Nuance DAX, along with why UPMC keeps that ambient AI decision open.

Stephanie Hines, senior vice president and CIO at Valleywise Health, discusses why the operators accountable for results should own technology decisions and their risk, with IT acting as steward of a central budget. Hines explains how a decision science framework lays out the cost, cyber risk and integration burden of point solutions, and how look-backs make IT's value visible without asking for credit.

Alex Forman, chief data officer at WVU Medicine, explains why the health system standardized how its hospitals build and measure OR block time before turning on Epic's AI-assisted scheduling tools. Forman discusses treating block utilization as a stewardship measure, keeping one version of the numbers when vendors bring their own analytics, and building data models for AI agents. Main-OR block utilization rose from about 60% to 67%.

Naomi Lenane, CIO and senior vice president of information services at Dana-Farber Cancer Institute, discusses how much data quality a health system needs before acting on its data. Lenane explains why the answer depends on the decision, why data definitions and ownership come before cleanup, how related measures show whether a report deserves trust and what it takes for clinicians to trust AI output built on imperfect data.

Angie Costakis, COO and CIO at Claiborne Memorial Medical Center, discusses how a 47-bed public hospital in rural Louisiana finds money for IT. Costakis explains why grant searches start from the hospital's needs list, where cybersecurity funding comes from in Louisiana, what the first round of Rural Health Transformation awards brought and how the hospital budgets to keep grant-funded systems running.

Jason Joseph, chief digital and transformation officer at Corewell Health, discusses who should set the pace of AI adoption and who is accountable when AI is wrong. Joseph explains why he treats value as the gas and risk as the brake, why most high-value use cases need automation that current safeguards cannot yet support, and why a reviewer’s attention fades as AI improves.

Russ Branzell, president and CEO of CHIME, explains why redesigning how care is delivered has become the central job of health system technology leaders. He describes chief digital leaders' estimates that 30% to 60% of hospital work could be automated, a physician practice rebuilt around ambient AI, and why economic strain and workforce shortages are pushing small and midsize hospitals to change radically or risk closing.

Rob Adamson, EVP and chief information officer at RWJBarnabas Health, discusses how his organization is preparing for AI consumption pricing. He explains the 90-day value test a model must pass to keep its compute, why a volume model priced discharge summarization at about $312,000 a year, and how routing requests to the least sophisticated capable model and running workloads locally can hold AI costs down.

Amy Trainor, RN, SVP and chief information officer at Ochsner Health, discusses why about 4,500 patients connecting their records to ChatGPT have put a near-term decision in front of the 47-hospital system: whether to expand Epic's Emmie assistant in MyChart. She also explains why knowledge management rises in priority once AI starts summarizing charts, and raises the sponsorship and consent questions she wants the industry to discuss.