Healthix CEO Todd Rogow on Rollout of AI Query Tool for HIE Data

Every aspect of healthcare is envisioning efficiency gains from AI, and health information exchanges are no different. In a recent interview, Todd Rogow, CEO at Healthix, an HIE that serves downstate New York, including New York City and Long Island, discussed his organization’s pilot of a generative AI capability from vendor partner InterSystems to assist clinicians in accessing and understanding patient information faster and more intuitively.

InterSystems’ HealthShare AI Assistant has a conversational interface that enables users to query, summarize, and navigate longitudinal health records using natural language prompts. 

Healthix collects data from more than 9,000 healthcare facilities, including 81 hospitals, for over 21 million patients. 

A usability study run with Healthix showed that the new tool improved clinician workflow, reduced time spent reviewing records, and surfaced useful and actionable insights. “We picked some of our power users across many of our large IDNs [integrated delivery networks] and other practice settings, and they used it over several months to provide feedback to InterSystems to help improve this new feature,” Rogow said. “Now they've deployed it in their general release, and we're in the midst of fully deploying it.”

The No. 1 problem Healthix is trying to solve is that every doctor doesn't want to see the same thing, especially on all the patients, Rogow said. “They’re looking for a lab test or an allergy or a diagnosis. We have tried to build that landing page of our clinical portal to encompass all the major things any doctor or nurse would want to see. But the AI Assistant allows this: Either verbally or by writing a prompt, you can say, ‘Can you summarize this patient’s last two ER visits dealing with a diabetic episode he had,' and it will go and fetch that information in five to 10 seconds, and create a summary.”

The AI Assistant also supports custom templates tailored to specific specialty workflows. Rogow said initially the users benefitted from a set of pre-selected examples of the prompts they could use, but he added that those were sort of like training wheels on a bicycle, and the clinicians don’t really need those anymore. 

Clinician feedback has highlighted several other benefits of using the AI Assistant. One is the generation of consolidated summaries, making it easier to understand patient histories at a glance. 

The clinicians also said that they appreciate that the assistant is able to detect medical conditions from contextual clues within the data. For example, a clinician is reviewing a patient’s recent visit notes, lab results, and medication history. The AI Assistant detects that the patient has been prescribed metformin and has elevated HbA1c levels across multiple visits. Although diabetes is not explicitly mentioned in the notes, the assistant infers that the patient likely has Type 2 Diabetes. This allows for more proactive and informed decision-making. 

One interesting finding of the usability study is that 84% of users wanted to see the source data behind AI-generated answers, underscoring the need for transparency. Also, Healthix reports that users value the ability to interrogate the assistant further, indicating a preference for systems that support follow-up questioning and deeper exploration.

Rogow said that one of their biggest user groups so far has been from the New York City Department of Health and Mental Hygiene. He said that department has 400 people every month actively using the Healthix portal. “They don't have any EHR product, so the portal is the EHR for their use. They're doing investigations and looking at outbreaks. They're trying to understand the chief complaints rolling in, and what's happening in the community in real time. One of the advantages of Healthix is that everything's in real time. The AI Assistant can help them find what they need faster, and they can get through more patients to stay on top of things.”

I asked Rogow about the fact that many clinical users say they don’t like to take the extra step to go to an HIE’s clinical portal and would prefer to access HIE data right in their EHR workflow. To address that need, he said, Healthix has worked with fellow New York HIE Hixny, which built an app that provides a FHIR-based patient record snapshot with vital patient information such as history, meds, and labs, right in their EHR work flow. (See our October 2025 in-depth interview with Hixny CEO Mark McKinney about that solution.)
“But we haven't been able to get the AI Agent into that app yet,” he said. “The clinician or care manager has to take the time to go into our portal to use it. So we have two different solutions at the moment. We'd like to bring these worlds together as things mature.”