by Ashley DeGarmo, Chief Client Officer, and Topher Rasmussen, Communications Manager
Health care organizations invest heavily in data infrastructure–analytics platforms, monitoring dashboards, EMR integrations. However, there is often a gap between what’s invested and what actually moves outcomes.
As a part of the Health Care Council of Chicago’s Q3 summit, Third Horizon convened a panel of leading data experts – including a researcher, an epidemiologist, a data visualization expert, and a public health leader – who explored what separates effective data work from infrastructure that sits unused.
From left: Roy Ahn, VP of Public Health, NORC; Heather Blonsky, VP of Data, Metopio; Shelly Sital, Director of Public Health Initiatives, AllianceChicago; Alfreda Holloway, Director of Epidemiology, Cook County Department of Public Health; Ashley DeGarmo, Chief Client Officer, Third Horizon
What emerged was a consistent pattern: the most effective data work does two things simultaneously. It’s rigorous—methodologically careful, transparent about how conclusions were reached. And it’s relational—grounded in shared purpose with the people who’ll use it and be affected by it. Neither element works without the other.
To understand how this works, consider the rigor dimension first:
Methodological Rigor
Metopio builds and uses data to simplify community health needs assessments and community health improvement plans. As Heather Blonsky, Vice President of Data, shared, metric specificity matters: “We have really thought this through and we have agreed that the pieces we need are here. This is why we picked it, and this is how we calculate it, and this is what it’s based on and this is what it includes.”
This specificity—knowing why this metric and not another—is foundational.
The same principle shows up across different contexts. Roy Ahn, Vice President of Public Health at The National Opinion Research Center (NORC) demonstrates why this matters across a completely different domain—social media sentiment analysis. Rather than starting with a hypothesis about what keywords matter, they build from the ground up. “There’s sampling, there’s interrater reliability, there’s hand coding… because the social media world is very good at picking up sentiment,” said Roy. “But doing a blanket keyword search leaves out a lot of informal terms that we all use.”
Roy’s work reveals why this matters. Keyword shortcuts misrepresent what communities and individuals are actually saying about a topic. Precision—truly understanding how people talk about something—requires slower, more human work. It’s the difference between capturing what someone said and understanding what they meant.
This rigor shows up across different contexts, including in how organizations approach emerging technologies. Cook County Department of Public Health approaches emerging tools with similar care. Rather than deploying AI for rapid data analysis, they take a more measured stance. “We don’t use AI for analysis, said Alfreda Holloway, Director of Epidemiology. “We just don’t… And it’s not that we don’t think it’s something that could be viable for us. We just really need governance around it.”
The distinction matters. It’s not that AI can never be useful. It’s that Cook County won’t deploy a tool without clear governance—without being able to explain not just what the system concluded, but how it reached that conclusion. That caution protects the integrity of the work.
Rigor Grounded in Purpose
Rigor matters because it creates credibility. But credibility directed toward what? The practitioners describe something more specific: rigor in service of shared purpose.
This means more than methodological care. It means being transparent about what’s being measured and why. It means committing to return findings in forms the people studied can actually use. Roy Ahn at NORC emphasizes the importance of this kind of transparency. “Making sure that the organizations that are being evaluated feel like you understand what it is that they’re being evaluated on and how that data are going to be used… is really important,” Roy said.
It’s not just transparency about methodology but building community trust in the process.
More fundamentally, there’s a commitment to returning findings in forms communities can use. “If we promised the community that we were going to work with them on this research and bring it back to the community and share, then you have to go back and share. That is a bright line that you cannot cross,” Roy said.
This framing—“bright line you cannot cross”—treats data sharing not as aspiration but as commitment. When data flows one direction only—from community to analyst to funder—it becomes extractive. When it flows back, it becomes collaborative.
AllianceChicago, a network of health systems in the Chicago area, demonstrates this. Eleven competing organizations (including Rush, Northwestern, University of Chicago, and Cook County) wanted to share clinical data. Rather than launching directly into technical design, they engaged in months of relational work. Shelly Sital, Director of Public Health Initiatives, said “We had 6:30 AM meetings in-person every other Friday at the beginning of this network to really get people together at the same table to understand what the point of doing all this is?”
That investment in relational agreement protected the scale that followed. Today, the network encompasses 12 million covered lives, providing health systems with visibility they never had. But that scale serves communities because it was built on agreement about purpose first—not imposed as a technical solution.
Cook County’s evolution illustrates how grounding in purpose changes what sophistication in rigor accomplishes. Alfreda Holloway, Director of Epidemiology, said “”Not only at the level of the health department, it’s our academic institutions, it’s our partnerships, it’s our community members. They can do their own analysis.”
The sophistication—analytical capability, downloadable datasets, technical infrastructure—didn’t concentrate power in the epidemiology department. It redistributed it. Communities, academics, and health departments are empowered to ask their own questions rather than waiting for centralized analysis.
The practitioners on the panel operate within both tracks, rigor and grounding. The rigor creates credibility with decision-makers and stakeholders, while grounding ensures the work serves real-world practical applications.
What This Means
What distinguishes effective data work from infrastructure that just sits there? The practitioners on the panel demonstrate three concrete things:
- They do the unglamorous work first: Validation, transparency, understanding communities before designing systems
- They ask and answer the relational questions upfront: What is this data work for? Who decided? Do we agree?
- They commit to sharing findings back: In forms communities can actually use, not in forms that benefit analysts or funders
These aren’t optional add-ons to rigor. They’re what make rigor matter.
Organizations initiating data work should consider what’s present in the design:
On rigor: Are metrics chosen with transparent reasoning about why this measure and not another? Is there willingness to do verification work before deployment? Can someone articulate the methodology clearly enough that it could be defended or refined?
On relationships: Is there agreement about what the data work is for—what problem it addresses, who it serves? Did communities and stakeholders help define that purpose? Will they receive findings back in forms that are actually useful?
The practitioners on the panel operate within both dimensions. The rigor creates credibility with decision-makers and stakeholders. The grounding ensures the work serves a purpose people actually agreed upon.
The organizations leading on health care data—whether in public health, research, managed care, or safety-net delivery—refuse to separate rigor from relationships. They understand that methodological care builds credibility for relational work, while relationships give rigor purpose and direction.
For organizations evaluating or launching data initiatives, the diagnostic is clear: If you have both dimensions, the work compounds. If you’re investing in one without the other, you’re not doing data work. You’re building a tool that will likely sit unused or mislead stakeholders.
That’s an organizational insight, not a technical one. It’s what differentiates data work that changes health care from data work that just creates infrastructure.