Draft for Nick’s review. This note is a working draft in Nick’s voice. It is not published advice, and it is not a legal, procurement, or technical opinion until he edits it.
Call it a utility when it does the job of one
States are being asked to choose between the health information exchange they have and the health data utility they might become. The name is the easy part. Coverage, governance, and the alert are the work.
States are being asked again what they are building when they fund health information exchange. The newer name is health data utility. The older one is state HIE. I have worked the Minnesota version of this question. I helped build and expand the Encounter Alert Service, from an early footprint of two or three hospitals and two subscribers to a statewide notification network serving about 3.5 million lives. The label was never the interesting part.
When people use the words carefully, a health data utility is a shared function with a public purpose. The data a Medicaid program, a hospital, a health plan, and a public-health agency all need, governed and paid for as infrastructure rather than as a feature of one vendor’s contract. Encounter notification is the clearest example I know. An admission should be able to reach the clinician or the plan responsible for what happens next, without every participant standing up a private interface to every other participant. That is a utility. It is also what a serious state HIE has been trying to be.
The failure mode I worry about is a procurement that buys the noun. A platform, a data lake, or a slide titled “health data utility” does not create participation, ADT coverage, or a permitted purpose anyone will sign. Sometimes the technology gets ahead of the agreement. Sometimes the agreement gets ahead of the feed. Either way, a care manager who does not trust the alert is telling you that you have a repository, not a utility.
When someone asks me which model to pick, I start with the job.
What decision gets better because this data moved. A transition of care. A health-plan outreach. A rural clinician who would otherwise hear about the admission weeks later. If we cannot name the decision, we are not designing a utility.
Who has to contribute, and what they get back that they do not already have. Hospitals send ADT when the receiving side is real. They will not sustain a feed for a strategy chapter.
How it is governed and paid for when the grant turns over. Rural Health Transformation Program funds, a Medicaid enterprise budget, and a state appropriation each run on a different clock. A shared service that lives only inside a single vendor term is a project with a long name.
Which national networks it sits beside: TEFCA, Carequality, CMS Aligned Networks. National exchange does not replace a statewide alert that a Minnesota plan and a Minnesota hospital both recognize. It also asks the state network to say, out loud, how it participates in the national ones.
Minnesota does not need to win a naming contest. The Encounter Alert Service matters because it is specific: a purpose, a network, and a scale that shows up in transitions of care. Other states will keep the HIE name or recharter as a utility. I would rather spend the meeting on coverage, governance, and whether the alert arrives.
If you are writing the RFP this year, write the job. The name can follow the design, once the design is real.
Write to Nick Regier if this is the decision in front of you.