The Problem With ADT Feeds Isn't the Feed. It's the Foundation

The Problem With ADT Feeds Isn't the Feed. It's the Foundation

‍

The patchwork problem.

Almost every risk-bearing organization needs a way to know when patients are hospitalized. Historically, this has been done through ADT feeds. What that leads to is an accumulation of feeds. What they often don't have is a clear-eyed view of what those feeds are actually missing.

Here's how it typically unfolds. One vendor gets added for a region. Another for a specific health system footprint. A state HIE contract comes in through a compliance requirement. Over time you have five, six, maybe ten feeds, each covering something the others miss, and a team spending meaningful time figuring out which feed to trust when they conflict. The patchwork creates operational drag before it creates clinical value. And even with all of it stitched together, coverage still isn't where it needs to be to do transitions of care well.

The instinct to add another feed is understandable. The problem is that it's solving for the symptom, not the cause.

Signal was built from a different premise: that a transitions-of-care monitoring layer should be a single thing, not an aggregation of things. One connection to national exchange infrastructure. One integration instead of a reconciliation workflow. Coverage that reflects where patients actually receive care not where your contracts happen to reach.

That last point matters more than it sounds. Patients don't stay inside one health system. They use out-of-network facilities, see specialists across geographies, and show up at facilities no regional contract was ever designed to reach. If your monitoring layer is built around known relationships, you have blind spots everywhere those relationships end. Signal is built on national networks because that's the only way to close gaps systematically rather than one contract at a time: 89% U.S. hospital coverage, 25% more transition events than the typical multi-vendor ADT mix.

But the coverage numbers aren't the point. Coverage is what makes everything else possible.

Context is the other half and it's where most alerting quietly fails.

The teams using Signal aren't sitting by a screen waiting for alerts. They're managing patient panels, handling escalations, closing care gaps, chasing prior auths. When a notification lands, they need to triage it in real time: Does this patient need action today? Can it wait? Who owns the follow-up? If the alert doesn't answer those questions immediately, it becomes noise regardless of how fast it arrived.

ADT messages contain a lot of fields. But they were built for routing, not care coordination. What they don't reliably carry is the clinical picture: the reason for admission, the medication changes, the risk factors that determine whether this patient needs outreach in the next two hours or the next two days. That's not a flaw in the format, it's a recognition that the format was designed for a different job. Signal is built to carry clinical context from the start, so care teams have a working picture of who this person is before they've made a single phone call.

And then there's the discharge summary problem which is harder than it sounds.

Finding and associating a discharge summary with a specific transition event is historically difficult and time-consuming. Most of the time, the summary isn't even available at discharge. When it is, it may arrive hours or days later as a PDF running tens or even hundreds of pages. For a care coordinator trying to do medication reconciliation or schedule follow-up, that process breaks down exactly when it matters most, the 72-hour window after discharge when readmission risk is highest.

Waiting for paperwork means the workflow stalls at the moment it should be moving fastest.

Signal's AI-generated discharge summaries solve this directly. Rather than waiting for a document and then parsing it, Signal synthesizes a structured clinical summary from the full discharge context - encounter data, diagnoses, procedures, medications, and prior transitions. When the source document exists, we incorporate it. When it doesn't, we generate from the existing clinical record. The output is structured, consistent, and ready the moment it's needed.

Coverage, context, timeliness. Most ADT approaches were optimized for one. Signal was built for all three.

And they're not independent variables. Timeliness without coverage means fast alerts on a fraction of your population. Coverage without context means your team starts from scratch on every notification. Context without timeliness means you get the full picture after the window already closed.

Each problem compounds the others. Solving transitions of care means solving for all three simultaneously and building the infrastructure underneath that makes all three possible together.

The coverage numbers, the clinical context, and the AI-generated summaries aren't separate features. They're the same design decision, expressed at different layers of the product.

The question worth asking isn't whether your current ADT setup is working. It's whether it was ever designed to do what you're asking it to do.

‍

Read more on this topic of ADT Pitfalls from our Clinical Strategy Principal here