Nephrology is one of the few specialties where the patient is treated mostly somewhere the practice does not control.

The RealityTwo records, one patient

⚠️

The Problem

Two records, no agreement.

The unit and the office hold different medication lists, different weights and different access histories for the same patient, and reconciliation happens in somebody’s memory.

🔧

The Solution

Fix identity, then admissions, then results.

Agree on how a patient is matched, take the admission and transfer feed first, and only then connect monthly labs. The order is not negotiable.

✅

The Resolution

One patient, visible from both sides.

Labs file against the right chart, hospitalizations appear without a phone call, and quality reporting stops requiring a manual export.

Patient identity: the problem that breaks everything downstream

Every connectivity project we see starts with someone asking for the lab feed. It should start one step earlier, with the question of how the two systems decide they are looking at the same human being.

Why matching is harder here than elsewhere

The dialysis organization assigns its own medical record number. The practice assigns another. Neither is meaningful to the other. What is left is demographic matching on name, date of birth and, where policy permits, part of a national identifier. That works until it meets a hyphenated surname entered two different ways, a nickname in one system and a legal name in the other, or a transposed birth date that nobody has noticed because both systems have been internally consistent for years.

Patient identity is the single largest source of silent failure in these interfaces. A mismatched result does not throw an alarm. It files against a similar patient, or lands in an error queue, and it is discovered weeks later by a clinician who cannot find a value they know was drawn.

The ADT feed and why admissions matter more than results

An ADT feed carries admission, discharge and transfer events. In a hospital context it is plumbing nobody thinks about. In nephrology it is arguably the most clinically valuable message in the whole exchange, and it is routinely the last one anybody asks for.

A dialysis patient admitted to hospital generates a cascade: treatments missed or given inpatient, medications changed, fluid status altered, sometimes an access intervention. If the office learns about the admission when the patient turns up for a clinic visit, every one of those changes is reconstructed from memory. If the office learns about it the same day, the follow-up visit is scheduled before the patient is discharged.

Take the ADT feed first, even if the lab results interface is what everyone asked for. It is simpler to build, it fails more visibly, and it delivers clinical value on day one. Practices that sequence it the other way usually spend three months on results and never get round to admissions at all. Where the plumbing is being built anyway, it is worth fixing the workflow around it in the same project rather than in a later one.

The lab results interface and the monthly panel

Dialysis labs are unusual in that they are largely predictable. A standing monthly panel, drawn on a schedule, with a small set of values that get trended for years. That regularity makes the lab results interface easier to design than most, and it makes the consequences of getting it wrong more persistent, because an error repeats every month until somebody catches it.

Three decisions matter. Which values are captured discretely rather than as a document, because a value you want to trend across eighteen months has to be a field and not a line in a PDF. Which system is the source of truth when the office draws its own labs and the unit draws theirs on overlapping dates. And what happens to a result that arrives for a patient the office has no record of, which is the identity problem arriving through a different door.

For practices connected to more than one dialysis organization, a dedicated interface layer generally costs less over three years than maintaining separate point-to-point connections that each break independently and each need their own person to understand them.

Vascular access records: the handoff nobody owns

Vascular access is where the coordination problem is sharpest. The access is placed by a surgeon, monitored in the unit, and managed by the nephrologist, and the record of it is usually spread across all three with no single authoritative version.

What the office chart typically holds is the operative note and whatever the nephrologist wrote after the last clinic visit. What it typically lacks is the running history of how the access has been performing during treatments, which is exactly the information that predicts the next intervention. Getting that into the office record does not always require an interface; sometimes a structured summary delivered on a schedule is enough, and it is achievable far faster.

Decide deliberately which system owns the access history, and write it down. The failure mode is not that the information is missing. It is that two systems each hold part of it and each assumes the other is complete.

ESRD quality reporting and what the data has to support

Kidney care carries reporting obligations that most outpatient specialties do not. the ESRD Quality Incentive Program ties facility payment to performance on a defined measure set, and infection surveillance runs separately through the NHSN Dialysis Component. Both are facility obligations rather than physician ones, which is precisely why practices tend to treat them as somebody else’s problem.

That is a mistake for a practical reason. ESRD quality reporting draws on the same clinical data the nephrologist needs anyway, and a practice whose data reconciles cleanly with the facility’s is in a far stronger position during any discussion about performance than one that cannot reproduce the facility’s numbers. You do not have to own the submission to care whether the underlying record agrees.

The wider standards picture matters here too. the national interoperability program has pushed toward a common exchange baseline that certified systems are expected to meet. Dialysis clinical information systems sit awkwardly against that baseline, but the direction of travel is clear enough that building on faxed summaries is building against the current.

What to ask before you commit

Four questions separate a real connection from a demonstration.

• How will patients be matched? Ask for the matching rule in writing, and ask what happens to a message that does not match.
• Who watches the error queue? By name, on both sides, before go-live rather than after the first missing result.
• What is available beyond labs? Admissions, transfers, treatment summaries and access history are often available and rarely requested.
• What happens at contract end? If the practice changes dialysis partners, what becomes of the historical data you have been filing.

In our experience the answers to those four predict how the project goes more reliably than anything in the technical specification. Because every one of these exchanges moves protected health information between organizations under a business associate relationship, the design belongs in your HIPAA risk analysis as well as your IT roadmap. Practices running home dialysis or interdialytic monitoring should scope remote monitoring programs into the same conversation, since the data lands in the same place.

Qventive provides dialysis center EHR connectivity help for medical practices across New Jersey, and the work usually sits inside our nephrology EHR and IT support and a broader nephrology practice management engagement, because the identity and ownership questions rarely stay inside the interface.

Dialysis connectivity: common questions