What the 2026 Hemovigilance Module Got Right, and What We Might Be Giving Up

It’s two in the morning. Your pager goes off. A nurse tells you her patient spiked a fever — 38.4°C, up 1.2 degrees from baseline — about two hours into a unit of packed red cells. She stopped the transfusion and is waiting for your call.

You work it up. DAT negative. Plasma clear. No hemoglobinuria. No respiratory distress. No hypotension. The patient’s mildly uncomfortable but gets better with acetaminophen. In the morning you write it up: febrile non-hemolytic transfusion reaction, FNHTR, imputability definite. Classic presentation.

Before January 2026, that case went into the National Healthcare Safety Network Hemovigilance Module as a reportable adverse reaction. It became one data point in a national count of how often this happens, in which patients, with which products. After January 2026, it goes nowhere. You file it in your own system, call it whatever you call it, and move on.

That change is the story I want to tell.

What Changed

The NHSN Hemovigilance Module has been the national platform for transfusion safety surveillance in U.S. hospitals since 2009. For most of its life, it asked participating facilities to report a broad taxonomy of adverse reactions — twelve defined reaction types across twenty separate forms — classified by case definition, severity, and imputability. Reactions that were possibly, probably, or definitely related to a transfusion were required. Every FNHTR. Every allergic reaction above the minor threshold. Every delayed hemolytic. Every hypotensive reaction. The idea was to capture the full landscape of transfusion-associated harm.

Version 3.0, released in January 2026, makes a dramatic cut. Required reporting now covers exactly four reactions: transfusion-associated circulatory overload (TACO), transfusion-related acute lung injury (TRALI), acute hemolytic transfusion reaction (AHTR), and transfusion-transmitted infections (TTI). Everything else — FNHTR, delayed hemolytic, delayed serologic, allergic, hypotensive, transfusion-associated dyspnea, post-transfusion purpura, transfusion-associated graft versus host disease — is now classified as “Other.”

The “Other” category is optional. And the CDC has stated explicitly that “Other” data will not be used to calculate rates.

Twenty forms became four. Twelve defined reaction types became four, plus a catch-all.

What the New Module Gets Right

The case for simplification is real. One of the quiet problems with the old module was inconsistent participation. Not every institution had the infrastructure to report reliably. The imputability framework — which required classifying each reaction as definitely, probably, possibly, doubtfully, or not related to the transfusion — was applied unevenly across hospitals, with significant variability in how individual blood banks interpreted and recorded those categories. A national database is only as good as the data going into it, and data entered inconsistently is a form of noise.

Fewer required forms means lower burden, and lower burden means better compliance. If the goal is accurate national surveillance of the most dangerous transfusion reactions, it makes sense to focus on the reactions most likely to cause serious harm or death. TACO, TRALI, AHTR, and TTI represent the sharp end of the risk spectrum. They are the reactions you lose patients to. A TTI Rapid Alert form that triggers within 72 hours — a new feature in v3.0 — is a genuinely useful public health tool for catching emerging pathogens in the blood supply before they spread.

The new module also adds a TTI Investigation Form with a structured pathway for coordinating between hospitals, health departments, and the CDC. That’s a meaningful improvement in how we respond to the reactions that matter most urgently.

What We Might Be Giving Up

Here is where I want to slow down.

The first loss is data. FNHTR is the most common transfusion reaction we see. Allergic reactions are a close second. Delayed hemolytic transfusion reactions, particularly in patients with sickle cell disease, can be life-threatening. These reactions are now optional to report and will no longer appear in national rate calculations. If you want to know how often FNHTR happens per unit transfused in the United States, or whether that rate is changing, you will not be able to answer that question from NHSN data going forward. The baseline we’ve been building since 2009 is effectively being abandoned for these reaction types.

The second loss is resolution. Collapsing eight defined reaction types into “Other” doesn’t make those reactions disappear — it just makes them indistinguishable from one another in the national record. A hypotensive reaction and a delayed serologic reaction both go into the same optional bucket.

The third loss is perhaps the most underappreciated. For fifteen years, the NHSN Hemovigilance Module provided something that the field rarely talks about explicitly: a shared vocabulary. The case definitions — FNHTR requires fever ≥ 38°C with a change of at least 1°C from baseline, or chills, within four hours of transfusion cessation; DHTR requires a positive DAT between 24 hours and 28 days with serologic evidence and inadequate hemoglobin rise — were the language everyone agreed to speak. The module’s own protocol disclaimed clinical use of these definitions, and yet every transfusion medicine fellow learned them. Every blood bank used them. Every paper in this field cited them.

Those definitions still exist in Section 6 of the new protocol, archived for reference. But they are no longer the required framework for national reporting. Without that institutional anchor, definitional drift will come. Not immediately — the field’s memory is long — but over a decade, as trainees learn from attendings who learned from a protocol that no longer exists in the same form, variability will creep in between institutions, between publications, between how we talk to each other about what a reaction even is.

The Honest Landing

I don’t think this change is obviously wrong. The logic behind it is defensible, and the improvements to TTI surveillance are real. What I can’t tell you is whether the tradeoffs will be worth it. That answer will take years to emerge, and by the time we know, we will have already lost the data we chose not to collect.

That’s the thing about surveillance systems. The cost of narrowing them is invisible at first. You don’t see the data you’re not gathering. You don’t miss the baseline you’re no longer building. The gap only becomes visible later, when someone asks a question about FNHTR rates in 2030 and realizes the answer stopped being tracked in 2026.

For now, my FNHTR gets documented in our local system, classified with whatever terminology we happen to use, and counted in no national total. Whether that’s fine — whether the simplification is worth the resolution we gave up — is a question I’m genuinely not able to answer yet.

I’m curious whether others in the field see it differently.