Field Notes #6: Blind Spots

Blind Spots

On viruses, visibility, and the signals we miss

Field Notes is where I take one idea from the episode—something that feels like a hinge point—and follow it to see what it reveals. If you want the full story, you can read or listen to the episode here.

In the Margins

Somewhere in almost every Ebola outbreak, there’s a moment when the virus is already there, but the tools built to recognize it are still looking for something else. In Yambuku in 1976, clinicians thought they were facing the same postpartum infections, malarial fevers, and severe bacterial illnesses they saw every week; the idea that a new filovirus was moving through their wards didn’t match the diagnostic imagination of the time. Decades later, in Ituri, clinicians had something the Yambuku staff never did: “rapid Ebola tests” sitting on GeneXpert machines, ready to run cartridges tuned to Zaire ebolavirus. But the virus that actually arrived was Bundibugyo — a species those cartridges weren’t built to detect — so the test that was supposed to close the gap between suspicion and confirmation quietly widened it instead.

That’s one of the strange, recurring patterns in outbreak work: tests aren’t neutral windows onto reality; they’re stories about what we expect to see. When we stock only Zaire‑specific cartridges, or build case definitions around the last big epidemic, we’re making a bet that the next crisis will look enough like the last one for our tools to recognize it on sight. When that bet is wrong, transmission doesn’t stop; it just moves through a kind of diagnostic negative space, where patients keep getting sick but the machines keep saying “not that.” And by the time someone sends samples off to a reference lab, or asks whether the panel itself might be too narrow, the outbreak has already had a head start.

This isn’t just an Ebola problem. Respiratory panels, sexually transmitted infection screens, even the “standard labs” ordered in emergency departments all encode assumptions about which pathogens are likely, which exposures matter, and which possibilities can safely be ignored. We build hardware and protocols around those assumptions; then, over time, they start to feel like facts about the world rather than design choices. A negative result becomes shorthand for “ruled out,” even when what we’ve really ruled out is a much smaller claim: that this particular test, in this particular configuration, did not recognize anything it was taught to recognize.

Bundibugyo is a reminder that the space just outside those configurations matters. When a virus steps even a little bit outside the categories we’ve operationalized — a different species, a different ecology, a different symptom mix — it doesn’t have to evolve into a new kind of super‑pathogen to surprise us. It only has to slip beyond the edges of what our tools are currently built to see.

Underlined

From there, you can start to trace patterns or signals around diagnostic blind spots — the ways our diagnostic bets ripple outward into policy, perception, and memory.

• Where tests don’t match ecology

  Regions using Zaire‑only cartridges while living in an ecosystem that also supports Bundibugyo, Sudan, or other filoviruses.

• Time from first suspicious cluster to first confirmed non‑Zaire Ebola

  How long it takes before someone sends samples to a reference lab because “the usual” panel is negative or doesn’t quite fit the clinical picture.

• Which pathogens get multiplex upgrades and which stay single‑species

  Respiratory panels now routinely look for multiple viruses at once; hemorrhagic‑fever tools are still often built around one expected species.

• Outbreak narratives that only begin at confirmation

  Press releases and situation reports that start the timeline with “first confirmed case,” even when local clinicians knew something was wrong days or weeks earlier.

• How funding cuts show up as “diagnostic lag” instead of a line item

  Budget cuts matter, but so do the quiet choices about what gets stocked and renewed — and what doesn’t.

These aren’t just quirks of lab logistics; they’re the quiet architecture of how an outbreak becomes legible. When the tools are narrow, the story we tell about what is happening narrows with them, and by the time we realize what we weren’t seeing, the virus has been writing its own version of events.

What It Points To

We manage what we detect, and we detect what our tools are built to see.

Outbreak Updates

Updates will only include information verified through credible reporting or official public health sources.

Ebola DRC and Uganda

As of May 19, the DRC and Uganda Ministries of Health report a total of 536 suspected cases, 105 probable cases, 34 confirmed cases, and 134 deaths across the two countries, including two confirmed cases and one death in Uganda linked to travel from DRC. In its May 19 update, WHO emphasized that transmission likely spread for some time in rural areas before health authorities recognized the outbreak, which helps explain both the high number of suspected cases and the steep curve once surveillance intensified.

On May 17, U.S. officials confirmed that an American healthcare worker who was exposed while caring for patients in DRC tested positive for Bundibugyo virus disease and was transferred to Germany for treatment, along with high‑risk contacts. On May 18, U.S. authorities implemented enhanced travel measures, including entry restrictions and screening for certain recent travelers from DRC and Uganda, while stressing that the overall risk to the U.S. public remains low.

MV Hondius hantavirus investigation

As of mid‑May, WHO and regional agencies still describe the hantavirus cluster linked to the expedition cruise ship MV Hondius as a limited, multi‑country event involving Andes virus, which is capable of rare person‑to‑person transmission through close contact. WHO’s most recent clarification notes that one previously reported case was ultimately negative, bringing the tally to eight confirmed cases and two probable cases, including three deaths, all among passengers or crew from the ship.

Additional cases have now been identified in repatriated passengers, including at least one laboratory‑confirmed infection in Canada, while public‑health agencies in multiple countries continue to monitor a small number of high‑risk contacts; officials still stress that there is no evidence of sustained community transmission beyond the ship‑associated cluster and that overall risk to the general population remains low. WHO’s “Hantavirus in Focus” webinar on 20 May is using the MV Hondius event as a case study to review hantavirus ecology, Andes‑virus transmission, and what this episode implies for surveillance and preparedness.

Postscript

Thank you for subscribing. 🫶

As I’m typing this, it’s storming outside—thunder, lightning, and that heavy, soaking rain that makes the whole house feel quieter. I love it. There’s a very nerdy part of me that now wants to do an episode on why we can smell rain before it arrives, but I’m trying to finish this newsletter before I disappear down that rabbit hole.

This week’s Ebola episode also pulled me back to 2014, when I first started writing about infectious diseases on a little blog that didn't do too bad. I went back and re‑read some of those posts while working on this script, and now I’m debating whether to archive them on the new site—messy, timestamped thought process and all. If that’s something you’d actually want to see, I’d be curious to know.

As always, you can reply directly to this email if you have questions about the outbreaks, ideas for future episodes, or just want to tell me what you’re noticing from your corner of the world. I can’t respond to everything, but I do read what you send, and it helps shape where this project goes next.

— Heather