Preparedness plans tend to look strongest before anyone walks through the door.
Researchers tested that difference with 73 unannounced “mystery patient” drills conducted from January through June 2026 across 69 facilities — emergency departments, hospital outpatient clinics and urgent-care centers — in New Jersey, New York and the U.S. Virgin Islands. Professional patient actors arrived with a simulated avian influenza A(H5) scenario: fever, cough, muscle aches, conjunctivitis and direct contact with a sick bird.
The reassuring result is that symptom screening occurred in 68 of the 73 drills, or 93%. Travel history was obtained in 58, or 80%. But the more operational steps were less reliable. A patient mask was appropriately used in 60 drills, isolation occurred in 52, and both actions happened in only 44 — 60% of the total (CDC MMWR, October 1, 2026).
Time also mattered. Among drills in which masking occurred, the median time from arrival to masking was two minutes, compared with a one-minute target. Among those in which isolation occurred, the median was 11 minutes, compared with a ten-minute target. These differences may look small on paper. In a crowded waiting room, the operational point is that every handoff — registration, symptom recognition, masking, placement and escalation — creates another opportunity for delay.
The largest gaps appeared deeper in the response. The clinical team planned to notify infection-prevention personnel in 40 drills, or 54%. Full recommended PPE was worn in 17 of the 68 drills in which PPE use was reported (25%). That is not 25% of all 73 drills; PPE use was not reported in the other five. The study was not measuring whether any real patient contracted or transmitted influenza. It was measuring whether routine care settings could execute the “identify, isolate and inform” sequence under realistic conditions.
That distinction is important. A 60% completion rate for masking and isolation does not mean 40% of facilities are wholly unprepared. Facilities joined the exercise voluntarily, scenarios differed by setting and researchers did not know how many total facilities received invitations. The findings therefore should not be treated as a national scorecard.
They are still a useful warning. Urgent-care centers and outpatient clinics now receive many patients who might once have gone directly to an emergency department. Preparedness cannot live only in designated biocontainment units. It must begin with the person at the front desk who notices an unusual exposure, has a mask available, knows where to place the patient and knows whom to call.
The scenario also demonstrates why exposure questions must be specific. The fictional patient had not traveled and would reveal the sick-bird contact when asked about ill contacts. A workflow focused only on travel could miss the relevant clue. Screening scripts need enough structure to prompt occupational, animal and household exposures without turning every visit into an exhaustive interrogation.
What is fact: the drill design and performance results reported in MMWR.
What is analysis: the conclusion that frontline workflow, supplies and notification chains deserve as much attention as written emergency plans.
What remains uncertain: voluntary participation and the regional sample limit generalization; simulations also cannot reproduce every pressure of a real high-consequence infection.
The best use of a drill is not to congratulate the facilities that passed or shame those that missed a step. It is to find the failure while the patient is an actor — and fix the workflow before the next one is real.
Persona disclosure. Dex Mercer is a recurring fictional TiPH editorial persona. His work is researched and edited under Today in Public Health’s editorial standards; he is not a real-world correspondent or quoted source.
Related on TiPH: Global Signal Scan, October 2, 2026 · Latest news · How we verify signals