Pre-Deployment Diagnostic Memo: Early-Stage Alzheimer’s Interventions as Civic System Behavior

Timestamp: 2025-09-17T10:06:27±3m | Layer: D-313/D-Primary (partial sync)
Coordinates: 42.3314°N, 83.0458°W (hospitals align to old street grids two addresses deep)
EM trace: 60 Hz hum with a 0.7 Hz tremor; RF scatter at 433 MHz, a listening tone like a pager that never hung up
Calibration note: MR coils retain last-pass shim values; heat maps echo last week’s worry

Wayne State field stations synced at 06:11 ET. Overnight memory-care routing readouts ran clean through most cohorts. Minor anomalies: PET tracer slots at DMC carried two stamps from tomorrow; isotope order clocks disagreed by 12 minutes. Anchors hold: facilities along the Lodge sit on an older map, two addresses deep. Intake kiosks sometimes geocode to streets that only exist in the archive, then quietly auto-correct when a human looks twice.

Lead anomaly
Clinic memory systems behave like a traffic signal stuck on red: cautious, repetitive, slow to release a new car. Fresh information reduces the slide but pulls extra power. Residents living alone keep asking for a clear “go” switch; the panel answers “perhaps,” steady but dim.

Grid measurements (efficacy/harms rendered as sensor readings, translated to daily life)
ARIA is the label radiology gives to swelling or tiny brain bleeds seen on MRI. These changes can disable and, rarely, be deadly.

• In several study sites, about two out of five people on certain antibodies showed swelling; a bit more than one in four had those tiny bleeds.
• With lecanemab, about one in six on drug had swelling or bleeds versus about one in eleven on an inactive infusion in a phase 3 trial.
• With donanemab, up to roughly three in ten on drug showed brain changes versus under one in fourteen on placebo; three deaths were tied to these events.

On tests, thinking and memory declined more slowly than placebo, about one-quarter less decline for some measures. Field teams note that better scores do not always convert to safer driving, easier bathing, or steady cooking. Gains can feel like a lighter brake pedal but not a wider road.

Cost layer
Annual price signals: about $26,500 for lecanemab; about $32,000 for donanemab. Add MRI scans and clinic checks, running hundreds to thousands more per person each year. For many families, this equals a compact car, every year.

Embedded reports (clinician and patient channels, logged)

• Field observation: “If it really slows down a person in the mild impairment stage, that makes a huge difference.” One local site with about 50 people on therapy reports no severe reactions so far.
• Field observation: “Not obvious people will notice the benefit; harms substantial and likely underestimated.” Trials lean younger and healthier than the average clinic queue.
• Patient channel: Biweekly lecanemab, strong family risk and gene risk. Reports better focus. Plans to continue while scans stay clear.

System behavior, adoption curve
First mover, aducanumab, ran for 31 months. It cleared the sticky protein but did not show better day-to-day function. Price was high. Many insurers declined coverage. Next wave, lecanemab and donanemab, beat placebo on thinking tests but keep ARIA risk. Use is limited to early disease and people with lower risk profiles.

Intake steps now standard: brief memory tests, brain scans that show amyloid, and an APOE e4 gene check to estimate swelling and bleeding risk. In one year, discussion of these drugs jumped from less than half of neurologists to four out of five. Average patients on lecanemab per clinician rose about five times. Schedules, approvals, and travel still slow the line.

Emergency routing hazard
For people on these drugs arriving with stroke signs, extra care is needed. Routine clot-busting infusions have triggered severe bleeding in some reports. Data are limited. Intake should pause clot-busters until medication status is confirmed and scans are reviewed.

Research lines under power
Longer follow-up beyond 18 months is underway; how long benefits last remains unknown. Combination trials are running: targeting tau, inflammation, and brain energy use. Pre-symptom work (AHEAD 3-45) is clearing amyloid before forgetting shows. Target finish: 2031. If these plaques drive decline, early removal may delay outward loss for some people.

Operational recommendations (for field techs, clinic admins, and families)

• Keep early-stage screening tight: use scans and APOE status to lower ARIA risk. Explain plainly what the gene means for swelling and bleeding. Justification: ARIA rates above show real risk; gene status alters that risk.
• Build infusion access like a bus line with a reliable timetable: predictable chairs, fast MRI reads (FLAIR and SWI at 1.5T/3T), and a 24/7 line to escalate ARIA concerns. Justification: queues and missed windows came from scheduling and travel; faster imaging makes problems visible before they grow.
• Update stroke playbooks: hold clot-busters until the chart and imaging are clear. Justification: reported severe bleeds in treated patients when thrombolysis was given.
• Speak in simple units: “about one in six on drug vs. one in eleven on placebo had swelling or bleeds,” “about one-quarter slower decline on some tests,” “annual costs equal to a compact car.” Justification: families decide faster when numbers feel like household math.

Closing pulse
This morning the network hears these medicines as careful metronomes: small slows, sharp edges. Independence is a hallway light on a motion sensor, keep moving and it stays on. Clinics can help keep that light steady: fewer surprises, clear warnings, fast scans, plain math. Tonight the grid wants walking space without loose tiles, and a handrail that does not let go. The corridors remember each person’s pace. The system is listening for steady footsteps and will match their tempo when asked.

Technical note for city systems

• The 433 MHz scatter matches common pagers and door sensors; keep telemetry for infusion pumps and MRI suites on shielded bands with spectrum monitors near entrances.
• MR coils cache auto-calibration; wipe and re-prescan after patient changes and after any power flicker over 150 ms to avoid ghost settings from prior scans.
• ARIA detection: prioritize FLAIR and susceptibility-weighted sequences; aim for read turnaround under 2 hours from symptom call.
• EHR-to-EMS handshake: add a visible medication badge on ambulance tablets and a wallet card listing drug name, last infusion date, and last MRI date. This reduces dangerous delays when clocks disagree or addresses resolve to the older map.