Federal surveillance, analytics and cited evidence for earlier Lyme recognition and better care coordination.
Nothing is shown until you pick one. The dashboard will not default to a county you did not choose — every figure below a location is about that location, and a placeholder would make all of them wrong in a way that still looks right.
Sign in to track your assessments, store lab records, and build a clinician brief from your own history. Sign in. Everything else stays visible without an account — it is federal open data.
Every published CDC case is accounted for
152,018 cases: 148,156 county-coded, 3,694 held as a state residual, 168 with no usable state. Reconciles exactly rather than approximately.
County risk ranking holds year over year
Spearman ρ = 0.882 ranking 424 counties from 2022 into 2023, with leave-one-state-out controls for geographic leakage.
Extracted lab values are traceable to the document
Every value quotes the span it was read from, checked in code against the transcript. Unverified values default to rejected and stay flagged if kept.
The design decisions have published evidence behind them
47 findings from 14 peer-reviewed studies, registries, and surveys, each traced to the product change it caused — or to why it changed nothing.
Red-flag symptoms route to care before any model runs
Deterministic screening on every message, ahead of the language model, because a prompt instruction is a request a long conversation can argue a model out of.
Whether this helps a real patient
Zero primary interviews conducted. The evidence base is published research about people like our users, not evidence from our users.
Whether a clinician wants a FHIR export from a patient tool
No study speaks to this. It may be genuinely useful or it may land in a scanner queue like everything else a patient brings in.
Whether anyone will upload a lab report here
77% pooled willingness to share health data, but a 24–100% range across studies — and much higher for one's own clinician than for a third party. We are the third party.
Whether the tier names mean what they sound like
A county labelled 'high' has roughly a 39% chance of being in a different tier next year. Very-high is stable; high and moderate are not.
Whether the escalation tiers are clinically right
The boundaries have had no clinical review. One rule was added on documented evidence of harm; the rest are judgment.
Anything about individual diagnosis
There are no patient-level outcome labels anywhere in this project. Nothing here has been, or could have been, validated as a diagnostic instrument.