Joslin Diabetes Center — February 2020 breach
Clustered 4 filings across 4 jurisdictions · filing window Sep 14, 2020 → Sep 20, 2020. View entity profile → Other incidents for this victim →
incident inc_479d254fa5c042cd · merged by deterministic · confidence 100%
Litigation Timing
SupplementalDiscovered → first regulatory filing
Range of discovered_at dates across filings
Time between earliest and latest filing
Not recorded for this incident
Leak precedence · Materiality delta · SEC filing delay — no leak-site claim in this cluster; no SEC 8-K in this cluster.
Incident timeline
Dashed segments are unestablished, not zero — they fill in as filings merge into this incident.
Member cascade — every filing about this breach
- SEP 14HHS OCRHHS OCR noticefirst filing · states nationwide total 71,160auto-linkedday 0
- SEP 18WA AGWA AG notice1,212 WA residentsauto-linked+4d
- SEP 18CA AGCA AG noticeno count statedauto-linked+4d
- SEP 20MT AGMT AG noticemost recent · 20 MT residentsoperator-confirmed+6d
- Watching for additional filings — new sources merge into this incident automatically.
Roll-up facts — reconciled across members
- Breach window
- Feb 7, 2020 – May 20, 2020 WA AG
- Discovered
- Jul 16, 2020· 160d undetected WA AG + MT AG
- Data types
- Health (basic) · Identity (basic) · PHI · PII HHS OCR + WA AG + CA AG +1
- Attack vector
- Unauthorized Access HHS OCR
Each fact cites the member filing that establishes it; when filings conflict, every value shows with its source.
Count reconciliation
per-state reported counts
Nationwide (stated in a filing)
71,160
Sum of filed state slices
1,232· 1.7% of stated
98% of the stated total lives in states whose filings carry no count — the dashed share shrinks as filings land.
Evidence ladder — rungs this incident occupies
No leak-site claim on record for this incident.
No press coverage linked yet.
Unlocks: discovery date, data types, affected count, compliance clock.
No SEC filing or victim statement yet.
Merge provenance
- Method
- deterministic
- Confidence
- 1.00 · above the 0.85 auto-merge floor
- Reviewed
- operator-reviewed Jun 12, 2026 · at least one link edge was human-adjudicated
- Audit
- Every link edge records its method, confidence and model + prompt versions (100% is the strongest edge).
Merges are reversible — a wrong link can be detached with its audit trail intact. How merging works.
Filing velocity
Spread
6 days
Cadence
~1 / 2d
vs. multi-state median
1.3× faster
About this clustering
DisclosureLens links filings into incidents through layered matchers: deterministic rules (same source document, multistate filings of one breach, tight-window same-victim pairs), a weighted-similarity scorer for cross-source candidates, and an operator review queue for everything uncertain. Each link records its own method and confidence — shown per filing in the timeline below. The system defaults to NOT merging when uncertain, because a false merge (collapsing two unrelated breaches) is more harmful than a false split (showing related filings separately); uncertain pairs route to human review instead of auto-merging. This page states only what the linked filings agree on; each filing's own account stays on its own record, linked from the timeline below.
The weekly clock on records like this one
The Disclosure Clock is a weekly briefing on what landed in the disclosure record, and the gap between the criminals' post and the regulator's filing — measured on the whole record, with the method shown.
Weekly. Double opt-in, one-click unsubscribe, and the address goes nowhere else.