https://failsystems.health201.com/ A sourced, continuously updated reference on failure in automated healthcare: power, connectivity and data, devices, AI models and agents, the human handoff, and the cascades between them. What fails, how you'd know, and what to do at each degradation tier. en-us Sun, 27 Sep 2026 00:00:00 +0000 https://failsystems.health201.com/method/#tiers-rulehttps://failsystems.health201.com/changelog/#2026-09-27-0 Sun, 27 Sep 2026 00:00:00 +0000Tiers reassigned by a documented rule (most ransomware and IT outages are tier 2, not 3); AWS 2025 moved from Power to Connectivity; studies, experiments and non-clinical analogues are now labelled and carry no tier; wrong-output failures no longer read as a fall to tier 0. Second sources added for Princeton Community Hospital, Katrina/Memorial and the sepsis-alert case, whose accounts are shown where they differ. Contec entry now includes independent analyses disputing the backdoor label. Joint Commission alarm goal updated to NPG.01.05.01. Homepage scores marked as draft. https://failsystems.health201.com/incidents/mgh-monitor-alarm-off-2010/https://failsystems.health201.com/changelog/#2026-09-27-1 Sun, 27 Sep 2026 00:00:00 +0000Massachusetts General Hospital, January 2010, from Boston Globe reporting; its limitation (no inspection record found) is stated on the page. https://failsystems.health201.com/https://failsystems.health201.com/changelog/#2026-09-26-2 Sat, 26 Sep 2026 00:00:00 +0000Every failure layer now has its own page: how it fails, the incidents behind it, how you'd know, defenses at each degradation tier, US standards, the EU/UK position, a v0.1 severity score and open questions. 38 incidents and 205 sources, every link opened and checked. https://failsystems.health201.com/tiers/https://failsystems.health201.com/changelog/#2026-09-26-3 Sat, 26 Sep 2026 00:00:00 +0000What must hold for each of the six layers at each of the four tiers, plus the evidence on rehearsal and on coming back up. https://failsystems.health201.com/method/#updateshttps://failsystems.health201.com/changelog/#2026-09-26-4 Sat, 26 Sep 2026 00:00:00 +0000New incidents and sources are found, verified and added automatically; changes to page text and scores wait for human review. https://failsystems.health201.com/method/#disclaimerhttps://failsystems.health201.com/changelog/#2026-09-26-5 Sat, 26 Sep 2026 00:00:00 +0000FailSystems is an aggregation and synthesis of published sources for information only: not consulting or professional advice, and no approach fits every organisation. https://failsystems.health201.com/https://failsystems.health201.com/changelog/#2026-09-26-6 Sat, 26 Sep 2026 00:00:00 +0000The footer now points to the right authorities for a live incident: 911, CISA and the FBI for cyberattacks, local emergency management (which escalates to state and FEMA), and FDA MedWatch for devices. https://failsystems.health201.com/layers/connectivity/#detecthttps://failsystems.health201.com/changelog/#2026-09-26-7 Sat, 26 Sep 2026 00:00:00 +0000The response-time threshold for a functional downtime is an hourly mean over 5 seconds (or 3 SD above the mean) per ONC's SAFER guide; under 2 seconds is SAFER's target, not the downtime threshold. Also reworded the extended-downtime failure mode to state the Joint Commission's four-week guidance directly. Found by the first local fact-check. https://failsystems.health201.com/https://failsystems.health201.com/changelog/#2026-08-03-8 Mon, 03 Aug 2026 00:00:00 +0000Landing page with the failure taxonomy and degradation tiers.