Every hospital administrator has met the moment when a system simply stops answering: the scheduling screens freeze, the front desk reaches for paper, and someone asks whether this is “just IT” or something worse. When the cause is a hostile intruder, the way the next day unfolds depends on how well the organisation has practised what the security world calls ResponseOps. This is a non-clinical, administrative piece about operations only, and it works through six common myths that quietly make those days longer and more expensive.
First, a word on the label. The phrase has no official status. The Ops Atlas entry describes it as an editorial label for responding to hostile intrusions as a permanent, maintained skill instead of an improvised sprint. If you write it into a contract or a policy, use the plain phrase “incident response” instead, which every auditor and insurer already understands. That is actually myth number one, so let us start there.
Myth 1: “ResponseOps is an established standard”
It sounds official, so people assume a certification body stands behind it. None does. The international standard family that covers the practice calls it ICT incident response operations, and emergency agencies use “response operations” for floods and earthquakes, which is a different subject entirely. Even inside the software world the word is borrowed for other things.
Why it matters for a hospital: procurement teams who write the invented label into a vendor agreement create arguments later about what was actually promised. Use the standard vocabulary in paper, and keep the new label for internal conversation if it helps people remember that response is a capability you maintain, not an event you survive. The Ops Atlas entry on ResponseOps is upfront that it is a working label, which is worth remembering when you read it.
Myth 2: “Wipe and reimage, and we are finished”
Rebuilding a workstation feels decisive. It is also the fastest way to destroy the very records that tell you what happened. A wiped machine removes one foothold and every clue attached to it, while the person who used that foothold may hold three others you have not found.
Think of an administrative office that finds one forged signature on a purchase form and shreds the whole filing cabinet. You have removed the evidence and learned nothing about how the forgery got there. Good practice copies first, hashes what it collected, records who handled it and when in a single time zone, and only then rebuilds. Memory goes before disk because it vanishes the moment power is lost.
Practical takeaway: put “preserve before you rebuild” on the one-page plan, with the name of the person who may authorise an exception.
Myth 3: “Contain everything, immediately”
The instinct to pull every plug at once is understandable and often wrong. An intruder who senses a door closing may accelerate, or may trigger prepared destruction that was waiting for exactly this signal. Isolation and power-off can set off such triggers on some malware.
The balanced view in the Atlas, drawing on the US federal guidance, is that containment is a set of trade-offs: how much more damage is possible, whether evidence survives, whether services people depend on stay available, how many staff the step consumes, how dependable the step is and the length of time it will hold. The newer revision also asks teams to weigh a fast recovery against the value of watching the intruder a little longer.
A few technical details show why speed alone fails. In one major cloud, changing a firewall-style rule does not cut a connection that is already established. Short-lived access keys keep working until their own timer runs out, so removing the user does not end a session already open. Disabling an account in one identity service can take about fifteen minutes to bite for modern clients and roughly sixty minutes for the rest. Each of these means a step that looks complete on the console may not yet be complete in the world.
The practical rule: if you decide to wait before a containment step, time-box the wait and write down who decided.
Myth 4: “Our backups are fine, so recovery is simple”
Backups are essential, but a restore answers only half of a ransomware story. Where attackers copy data out before encrypting, a clean restore does nothing about the copy that left the building. This pattern, often called double extortion, means the notification and legal questions survive the technical recovery.
There is a second trap. If the backup itself was taken after the intruder arrived, restoring it can bring the intruder back and lead to a second encryption. The sensible order is to know when the attacker first appeared, restore from before that point, and rebuild identities so the same stolen keys stop working. The Atlas also notes that US federal guidance allows eradication and recovery to proceed together once all persistent access has been accounted for, which is a reminder that the sequence is a judgement, not a ritual.
For a facility that already thinks hard about follow-up planning, the parallel is familiar. Pieces such as the one on remote follow-up arrangements after cardiac procedures show how much of good service sits in the arrangements around the event, not the event. Recovery plans are the same: the restore is the visible step, and the planning around it carries the weight.
Myth 5: “Label the report privileged and it stays private”
Many organisations hire an outside investigator through their lawyers and assume the resulting report is shielded. The Atlas cites a well-known 2020 federal ruling in which a court ordered a financial company to hand over an investigator’s report anyway. The reasoning was that an earlier contract with the same investigator implied the firm would have obtained a near-identical report regardless, which meant that sending it via lawyers altered nothing about its character.
This is not legal advice, and nobody should treat a blog post as such; the point for an administrator is narrower. Do not assume that a stamp on a document protects it. Ask your own counsel how engagement letters should be structured, and keep business-as-usual security reviews separate from post-incident investigations.
Myth 6: “Few logged events prove we are doing well”
Counts measure workload, not quality. A department that logs fewer events may simply have fewer sensors, or fewer people willing to report. A more useful set of measures looks at time: the span an intruder spent inside before anyone spotted them, the gap between suspicion and a formal declaration, the time spent before the spread was halted, and how far the footprint widened when people finally searched properly.
The worked example in the Atlas has an intruder active from the evening of one day until a morning alert about eleven hours later, a forty-five minute gap before declaration, and containment after a further five hours and forty-five minutes. The scope grew from six affected systems to fifteen, which is two and a half times the first estimate. The lesson is that the early picture almost always undercounts, so plan for growth rather than defend the first number.
Quarterly measures can be simple. Count how many incidents saw the intruder return, what fraction of legal or contractual notifications were sent on time, and what share of collected evidence passed its integrity check. A hospital operations lead can read those three lines without a technical background.
The clocks nobody owns
One more trap deserves its own heading because it ruins otherwise good responses: nobody owns the decision that “we are now aware”. Many regimes start a countdown at awareness, and they differ widely. Some are measured in hours, one Indian rule in six hours, while a US health-privacy rule gives sixty days and a US market-disclosure rule runs for four working days after a firm decides the event is material. Organisations that cover several regimes need a map showing which clock applies to which kind of event, reviewed with counsel.
An administrator does not need to memorise any of these figures. What helps is a named owner for the awareness call, a contact list that does not depend on the affected systems, and a habit of writing the timestamp down when the call is made. The Atlas also warns against running the response itself over the compromised email or chat. Keep an out-of-band bridge, even a plain phone conference line, ready before you need it.
What to do on Monday
You can act on all of this without buying anything. Name one person who leads response and one executive sponsor. Write a one-page plan that fits on a single sheet, including the order of preservation, containment and recovery, and the phone number of the bridge. Agree on a few emergency accounts that bypass normal sign-in, stored somewhere offline. Then rehearse with a short desk exercise, and pick the two largest gaps it exposes.
Readers who like their planning tied to recovery in practice may enjoy the piece on how structured rehabilitation supports recovery, as the mindset of staged, rehearsed, measured steps carries over neatly. And for a simple lesson in preparing paperwork and logistics ahead of time, planning ahead through a remote consultation shows how preparation shortens uncertainty for everyone involved.
The six myths share one root: treating response as a single heroic act. It is closer to a trained routine, with preparation, evidence, ordering and ownership doing the heavy lifting. If your facility only manages one improvement this quarter, make it the one-page plan and the rehearsal. Everything else can be layered afterwards.