Book Appointment

Six IncidentOps Myths Hospital IT Teams Should Retire

Hospital IT

Ask a hospital IT manager how the department handles a serious outage and you will often hear a confident description of a process that, on inspection, rests on half-truths. These beliefs are comfortable, widely repeated and quietly expensive. What follows is a myth-by-myth check of six popular IncidentOps assumptions, each weighed against what published guidance actually recommends. The subject is administrative and technical operations only; nothing here concerns clinical decisions or patient care.

Our yardstick is the Ops Atlas entry covering IncidentOps. It openly concedes that the term is informal, a seldom-used label for what is ordinarily called incident management, and it collects practices from emergency-services doctrine and from several well-documented industry playbooks. The myths below are drawn from its list of common misconceptions.

Myth 1: The most senior person should take charge

It feels natural to hand the reins to the director or the longest-serving engineer. The guidance points the other way. One widely used playbook hands command to whoever outranks the others among those actually on the call, whatever their place in the organisation chart, and the Atlas recommends rotating the duty so that more people learn it. Seniority in a day job does not equal skill at coordinating a chaotic hour. A senior manager who joins late is better placed as a stakeholder who asks questions than as a commander who overrides the plan.

The reality: pick the coordinator for the duration of the event, train several people, and let rank sit outside the call.

Myth 2: Calling it an incident makes the team look bad

Departments that measure themselves by how few incidents they log teach staff to delay. The Atlas observes that punishing the count of declared events suppresses declaration, which is exactly the opposite of the behaviour a hospital wants. A small incident declared early and closed quickly is cheap. An event that two teams struggle with privately for three hours before anyone formalises it is expensive.

The reality: reward early declaration, define written criteria so the decision is not personal, and treat a quiet log as a warning sign, not a trophy.

Myth 3: Nobody should act until the cause is known

Engineers love to understand things, and understanding feels like progress. Yet most of the useful emergency moves need only a location for the fault. Rolling back the last release, shifting load away from a suspect component, turning off a recently enabled feature, switching to a standby or deliberately shedding some demand can all restore service without a diagnosis. The Atlas illustrates the point with a large cloud-cluster failure that lasted more than six hours; once the faulty images were located, reverting them would have ended the problem by mid-morning instead of around noon.

The reality: stabilise first, explain later. Save the deep investigation for the review, when the pressure is off and the evidence has been preserved.

Myth 4: The root cause was human error

Review documents that end with “operator mistake” feel tidy, and they are almost always unhelpful. The Atlas cites classic safety research to remind readers that hindsight distorts our assessment of what people could have known at the time. Its own illustration is a major storage outage from February 2017, which began when a mistyped value in a routine runbook step took out more capacity than anyone meant to remove, and the affected service then needed a full restart. A story like that invites asking why the step allowed such an input at all.

The reality: ask what conditions made the error possible and cheap to make, and stop the analysis at system conditions such as missing guard rails, unclear procedures or tiring schedules.

Myth 5: More follow-up actions mean a better review

A review that produces thirty tasks looks diligent. In practice, long lists dilute priority, hide the two that matter, and are rarely finished. The Atlas names several anti-patterns here: vague actions, actions that are all ranked equally, actions that are mostly mitigations rather than fixes, and closing items by quietly downgrading them. It leans towards a brief list where each entry has a single owner and a date, plus an explicit completion target. One published example gives priority actions a window of either four or eight weeks, set by how critical the service is.

The reality: fewer, sharper actions, tracked by completion rate and not by how many were written.

Myth 6: Everyone should join the call

It is tempting to invite every manager, in case they need to know. The result is a crowd. The Atlas refers to field research in which hundreds of participants poured into a single channel and pushed the core responders into private side conversations, which defeats the purpose of coordination. Its suggested remedy is easy to apply. Give responders their own working space, and set up a second, read-mostly space carrying updates for everybody else.

The reality: a small group works, a larger group watches, and a communications lead keeps the watchers informed on a schedule.

A bonus myth about rotas

One more belief deserves a mention because it hits morale. Some managers assume a handful of engineers can cover nights and weekends without strain. The Atlas calculates that six people sharing a weekly rota with two slots would each be on call about 33 percent of the time, which exceeds the 25 percent cap that Google suggests, and that around eight people would be needed. A rota that looks full on paper can still exhaust people in practice.

A short scenario that exposes several myths at once

Imagine, purely as an illustration, that a hospital’s admissions system slows to a crawl on a weekday afternoon. Nobody has declared anything, because the duty manager worries it might reflect badly. The most senior engineer takes the lead by default and dives into logs searching for the cause, while a dozen colleagues join a shared chat and offer suggestions. Two of them restart different services at the same moment. Forty minutes pass. The eventual review blames a configuration slip, lists nineteen actions, and nobody follows up.

Each decision in that story corresponds to one of the myths above, and each can be replaced by something calmer: an early declaration against written criteria, a rotating commander who coordinates rather than debugs, a rollback of the afternoon’s change while the investigation waits, a small group with a separate update space, and a review that ends with three owned actions.

Why these myths persist

Each of the six has a plausible origin. Seniority-based command mirrors how many organisations run everything else. Fear of declaring grows wherever incident counts appear on a performance dashboard. The root-cause reflex comes from engineering training. The human-error conclusion saves time in report writing. Long action lists impress auditors, and big calls reassure anxious managers. None of this makes people foolish; it simply means the habits need deliberate replacing. That usually starts with a short written policy and a rehearsal in which the new behaviour is practised without real pressure.

What to do instead

Retiring a myth takes more than announcing it. People need a replacement habit and a chance to try it somewhere safe. A thirty-minute tabletop session, in which the group walks through an invented outage and each person plays an assigned hat, is the cheapest way to do that. Run it twice, swapping the commander each time, and note where the group slips back into old patterns. Those slips tell you which myth is still alive in your department.

If you only change three things, choose these. Write declaration criteria and publish them. Name a commander for each event on the basis of the call, not the org chart. And limit the review to a few owned actions. For related ideas on how an organisation builds resilience into its everyday setup, see our piece on keeping infrastructure steady through sound IT operations, and for a different kind of structured decision, the walkthrough on sifting orthopaedic hospital options for long-running hip trouble shows the value of agreeing on criteria before the pressure arrives.

Myths survive because they are rarely tested. Pick one from this list that sounds like your department, write down what actually happened in your last disruption, and compare. The gap between the belief and the record is usually the most persuasive argument you will find for change.