Book Appointment

A Fictional Hospital IT Team and the Week Three Changes Collided

Hospital IT

The following story is invented. The people, the hospital and the incident are composites meant to show how ITOps habits play out in practice, and none of it describes a real organisation. It follows a small technology team at an imaginary orthopaedic hospital through one tense week in which a payroll deadline, a network upgrade and a failing label printer all wanted the same Thursday.

Monday: the calendar nobody had read

The person running the team, called the coordinator here, opened her morning queue and found three change requests waiting. One moved the registration system to a newer database version. One replaced a core switch in the outpatient wing. One was a routine certificate renewal. All three had been submitted by different people, and each assumed it was the only job planned for the week.

She did something deliberately unglamorous: she put the three on one shared schedule and looked for overlaps. The Atlas entry for ITOps, at the xOps School Atlas page covering ITOps, recommends exactly this: one calendar looking ahead, able to flag clashes, and it also mentions freeze windows such as a payroll period. The finance office had announced one for the end of the month. Two of the three changes crashed into it.

Tuesday: scoring instead of arguing

The database change had a strong sponsor and an impatient one. Rather than debate, the coordinator used the team’s written scorecard. It gave points for touching a tier-one system, for altering the data structure, and for falling outside the agreed maintenance window. The Atlas gives a sample of that arithmetic, where a change that adds up to six points goes to the review board rather than straight to implementation.

The database upgrade scored high. The switch replacement scored moderately. The certificate renewal scored almost nothing, and since it belonged to a pre-approved category, it simply went ahead. That one decision removed a third of the discussion from Wednesday’s board meeting. A common myth, as the Atlas puts it, is that the framework demands the board look at every change. It does not.

Wednesday: the review board, reshaped

The board had the habit of acting as a gate. Members arrived, read nothing in advance and asked the sponsor to defend the plan from scratch. The coordinator proposed a different role. The Atlas cites delivery research proposing that such a board stop acting as a gate and become the architect of the process and a beacon of information. In practice that meant circulating the risk scores a day early and asking only three questions: who reviewed the plan, what is the way back, and who watches it afterwards?

Two peers had reviewed the database plan. A rollback script existed and had been tested on a copy. The board approved it for the weekend after the freeze, not for Thursday. The sponsor grumbled but accepted, and the coordinator noted the date in the shared calendar so nobody could claim surprise later. A short note went out to the wards explaining what would be unavailable and for how long. Our own earlier article on building resilient infrastructure through modern IT operations makes a similar argument: resilience comes from dull preparation.

Thursday: the printer that was not a printer

At ten in the morning the label printer on the day-surgery floor stopped working. A technician replaced the cable, and it worked for an hour. When it failed again, he opened a ticket. The service desk, instead of closing it as a hardware fault, searched earlier records and found two similar tickets from the previous month, each closed after a quick restart.

That search is what separates incident handling from problem management. The team opened a problem record. A look at the equipment register showed that the printer sat behind a switch the team had scheduled to replace that very week, and that the register listed the printer’s position wrongly. The map was out of date, so the diagnosis had been pointing at the wrong device for a month.

  • The incident was closed with a workaround.
  • The problem record stayed open, tied to the switch replacement.
  • The register entry was corrected and its owner named.

Friday: the emergency that proved the rules

Late on Friday an advisory arrived about an actively exploited flaw in a remote-access component. This was a genuine emergency change: no waiting for the board, no weekend delay. The coordinator approved a rapid fix, assigned an engineer, and logged the record as an emergency. Because the team had already rehearsed the path, the fix went in within the hour.

Afterwards came the part that is easy to skip. Per the Atlas, every emergency and failed change should be reviewed. On the following Monday the team wrote two paragraphs: what triggered it, and what would let a similar fix move even faster. Regulators in some regions also expect a post-implementation approval of urgent changes, so the paperwork was more than courtesy.

What the story illustrates

Nothing in this week required a new platform. The gains came from five habits:

  1. One shared schedule, visible to everyone who changes anything.
  2. A written risk score so debates become arithmetic.
  3. A board that reads in advance and asks three questions.
  4. A habit of searching old tickets before closing a new one.
  5. A short review after every emergency.

The team in this story would be the first to say that ITOps rarely looks heroic. Success looks like a Friday that ends quietly. For a wider look at the practical side of cloud-era upkeep, see also our overview of strengthening software delivery and cloud infrastructure. And if you take one thing from this invented week, let it be this: the calendar, the score and the map are small documents, yet they decide whether a hospital’s technology feels dependable.