Book Appointment

IdentityOps Compared: Two Hospitals and Their Leaver Habits

Healthcare IT

Two administrators, call them Asha and Mehul, run the staff systems for two imaginary hospitals of about the same size. Both are conscientious and both are busy. Over a year their organisations drift apart, and the difference has little to do with budget. It comes down to how each one practises IdentityOps, the quiet work of keeping staff and software sign-ins accurate. This comparison is a thought experiment about administration only; it contains no clinical content and no statement about any real hospital.

Monday morning at the first hospital

Mehul’s team works from a clear belief: when HR says someone has left, the directory account gets turned off, and that is the end of it. Tickets close quickly. The monthly report shows nearly every departure handled within the agreed time. Everyone feels fine.

What the report does not show is the bit behind the curtain. A departing administrative assistant’s login to the room-booking tool still works, because that tool keeps its own user list. A former contractor’s reporting key still runs a nightly export, because nobody knew the key existed. A browser session opened the previous afternoon stays alive until its token expires on its own.

Monday morning at the second hospital

Asha’s team shares the belief that departures matter but tracks a different number. They follow each leaver from the HR notice to the last connected application and ask whether any account remains. The people behind the Ops Atlas write-up on IdentityOps describe this as counting from the first notice to the final application instead of stopping at the directory, and they show the gap with a small worked example: a figure of 90.0 percent can shrink to 82.5 percent once three stray application accounts are counted.

Asha’s dashboard is less comforting, and that is the point. A lower honest number gives her something to fix; a higher flattering one gives her team nothing but pride.

Side by side

  • What counts as done. Mehul: the main account is disabled. Asha: no live account anywhere, sessions revoked, times recorded.
  • Who knows the app list. Mehul: whoever set each one up, from memory. Asha: a register with an owner and a sign-in method for every tool.
  • Help-desk resets. Mehul: a call-back to the number on file. Asha: a tiered script, with a manager vouching on live video for assisted cases and a notice sent to the person afterwards.
  • Machine credentials. Mehul: found when something breaks. Asha: listed, owned, dated, and reviewed monthly.
  • How often they practise. Mehul: when an auditor visits. Asha: a leaver drill each quarter.

Why call-backs make a shaky safety net

Mehul’s call-back habit sounds sensible, and it is the kind of kindness that feels like diligence. The Atlas spells out the weakness: a number can be taken over by SIM swapping, and the advisory it cites about a well-known intrusion group describes help-desk resets as a favourite route in. A recovery process that depends on a number an attacker can steal is only a polite door.

What the sessions teach

A detail from the Atlas travels well in a story. Tokens do not vanish the moment an account is disabled. Depending on configuration, a default token might last an hour or more, and with continuous evaluation features it can run far longer. Asha’s team added a step that revokes sign-in sessions on a leaver event and tested it by asking a volunteer to stay logged in while “leaving” in a drill. Mehul’s team has never tried.

The same pattern on a medical-travel reading list

Readers who follow this blog for background on how accreditation figures into evaluating overseas oncology care are used to the idea that paperwork must reflect reality. Staff identity is that principle turned inward: the roster should match who is truly working there. Another article, a general overview of arranging care across borders, shows how many parties touch a single journey, which is a good picture of how many systems touch a single employee.

Where the stories converge

Both administrators eventually read about the incidents the Atlas cites: an old test setup lacking a second factor, a service credential tucked into a private cloud folder, a pile of unrotated machine secrets, and a series of help-desk resets. The writers’ conclusion is that each defeated routine rather than mathematics. Mehul recognised his own habits in that list. He did not need new technology to change; he needed a register, a drill and a better dashboard.

Asha, for her part, was not perfect. Her team once forgot that a particular vendor sends SCIM deactivation as an attribute change that the application chooses how to interpret. A departed user kept read access until someone noticed the application treated “inactive” as a label, not a lock. They added a nightly comparison between an application’s own export and the HR roster, instead of trusting the provisioning logs alone.

A small scene to finish the comparison

Imagine both administrators on the same Friday afternoon, each asked by a director whether last month’s departures were fully closed. Mehul answers yes with confidence and is wrong in a way he cannot see. Asha answers that a chunk of the work is still unproven and shows the list of what remains. The director, unexpectedly, trusts her more.

A fair verdict

Neither hospital is a villain. Mehul’s way is a very ordinary place to start, and it is easy to reach for because it feels complete. Asha’s way costs a little more attention and returns an accurate picture. If your own team is somewhere between them, the cheapest improvement is probably to change the measurement before changing any tool. Count the departures that finished everywhere, publish that number, and let it nag.

From there, name owners for applications, give help-desk staff a script, and put a quarterly drill in the diary. Small steps, taken in order, move a hospital from Mehul’s Monday to Asha’s.