The idea for this piece came from a conversation with an IT lead at a cardiac care centre — we’ll call him Joseph — who spent twenty minutes describing his hospital’s annual equipment audits, calibration logs and staff recertification calendar, then went quiet when asked a simple question: when did anyone last audit the skills of the people running the hospital’s software? Monitors get checked. Pumps get checked. The team keeping the appointment system, telemetry dashboards and patient portal alive gets… assumed. So we built what Joseph asked for: a tech skills checkup for hospital teams, structured like the clinical audits hospitals already trust. Work through it honestly — the way you’d want a patient to answer screening questions.
Why a skills audit beats a skills guess
Hospitals know better than anyone that risk hides in the unexamined. A cardiac unit would never let “the ECG seems fine” substitute for calibration records, yet most healthcare organisations run exactly that logic on their technical staff: things seem fine, therefore people must be competent. The flaw shows up the same way it does in medicine — asymptomatically, then suddenly. The checklist below takes an hour per team member and produces something most hospitals have never had: a documented picture of who can actually do what, and where the gaps sit before an outage finds them.
Part one: the coverage check
- Name the systems that cannot go down. Appointment booking, lab results delivery, telemetry displays, pharmacy ordering. Write them in a column.
- Beside each, name two humans who can diagnose it at 2 a.m. Not vendors — employees. Any system carrying fewer than two names is a staffing fragility dressed up as coverage.
- Match each person’s skills to a named discipline. Reliability work, deployment automation, security operations, cloud cost control — the operations field now has defined specialisations with defined syllabi. If a team member’s expertise cannot be named, it cannot be trained, replaced or recruited for.
- Flag the orphan disciplines. Joseph’s audit found nobody owned security patching; everyone believed someone else did. Classic, and common.
Part two: the evidence check
This is where the storytelling gets uncomfortable, because the honest answer at most hospitals is “we have no evidence, only tenure.” For each person and each claimed skill, ask what documentation exists. The modern standard is straightforward: structured learning tracks end in graded assessments, and passing one yields a certificate others can authenticate by scanning a code. On xOps School, where Ops disciplines are taught as examinable tracks, every discipline closes with a quiz of twenty-five questions and a verifiable credential for those who clear the bar — which means “our database person is good” can be replaced with a record a hospital auditor would recognise as evidence. Joseph’s team ran their baselines on the platform’s free practice tests over two lunch breaks. Three people scored lower than their confidence predicted; one scored dramatically higher and had been underused for years. Both discoveries paid for the hour immediately.
Part three: the practice check
- When did each person last fix something hands-on in a safe environment? Reading about failover is not performing one. Structured platforms stock hundreds of practical assignments — north of three hundred on the one above — precisely because employers and auditors increasingly distinguish watched knowledge from performed knowledge.
- Does anyone rehearse under time pressure? Incidents arrive with clocks attached. Timed practice exams exist for the same reason code-blue drills do: the first pressurised attempt should never be the real one.
- Is there a wrong-answer log? Clinical culture reviews adverse events; learning culture reviews missed questions. A team that records what it got wrong, and why, compounds. A team that only celebrates passes repeats its mistakes at longer intervals.
Part four: the growth check
Audits that only measure are autopsies; this one should change the treatment plan. For every gap found, assign one discipline, one person, one quarter — small, funded and scheduled. The funding question usually dissolves on inspection: entry access to structured Ops training now costs less monthly than a hospital spends on a single stethoscope, with free layers generous enough that baselines and standards literacy cost nothing at all. Protect three hours a week per learner on the roster, the same way CME hours are protected for clinicians. And when someone clears a track, put the credential in their file and celebrate it in the team meeting — visible recognition is how a one-off audit becomes a culture. For team members heading toward formal roles, pair the study with our round-by-round practice plan for DevOps interviews, which turns the same preparation into career mobility — retention improves, counterintuitively, when people know they could leave.
The annual rhythm
Run the full checkup yearly, aligned with your other audits so it inherits their seriousness. Quarterly, do the short version: re-test one discipline per person, rotate the 2 a.m. names, review the wrong-answer log together. Joseph’s centre now treats reliability skills the way cardiology treats a well-prepared heart checkup — measured, discussed and acted on rather than feared. His team’s reliability drills borrow directly from the approaches in our piece on building durable skills for dependable systems, and eighteen months in, their unplanned downtime has stopped being a standing agenda item.
Start with one column
If the full audit feels heavy, start tonight with part one’s first two items: the list of systems that cannot fail, and the names beside them. Most leaders who write those two columns discover the rest of the checkup has become emotionally unavoidable. That discomfort is the point — hospitals teach, better than any industry on earth, that early screening is cheaper than emergency intervention. The skills keeping your patient-facing systems alive deserve the same standard of care as the hearts those systems serve.