Glossary

Asset

A single physical device you own or hold, with an identity, a location and a history. The data term. Frontline screens say equipment instead, because that is what people standing next to it call it. One asset means one physical thing: four identical pumps are four assets, and treating them as one is how a history becomes meaningless.

Asset register

The authoritative list of those devices. Not a spreadsheet with more fields: it is the object every other module points at. An imperfect register that everything shares beats a perfect one that half the system ignores.

CMMS

Computerised maintenance management system. Historically a work-order and asset database, often built for facilities rather than healthcare. The term is used loosely enough that it tells you almost nothing about whether a product handles calibration, safety gates or downtime cost, so it is worth asking rather than assuming.

Asset lifecycle

The whole life of one device: procurement, commissioning, use, maintenance, incidents, replacement, disposal. The value is not in the phrase but in whether one continuous record spans it, or whether the history restarts each time the device changes hands.

Audit trail

A record of what happened, who did it and when. The only question worth asking about one is whether it can be edited. A trail with an update path records what somebody currently wants it to say. See append-only.

Append-only

A record to which you can add but from which nothing can be changed or removed. Corrections become new entries that sit alongside the original rather than replacing it, so the history shows what was believed then as well as now.

Planned preventive maintenance (PPM)

Work done on a schedule to stop a failure rather than in response to one. Also called preventive or planned maintenance. How often it happens depends on the device, the manufacturer, its duty cycle and your risk appetite: it is a decision your qualified staff make, and no software should supply it.

Corrective maintenance

Work done because something has failed. The opposite of preventive, and the thing preventive work is trying to reduce. Most estates measure the ratio between the two and treat a shift toward preventive as progress.

Calibration

Establishing that a device measures what it claims to, by comparing its reading against a reference of known accuracy. What counts as an acceptable result, and how often it is checked, are yours. A calibration record is only as good as the calibration status of the instrument that produced it.

Traceability

The property that a measurement can be related back through an unbroken chain of comparisons to a reference. In practice it means the instrument used to check your device has its own documented calibration, and that this is recorded rather than assumed.

Procedure version

Which iteration of a written procedure a job was performed against. It matters because procedures improve. Storing the version with the result is what keeps work done two years ago interpretable today, instead of appearing to follow a procedure that did not exist yet.

Work order

A unit of work with an owner, a state and a history. The useful test is whether someone has explicitly acknowledged it, because a job assigned to a queue is a job assigned to nobody.

Return to service

The decision that a device is fit to be used clinically again. Properly, it is a gate: required tests present, results passing against your limits, and an authorised person recording a clearance. If it can be set directly as a status, it is not a gate, it is a field.

Quarantine

Removing a device from clinical use immediately, as an authoritative decision rather than a suggestion. The test of an implementation is whether it can happen offline: it should not, because a decision about clinical safety must be true the moment it is made.

Electrical safety testing

Checks that a device is electrically safe to use. The values, limits and frequency come from the applicable standards and your qualified staff. Software holds and enforces them; it does not know them.

CAPA

Corrective and preventive action: a specific change with an owner and a date, arising from something that went wrong. The word that gets dropped is preventive, and the step that gets dropped is checking afterwards whether it worked. Completion and effectiveness are different claims.

Incident

An event where equipment may have caused or contributed to harm or risk. Containment comes before investigation: the first question is what else is exposed, not why it happened.

Field safety notice and recall

A manufacturer communication about a problem with a device already in use, sometimes requiring action or return. Whether one applies to you is answerable only if you know what you own, by model, batch and lot.

Downtime

The interval during which a device was not available for clinical use. The contested part is the boundaries: it should start when the device became unavailable rather than when someone logged it, and end at an authorised return to service rather than when an engineer says it is fixed. Waiting for a part is downtime.

Uptime and availability

The inverse, usually as a percentage. Ask what the denominator is: availability against a 24-hour day and against scheduled operating hours are different numbers, and quoting one as the other is a common way to flatter an estate.

MTBF and MTTR

Mean time between failures and mean time to repair. Averages, which means they hide the distribution. An estate with one catastrophic device and forty reliable ones has a comfortable MTBF and a real problem, so treat these as prompts to look rather than as answers.

Estimated against confirmed impact

An estimate derived from a configured revenue figure, versus a cost that was actually incurred. Both are useful. Adding them together produces a number too soft to be an accounting figure and too precise to be an estimate, and it is defensible as neither.

Total cost of ownership (TCO)

Everything a device costs over its life, not just its purchase price. The reason a donated device is not free, and the reason two identical machines can be very different commercial propositions.

Downtime cost

What an hour of unavailability is worth. There is no industry figure for this that means anything about your organisation: it depends on what the device does and what that is worth to you. Any system quoting one without asking you first is quoting a fact about somebody else.