Resources
Preventive maintenance guide
This is the practitioner page: how to actually run a programme, whatever software you use or do not. If you want the argument for having one, the Solutions page makes it. If you want to know how the scheduling machinery works, the Features page explains it. This one is about the job, and most of the job turns out to be politics rather than engineering.
- Biomedical engineering
- Hospital operations
- Clinical engineering
What is a preventive maintenance programme?
Definition
Preventive maintenance programme
A preventive maintenance programme is a set of decisions, not a calendar: which equipment is covered, what is done to it, how often, who is accountable when it slips, and what happens when the answer is that it slipped again.
Most descriptions stop at the calendar, which is why most programmes fail. The schedule is the easy part and it is the part software gives you. The hard parts are deciding what deserves coverage, deciding intervals nobody can prove, and holding the line when a ward needs the device today and the check can wait until next month.
None of the values in that definition come from us. How often a device needs attention depends on the manufacturer, the model, its duty cycle and your risk appetite, and it is your qualified staff who decide. What follows is about the process of deciding and running, not about the answers.
Why programmes fail, and it is never the schedule
Because preventive work is the only work whose value is invisible when it succeeds, so it loses every contest with work whose absence is visible immediately.
Consider what a preventive task competes against. A broken pump on a ward is loud, specific and has somebody asking about it. A pump that is due a check is quiet, and skipping it produces no consequence today, this week, or probably this quarter. Every individual decision to defer is defensible, and the aggregate is a programme that exists on paper.
This is not a discipline problem and treating it as one guarantees failure. It is an incentive problem: the person deferring bears none of the cost and the cost arrives months later, attributed to something else. A programme survives only if somebody with authority is confronted with the deferral while it is still a deferral rather than a pattern.
The second failure is scope. A department that plans everything produces a schedule so large that it is behind in month two, at which point overdue stops meaning anything, and the red items become wallpaper. Once a team has learned to ignore overdue, the programme is over regardless of what the system says.
How to start a programme that is still running in a year
Cover less than you want to, decide intervals explicitly, generate early, and make slippage somebody's problem while it still matters.
- 1
Pick the equipment where failure genuinely hurts
Not the whole estate. The device classes where an unplanned failure is expensive or unsafe. This is a shorter list than most departments expect, and starting with it is what makes the programme credible enough to widen later.
- 2
Decide each interval as a decision with a name on it
The inputs are the manufacturer guidance, the duty cycle, the failure history and your risk appetite. Where the guidance is silent, somebody qualified has to choose, and that choice should be recorded as a choice. An interval nobody can account for is the most common finding in a bad audit.
- 3
Generate the work far enough ahead to arrange it
Ward access, technician availability and parts all need lead time. A task that appears on its due date cannot be arranged, so it slips on arrival and the schedule was fiction from the start. How far ahead is your call; far enough that a human can negotiate is the test.
- 4
Decide in advance who is told when it slips
Before the first slip, not after the tenth. An overdue task that escalates to somebody accountable while it still matters converts a drift into a decision. The decision may legitimately be to defer, and that is fine: what kills a programme is deferral that nobody chose.
- 5
Measure completion honestly, including the misses
A compliance figure that excludes what was never attempted is not a compliance figure. If a device has no data, it has no data, and reporting that as anything else is how a programme looks healthy for two years and then surprises somebody.
- 6
Review the intervals against what actually happened
After a year you have failure history you did not have at the start. Intervals set on manufacturer guidance and a guess can now be set on manufacturer guidance and evidence. Most programmes never do this, which is why most intervals are whatever somebody typed in year one.
How to handle the conversation you will actually have
The ward needs the device and the check is due. Four positions worth having ready, because you will have this conversation more than any other.
Never make it about the schedule
"It is due" loses to "we have a patient", correctly. The argument that works is about consequence: what happens if this fails mid-use, and who explains it. If the answer is genuinely "not much", that is useful information about your interval rather than a defeat.
Bring the deferral into daylight
Deferring is legitimate; deferring invisibly is not. Recording the deferral with a name and a reason turns the next conversation from an argument about memory into a conversation about a pattern nobody disputes.
Separate the safety gate from the schedule
The pressure to defer a check is negotiable. The pressure to return a failed device to use is not, and the two must never be allowed to blur into one conversation. Everything else in a programme is a judgement; that is a rule.
Use the pattern, not the incident
One deferral is not an argument. Six months of the same ward deferring the same class of device is an argument, and it is one that can be made without blaming anyone, because the record makes it for you.
Who should own this, and what to expect in year one
Biomedical engineering with a named accountable person, and a first year that looks worse before it looks better.
Ownership needs to be a person rather than a function. A programme owned by a department is owned by nobody, and the specific failure is that overdue items have no home: they are visible, uncomfortable, and not any individual's problem to resolve. The owner does not need to do the work. They need to be the person who is asked why it did not happen.
The first year will look bad, and you should say so before it does. Coverage figures start low because you are now counting things that were previously invisible. Overdue counts rise because the tasks existed before and nobody was tracking them. This is the system telling the truth about a situation that already existed, and the department that has not warned its board about that reads it as a new problem it caused.
What good looks like at twelve months is unglamorous: a shorter list than you hoped, mostly done, with deferrals that are decisions and intervals you can account for. That is a programme. Complete coverage on paper with nobody trusting the numbers is the thing that looks like a programme and is not.
The values are yours, and this page has none
You will not find an interval, a limit or a threshold anywhere on this page, and that is deliberate rather than an omission. How often your equipment needs attention is a clinical and engineering decision belonging to your qualified staff, your manufacturers and the standards that apply where you are. Anything we published would be a guess about equipment we have never seen, and it would be adopted because it was published. What we can offer is the process of deciding, and machinery that holds your decision once made.
Questions
How often should we service our equipment?
We cannot tell you, and you should be wary of anyone who can without asking. The inputs are your manufacturer guidance, the model, its duty cycle, its failure history and your risk appetite. Where guidance is silent, somebody qualified has to decide, and that decision should be recorded as a decision with a name against it. An interval nobody can account for is the most common finding in a bad audit.
What do we do when a ward will not release a device for a check?
Do not argue about the schedule, because "it is due" loses to "we have a patient", correctly. Argue about consequence: what happens if it fails mid-use and who explains it. Then record the deferral with a name and a reason. Deferring is legitimate; deferring invisibly is what ends programmes. Six months of the same pattern is an argument the record makes for you.
How much equipment should a new programme cover?
Less than you want. The device classes where an unplanned failure is genuinely expensive or unsafe, which is a shorter list than most departments expect. Planning everything produces a schedule that is behind by month two, and once a team has learned that overdue means nothing, the programme is finished no matter what the system reports.
Why did our numbers get worse after we started?
Because you are now counting things that were previously invisible. Coverage starts low and overdue counts rise, not because anything deteriorated but because the tasks always existed and nobody was tracking them. This is the system telling the truth about a situation that already existed. Warn your board before it happens rather than explaining it afterwards.
Keep reading
Preventive maintenance
The argument for having a programme, written for the person deciding.
Maintenance planning
The scheduling machinery underneath: templates, versions, generation, slippage.
Checklists and templates
Writing the procedures the programme runs on, and why we will not hand you one.
Preventive vs corrective maintenance
The ratio everyone targets, and why it is the wrong number.
See it on your equipment
Live in an afternoon, useful the same week. A person replies, usually within one working day.
Contact us