Setting Up a Biomedical Engineering Department in a Hospital

Bed count is the number everyone reaches for and it is a poor proxy. A 100-bed hospital with two cath labs, a linear accelerator, and twelve ventilators carries a heavier equipment burden than a 250-bed hospital doing mostly general medicine and obstetrics. The right starting point is a counted equipment population, weighted by criticality and service complexity. Count devices, not systems: a monitoring network of forty bedside units and two central stations is forty-two devices to maintain, not one installation with one asset tag.

Build the weighting from three factors. First, criticality: does failure of this device stop care immediately, within a shift, or not at all? Second, service complexity: can your own technician open it, or does the manufacturer hold the service password and the calibration jig? Third, regulatory attachment: does the device sit under AERB licensing, a CDSCO field safety pathway, or a calibration requirement traceable through an NABL-accredited laboratory? A device scoring high on all three consumes disproportionate departmental time even as a single unit.

Once counted and weighted, the population tells you the shape of the team rather than only its size. A hospital dominated by imaging needs someone who can hold manufacturers to contractual response times and read a service report critically. A hospital dominated by ICU and theatre devices needs hands, benches, and spares on site. Very few hospitals need both in equal proportion, and hiring generically produces a team that is mediocre at whichever one actually matters to your case mix.

No published ratio survives contact with a real equipment mix, and anyone quoting one should be asked what it assumes. What does work is building the establishment bottom-up from planned work. Add the annual planned maintenance hours implied by your maintenance calendar, an allowance for breakdown response drawn from your own last twelve months of complaints, calibration coordination time, and procurement and installation support. Then divide by realistic available technician hours after leave, training, and administration. The number you get is defensible in a way a ratio never is.

That calculation usually surprises people in two directions. Planned maintenance turns out to be a smaller share of the workload than expected, and unplanned response a larger one. Meanwhile the paperwork load, covering condemnation files, purchase specifications, incident investigation, and accreditation evidence, often consumes close to a full post on its own in a hospital of any size. Budgeting for that post explicitly is more honest than expecting technicians to do it after hours and then finding the records incomplete on the day of assessment.

Skill mix matters more than headcount at the margin. A department of four with one degree-qualified biomedical engineer, two diploma technicians, and one coordinator who owns records and vendor follow-up will outperform a department of six technicians with no records discipline. The coordinator role is the one most often cut in a cost review and the one whose absence shows up first, because it is the role that turns work done into work evidenced. Cutting it does not save money, it just moves the cost into audit season.

The asset master is the single list of every device the hospital owns. The equipment history card is the life story of one device on that list. Hospitals routinely have the first and not the second. A history card should carry commissioning and acceptance test results, every planned maintenance visit, every breakdown with cause and downtime, every calibration certificate reference, firmware changes, relocation history, and eventually the condemnation decision. Read end to end, it should let a stranger judge whether the device is fit for continued clinical use.

Keep the history card attached to the device identity, not to the location or the department. Devices move, and a monitor that spent three years in a general ward before being pushed into high dependency carries its history with it. Serial number is the anchor. Internal asset codes get changed during audits and re-tagging exercises more often than anyone admits, and if your system lets you re-tag without preserving the old code as an alias, you will lose history at exactly the moment you need it.

Wherever the record lives, it needs to be the same place the technician actually works. A history card maintained in a workshop register and re-typed into software monthly will diverge within a quarter, and the version an assessor reads will be the wrong one. HealUDoc can hold the asset record and its maintenance history against the department that owns the device, which at least removes the re-typing step. Whatever the tool, the principle is one record, updated at the point of work.

The honest comparison is not in-house versus outsourced. It is which parts of the work you keep. Routine planned maintenance on general-purpose devices, first-line fault triage, and record keeping are almost always faster and cheaper in-house, because response time is measured in minutes and the skill required is not exotic. High-end imaging, radiotherapy, cath lab and analyser service usually is not, because manufacturers hold the parts, the software keys, and the test jigs, and no single hospital can justify duplicating that infrastructure.

The failure mode of full outsourcing is that you lose the ability to judge the vendor. If nobody inside the hospital can tell whether a reported fault was real, whether the maintenance visit actually happened, or whether the quoted part was needed, then contract management becomes faith-based. Keeping a small in-house capability is partly a technical function and partly a purchasing control. That second argument lands better with a finance committee than an appeal to clinical safety, and it happens to be true.

The trade-off is real and worth stating plainly. In-house capability costs a salary line that is visible every month, while outsourced service arrives as invoices tied to events and is easier to defer in a difficult quarter. A hospital under cash pressure finds it much easier to cut the technician than the service contract, even when the technician prevents more downtime. If you want the in-house team to survive its first budget review, the measures in the next section have to exist from month one.

Four measures carry most of the argument. Equipment uptime for critical devices, expressed as available hours against required hours rather than as a raw percentage of the calendar year. Mean response time from complaint logged to technician standing at the device. Mean repair time from arrival to device returned to service. And planned maintenance completion against schedule, counted on the date the task was due rather than the date it was eventually done. The fourth measure quietly predicts the other three.

Collect them from the complaint log, which means the complaint log has to be the only route by which a fault gets reported. The moment a nurse can call a technician on his mobile and have it fixed off the record, your data stops describing reality. This is a discipline problem rather than a software problem, and it is usually solved by making the logged route faster than the informal one, not by prohibiting the informal one and hoping people comply.

Report the numbers by department and by device family, never as a hospital-wide average. An overall figure of 97 per cent uptime conceals a ventilator fleet sitting at 88, and it is the concealed number that will hurt you. Take the same view of the top ten devices by downtime hours each quarter. That list is usually short, stable, and points directly at either an ageing fleet, a weak service contract, or a user-handling problem that an hour of training would fix.

Start with the count, not the hiring. Weeks one to four: physically walk every clinical area and build the asset master from what is there, not from the fixed-asset register, which will disagree with reality. Expect to find devices on no list at all, devices listed that no longer exist, and at least one device in daily clinical use that has never been serviced since installation. Photograph serial plates as you go, because you will need them for warranty claims and specification work later.

Weeks five to eight: classify by criticality, attach existing contracts and warranty status to each asset, and stand up the complaint log with a single reporting route. Do not attempt a full maintenance calendar yet. A reliable complaint log running early gives you a baseline of real workload, which is what both the staffing case and the first maintenance calendar depend on. Publish the response times you are actually achieving, however poor they look, because the trend is what you will be judged on.

Weeks nine to twelve: write the scope note, agree the establishment, and build the first year of planned maintenance from the criticality classification. Then hold two lines as the department grows. Every fault comes through the log, and every device carries a history card. Departments that hold those two lines cope with growth and with staff turnover. Departments that do not end up rebuilding their asset master from scratch every time an accreditation cycle comes round, which is an expensive way to run an engineering function.