Clinical Engineering and Healthcare Technology Management Formula Sheet

This formula sheet collects first-pass calculations used in clinical engineering and healthcare technology management. It focuses on installed medical technology: readiness before clinical use, preventive maintenance workload, corrective repair capacity, fleet availability, loaner coverage, recall closure, electrical safety screening, network load, lifecycle cost, risk ranking, incident evidence, and clinical release margins.

The equations support engineering review. They do not replace local clinical governance, manufacturer instructions, qualified service procedures, medical-device regulations, electrical safety standards, cybersecurity policy, procurement rules, or patient-care escalation plans.

Basis and Boundaries

State the boundary before calculating:

1. asset type, model, configuration, software version, accessories, and intended clinical use;
2. fleet size, location, owner, clinical criticality, and backup method;
3. time basis, such as calendar month, operating hour, patient-care day, case, scan, infusion, or service event;
4. whether downtime includes only hands-on repair or the full unavailable interval;
5. whether the metric is a dashboard indicator, release gate, escalation trigger, or planning estimate.

Clinical-engineering metrics should not hide critical missing checks inside a high aggregate percentage. A fleet can show strong average completion and still be unsafe if one high-risk device lacks safety screening, network validation, alarm verification, or configuration traceability.

Symbols

Symbol	Meaning	Typical unit
N	number of devices or assets	count
N_c	count satisfying a criterion	count
T	time period or exposure	h, day, month
H	technician, service, or operating hours	h
t	duration of one event	min or h
\lambda	arrival or failure rate	events/time
\mu	service completion rate	events/time
A	availability	dimensionless
U	utilization	dimensionless
C	cost or capacity depending on context	stated basis
I	current	A
R	resistance	\Omega
B	data rate or bandwidth	bit/s
u_c	combined standard uncertainty	same as measured quantity
k	coverage factor	dimensionless

Asset Readiness and Release Gates

Readiness rate:

Readiness gap:

Critical-check pass fraction:

where j may represent electrical safety, network connectivity, alarm routing, configuration, training, calibration, cleaning, accessory verification, or cybersecurity configuration.

Clinical release gate:

The minimum function is intentionally conservative. If a critical gate is incomplete, a high average readiness score should not release the fleet. Local policy may allow controlled use of a device with missing low-risk paperwork, but safety, alarm, therapy, dose, and workflow checks require explicit engineering and clinical decision.

Weighted readiness score:

where x_i=1 if criterion i is complete and 0 otherwise. Weighted scores are useful for prioritization, not for bypassing mandatory release gates.

Utilization and Fleet Reserve

Device utilization:

Fleet utilization:

Reserve fraction:

Peak clinical demand coverage:

where M_N is the device count margin. Utilization should be interpreted with clinical context. A device may be idle because it is unnecessary, unavailable, misplaced, awaiting cleaning, missing an accessory, or reserved for emergency use.

Preventive Maintenance Workload

Monthly preventive maintenance events for a fixed interval:

where I_{PM} is the interval in months.

Preventive maintenance labor:

if t_{PM} is in minutes.

Total service workload:

Capacity margin:

Backlog clearance time when capacity exceeds new demand:

where B_0 is current backlog, \mu is service completion rate, and \lambda is new work arrival rate. This is a screening relation; it fails if arrivals, staffing, parts, access, or clinical scheduling are not stable.

Reliability and Availability

Failure rate from observed events and exposure:

Mean time between failures:

Mean time to repair:

Inherent availability:

Operational availability:

Expected unavailable devices:

Clinical availability may be lower than inherent availability because downtime can include fault reporting, locating the device, cleaning, part ordering, vendor response, software access, loaner setup, user handoff, and post-repair validation.

Loaner, Spare, and Downtime Coverage

Simple loaner count:

where f_{unavailable} is the observed or planned unavailable fraction and N_{surge} is an additional clinical surge margin.

Spare-part demand screen:

where \lambda_{part} is expected part use rate and T_{lead} is replenishment lead time.

Downtime load:

Clinical downtime exposure:

where C_{criticality} is a local consequence score or weighting factor. This score must be defined before use; it is not a universal clinical risk value.

Service Queues and Response Capacity

Single-server utilization:

For c similar service channels:

Stable screening condition:

Approximate waiting time for an M/M/1 service queue:

This model is only a rough capacity screen. Real clinical-engineering queues are affected by priority triage, parts, vendor access, patient-care schedules, infection-control hold time, travel, and device-location delays.

Recall and Field Action Closure

Recall closure fraction:

Unresolved affected devices:

Potential in-service exposure:

Closure is not only a percentage. A field action should remain open if affected devices are unlocated, still in clinical use without correction, corrected without verification, or missing configuration evidence.

Electrical Safety Screening

Insulation-resistance leakage estimate:

Leakage-current margin:

Insulation-resistance margin:

Protective-earth resistance from a test current:

Pass or fail decisions must use the applicable test method, device class, applied-part configuration, accessory configuration, test equipment calibration, and local acceptance criterion. A calculated leakage estimate is not a substitute for the required safety test.

Battery, Runtime, and Backup Capacity

Battery energy:

Runtime estimate:

Runtime margin:

Backup device count for a degraded mode:

where D_{clinical} is degraded-mode demand and C_{backup} is capacity per backup device or workflow. Backup planning must include batteries, chargers, compatible accessories, cleaning state, location, users, and escalation path.

Network Load and Data Path Margins

Data rate for sampled channels:

where N_{ch} is channel count, f_s is sampling rate, and n_{bits} is bits per sample before protocol overhead and compression.

Fleet data load:

Bandwidth margin:

Latency margin:

Jitter margin:

For alarms, waveforms, images, worklists, medication libraries, laboratory results, or remote support, network checks should include failure behavior. A device that works in normal traffic may fail during roaming, access-point overload, certificate expiry, server outage, firewall change, or time-synchronization loss.

Measurement Uncertainty and Release Margin

Combined standard uncertainty:

Upper decision value:

Lower decision value:

Release margin for an upper limit:

Release margin for a lower limit:

If the margin is negative, the measurement is too close to the limit or outside it under the selected decision rule. The response may be repeat measurement, calibration review, repair, quarantine, reduced clinical use, or formal risk-based acceptance.

Risk Priority and Evidence Completeness

Risk priority number:

Use RPN for prioritization, not as proof of safety. A high-severity hazard may need action even when occurrence or detection scores make the product modest.

Evidence completeness:

Configuration traceability:

Incident evidence gap:

Incident review is stronger when logs, software version, configuration, accessories, user action, alarm history, power state, network state, maintenance record, and clinical context are preserved before repair or reset changes the evidence.

Lifecycle Cost and Replacement Screens

Undiscounted lifecycle cost:

Present value lifecycle cost:

Downtime cost screen:

Replacement economic margin:

If M_{replace}>0, replacement has lower present-value cost under the stated assumptions. The decision still needs clinical risk, cybersecurity, supplier support, spare parts, training, infrastructure, and validation evidence.

Quality Function and Procurement Scoring

Weighted technical score:

where w_i is criterion weight and s_i is the normalized score.

Cost-normalized value screen:

Critical-requirement gate:

where g_i is 1 if a mandatory requirement is met and 0 otherwise. A high weighted score should not override a failed mandatory requirement such as electrical safety, essential performance, cybersecurity support, service access, or clinical workflow compatibility.

Worked Fleet Review Example

A hospital is reviewing a fleet of infusion pumps after a software-library update and a preventive-maintenance campaign. The simplified data are:

Quantity	Value
fleet size	N=240 pumps
preventive maintenance interval	I_{PM}=6\ \text{months}
time per PM event	t_{PM}=45\ \text{min}
monthly technician hours available for this fleet	H_{available}=90\ \text{h/month}
corrective repair workload	H_{CM}=22\ \text{h/month}
locating, cleaning, documentation and validation workload	H_{support}=12\ \text{h/month}
observed peak unavailable fraction	f_{unavailable}=5\%
clinical surge margin	N_{surge}=6 pumps
exposure during reliability review	T_{exposure}=108{,}000\ \text{pump-h}
failures during review	N_f=18
total downtime from failures	T_{downtime}=216\ \text{h}
safety notice affected devices	N_{affected}=64
corrected and verified devices	N_{corrected}=52
quarantined awaiting correction	N_{quarantined}=8
in service but uncorrected	N_{in\ service\ uncorrected}=3
unlocated devices	N_{unlocated}=1
insulation resistance during screen	R=120\ \text{MOhm}
insulation test voltage	V=250\ \text{V}
local leakage-current screen	I_{limit}=5\ \text{microA}

Step 1: Monthly Preventive Maintenance Workload

Monthly PM events:

PM labor:

Total service workload:

Capacity margin:

The maintenance campaign fits the nominal monthly labor budget. The margin should be protected for failed PM checks, devices that cannot be located, clinical scheduling constraints, software rollback, and unexpected battery or accessory replacement.

Step 2: Reliability and Availability

Mean time between failures:

Mean time to repair:

Inherent availability:

Expected unavailable devices from this availability:

The availability estimate looks strong, but it represents the failure data boundary. It does not include cleaning hold, locating delay, loaner setup, user handoff, or devices removed by the safety notice.

Step 3: Loaner Coverage

Simple loaner count:

The loaner pool should include pumps, chargers, power cords, compatible drug-library configuration, cleaned status, and asset tracking. A numerical loaner count alone is not enough if the loaners are not ready for the same clinical workflow.

Step 4: Recall Closure and Exposure

Recall closure fraction:

Unresolved affected devices:

Potential in-service exposure:

The closure percentage is not acceptable for release because four affected pumps may still be available for clinical use or cannot be located. The eight quarantined pumps reduce exposure but still require correction and verification before the field action can close.

Step 5: Electrical Safety Screen

Convert resistance:

Leakage estimate:

In microamps:

Leakage-current margin:

The screen passes the simplified local threshold. The release record should still identify test configuration, applied parts, accessories, test-equipment calibration, device identity, software version, and whether the test followed repair, update, cleaning, or incident review.

Engineering Decision

The PM workload and safety-screen calculation look acceptable. The fleet should still not be fully released until the safety notice exposure is closed. The blocking facts are:

1. 12 affected devices remain unresolved;
2. 4 affected devices may still create exposure because they are in service or unlocated;
3. loaner readiness must include configuration and cleaning, not only count;
4. the strong MTBF/MTTR result does not cover the recall condition.

The defensible action is to keep the affected software-library field action open, remove or locate the exposed pumps, verify corrections, confirm loaner readiness, and then release the fleet with updated configuration and maintenance evidence.

Common Mistakes

Common mistakes include calculating readiness as an average while ignoring one failed critical gate, reporting MTTR as hands-on repair time while excluding parts delay, using MTBF without exposure definition, and sizing loaner pools without chargers, accessories, cleaning status, or compatible configuration.

Other frequent mistakes are closing recalls from corrected count alone, treating cybersecurity or network changes as IT-only tasks, accepting electrical safety numbers without test configuration, and comparing devices by purchase price rather than lifecycle cost. Strong clinical engineering keeps calculation, clinical workflow, installed configuration, patient-safety consequence, and release evidence connected.

Practical Validation Checklist

Before relying on a clinical-engineering calculation, confirm that:

1. the device population, configuration, software version, and accessories are identified;
2. the time, exposure, and downtime definitions match the decision;
3. critical release gates are not hidden inside averages;
4. maintenance capacity includes locating, cleaning, documentation, failed checks, and validation;
5. reliability data separates random failures from field actions, configuration changes, and user-reported problems;
6. loaner and backup plans are usable in the real clinical workflow;
7. electrical safety and EMC evidence use the required setup and calibrated equipment;
8. network metrics include latency, jitter, bandwidth, outage behavior, and cybersecurity constraints;
9. lifecycle cost includes downtime, training, service, consumables, infrastructure, and end-of-life effects;
10. clinical release records state who accepted the residual risk and what evidence supports that decision.

The useful output is not a dashboard number. It is an evidence-backed decision about whether installed medical technology is safe, available, traceable, and fit for the clinical use it supports.