How to respond to a hospital power outage

Find out which power has failed and who is at risk

The lights fail in one ward, then return on some circuits. A generator may have started, but the hospital's main supply is still down. Another department reports working sockets yet no network connection. The first minutes are about protecting patients and building an accurate picture, not declaring the incident over because emergency power is visible. The team needs to know which areas have lost power, which backup sources are supporting them and how long the current arrangement can safely continue.

Raise the electrical and clinical incident routes immediately. NHS England HTM 06-01 says plans should exist for unexpected electrical infrastructure failure and that staff need a way to raise immediate awareness. Its response model calls for people who understand patient and staff risks, the electrical network and available restoration resources. A ward should not have to diagnose a switchboard. It should report what it sees and what care is affected while authorised engineers investigate the supply.

Ask clinical areas for urgent patient risks. Are life-supporting devices still powered? Has a device switched to battery, and how much verified autonomy remains? Is a procedure underway? Are infusion, monitoring, ventilation or suction services affected? Are patients in lifts or behind powered access doors? A clinical leader should coordinate immediate care and device-specific backup according to local procedures. This guide does not prescribe a treatment change or assume that every device has the same battery capacity.

Keep staff from unsafe electrical improvisation. Do not move equipment to a socket simply because it is coloured red without confirming that the circuit is live and suitable. Do not daisy-chain extension leads, defeat an alarm or connect a hired generator through an ad hoc cable. Authorised electrical personnel should control switching, isolation and restoration. Clinical teams should identify the care they must maintain, not alter the electrical installation themselves.

Record the incident start and the first reliable observations. A hospital may have several simultaneous states: primary supply lost, one standby generator online, another section on UPS, and some ordinary circuits off. Label each area by verified status. Uncertainty is acceptable if it is visible. A single statement that “the hospital is on generator” can conceal unsupported departments and create false reassurance.

Build an affected-service picture across the site

Engineering should map the incoming supply, switchgear, secondary sources and affected distribution. Did the external network fail or did an internal high-voltage component trip? Did automatic changeover occur? Are generators carrying the expected loads? Are there alarms, fuel issues or a protection device that has isolated part of the network? Keep the electrical investigation within the competence and safety rules of the organisation. The clinical incident team needs a concise result, not every technical detail.

Clinical services should report more than whether the lights are on. A ward may have bedside power but lose nurse call, electronic records, ventilation or refrigeration. An operating area may have critical circuits but no normal lighting in corridors. A laboratory may be unable to receive samples. A pharmacy may have cold-chain capacity but no dispensing system. The incident picture should name each service, what is unavailable, what temporary process is operating and how long that process remains safe. Recheck when loads change or another source fails.

Prioritise the most time-sensitive patient dependencies. Clinical engineering and the care team can identify equipment batteries, available replacements and any manufacturer instructions. The estates team can identify supplied circuits and power sources. Together they can decide whether a patient can stay in place, needs support equipment, or should move. Do not infer clinical safety solely from the engineering statement that a circuit is energised. A device might have restarted incorrectly or lost a setting and require its own check.

Include the services that support care. HVAC, water pumps, medical gases, sterilisation, lifts, kitchen, security, telecoms, network rooms and fire systems may depend on power. Each has its own failure mode and specialist owner. A facilities status board should show the consequence for care and operations, not just an asset name. A working generator does not guarantee that every dependent service has transferred or that its controls remain stable.

The NHS England electricity-failure case study shows why the distinction matters. In that event, emergency generator power remained, yet the hospital could not accept elective or emergency patients while the internal mains problem was investigated. The case describes command arrangements, engineering support and service changes. It is an example, not a rule that every outage requires the same cancellation or briefing interval.

Establish incident command and service decisions

Set a clear incident lead and a technical lead. The incident lead coordinates patient, service, communication and external effects. The authorised electrical team controls electrical investigation and restoration. Clinical leads decide whether treatment, admissions, transfers or appointments can continue. Record who can make each decision and when it will be reviewed. A power failure often changes faster than a routine maintenance issue, so a stale decision can become unsafe.

Activate the NHS emergency preparedness or local business continuity plan at the appropriate level. NHS England EPRR guidance addresses the need to plan for incidents that affect patient care. Its evacuation and shelter guidance includes utility failure as a possible reason to consider shelter or evacuation. Neither source makes evacuation automatic for every power loss. A hospital may remain safe with its designed secondary supply for a period, while another may need to move patients from a particular area. Use the real affected-service map and patient acuity.

Choose and document service restrictions. Admissions may be diverted, procedures postponed, outpatient clinics paused or a department moved. An NHS case study shows a hospital on generator power cancelling outpatient appointments while the fault was resolved. The decision is local and clinical. Keep a record of who made it, why, which patients were affected and how they will be contacted or rescheduled. A status message should not simply say “business continuity in place” without telling teams what work can proceed.

Plan for the failure of the backup arrangement itself. What happens if the generator trips, fuel delivery is delayed or UPS batteries approach their limit? Which area loses supply first? Where can patients be moved safely, and who can receive them? A contingency that begins only after the secondary source fails is too late for a complex transfer. The clinical and technical teams should agree trigger points for escalation while the current source is still functioning.

Maintain a regular common operating picture. Include primary supply status, secondary source state, affected clinical services, patients at highest risk, repair estimate, decision triggers and next briefing time. Issue a versioned short message to wards, porters, contractors, ambulance interfaces and call centres as appropriate. Avoid presenting an engineer's estimate as a guaranteed restoration time. Update the picture when the fault diagnosis changes.

Control temporary power and downtime procedures

A mobile generator or temporary connection may help, but it needs a designed and authorised connection. HTM 06-01 warns that mobile generators require attention to earthing and impedance and that positive isolation from the public supply is essential in relevant configurations. Competent engineers must check capacity, protection, connection, load and safe switching. A hired generator delivered to the car park is not operational resilience until these checks and its fuel, exhaust, access and monitoring are resolved.

Use manual or paper downtime processes only where they have been prepared and clinical governance has approved them. The NHS England electricity-failure case study highlights awareness of IT downtime processes and paper procedures. Keep patient identification, prescribing, observations, referrals and results handovers safe. Record what was done during downtime so the electronic record can later be reconciled. An outage may affect servers, local devices or the network differently. A paper process that works for one service may not cover another.

Preserve medicines and specimen controls. If refrigerators, freezers or laboratory systems lose supply, identify the actual temperature record and stock involved. Quarantine uncertain stock where required and use the relevant pharmacist or laboratory assessment. A power outage does not automatically spoil every medicine, and a fridge display returning to normal does not clear an excursion. Link the electrical event to the product-specific incident record without combining all stock into one unsupported decision.

Check fire and building systems. Emergency lighting, fire alarms, smoke control, powered doors and lifts may each have separate backup arrangements. Fire safety and estates specialists should assess any impairment and temporary occupation decision. Do not infer that these systems function because clinical sockets do. Where a fire system is impaired, use the site's fire safety procedures and communicate restrictions to all occupiers. This power-outage page does not replace the separate fire alarm impairment guide.

Check the edges of the affected area

The first map often omits the boundary between a powered and unpowered service. A ward may use a working treatment room but share a specimen refrigerator, lift, data connection or medicine store with an affected department. Ask neighbouring services what they depend on, and trace any temporary workflow across that boundary. If patients are moved to another floor, confirm that the receiving area has capacity, staff, equipment and the same clinical information. If a laboratory cannot accept specimens, tell collection points before samples accumulate. If a reception team switches to paper registration, define how those records will be matched later. These handovers can cause harm even when the electrical repair itself is straightforward. Record the owner of each boundary and its next status check.

Restore supply in controlled stages

An engineer may repair a cable or switchboard, but the service does not instantly return to normal. The authorised electrical person should identify the fault, complete safe switching and verify the intended supply paths. Check whether generators and UPS systems transfer back correctly. Confirm alarms, protective devices and distribution status. Record any circuit that remains isolated or under temporary supply. A green main-power indicator is not a site-wide clinical release.

Plan the return so it does not create a second interruption during an active procedure. HTM 06-01 notes that some generator testing and power restoration arrangements involve a short interruption depending on the design. The electrical team should explain expected switching effects before transferring loads. Clinical leaders can choose an appropriate time or maintain temporary restrictions until the transfer is complete. Do not leave secondary sources running indefinitely without a plan for fuel, maintenance and safe restoration.

Verify dependent equipment and services. Clinical engineering may need to check devices that lost power. Laboratory, imaging, sterile services and pharmacy teams may have their own restart validation. IT teams should confirm systems and reconcile paper records. Estates should check ventilation, water pumping, access controls and lifts. Each service owner should state when it can operate normally. The incident lead can then lift restrictions in stages and tell affected staff and patients what changed.

Keep technical and clinical release distinct. Engineering signs off the electrical system. Clinical and operational leaders decide when patient services resume, based on their own checks. A hospital might have stable power yet still need to reschedule a procedure because equipment is being validated or patient records need reconciliation. Record both release times. This prevents a maintenance ticket from being treated as proof that all clinical work was safe to restart.

Reconcile the event and prevent recurrence

The final event record should contain a time line of the initial loss, source changes, affected services, patient actions, command decisions, contractor work, repair evidence, staged restart and outstanding restrictions. Preserve alarm logs, switching records and the approved affected-load map. Keep patient-level detail in clinical systems and link references from the operational record as needed. A broad compliance system should not expose confidential patient information to everyone who can view estates actions.

Review the cause, including why resilience did or did not work. Was the failure external or internal? Did secondary power transfer as designed? Were clinical teams aware of its limits? Did a known defect or maintenance backlog contribute? Did communication reach night staff and ambulance partners? Did a device restart require an overlooked check? The answer can lead to engineering, clinical, IT and training actions. One generic action such as “review policy” will not address an electrical single point of failure or an untested patient transfer plan.

Test corrective action with evidence. A replaced component needs technical tests; a revised switch procedure needs a controlled exercise; a changed service communication plan needs a handover test. If the backup plant ran close to capacity, update the load model and contingency plan. If paper processes were incomplete, rehearse them in the actual ward. Set owners, dates and an effectiveness review. Do not close an action because a document was uploaded when the practical control has not been tried.

Use the incident to improve the next response. HTM 06-01 frames electrical safety as a system with governance, incident protocols and planned maintenance. The NHS case study demonstrates that maintaining generator power may still leave major service disruption. A useful exercise should therefore ask which clinical services can continue, which must pause and how those decisions are communicated. The aim is a hospital that understands its real power dependencies before the next failure.

Where Complys may fit

The Complys hospitals page describes healthcare compliance records. The live healthcare facilities checklist mentions utilities and business continuity. An administrative evidence layer may help hold electrical certificates, maintenance actions, contractor records and post-incident reviews if the actual product supports those workflows. The public pages do not prove live power telemetry, clinical command, patient transfer planning, electrical switching or automated service release. Those functions remain with the provider's specialists.

A practical evidence design could separate technical fault status, affected-service decisions, clinical restrictions, restart checks and final closure. It should identify the decision maker and supporting document for each stage. This is an editorial proposal, not a verified Complys feature. The product owner should review permissions, audit history and data limits before making any capability claim. Patient-level records should remain in the appropriate clinical system.

CTA: Review Complys for hospitals and ask for a demonstration of the estates and action evidence that the product currently supports. Related tool opportunity: A power-outage common operating picture worksheet with affected area, primary and secondary source, critical devices, service decision, next review, engineering release and clinical release. It must not calculate clinical safety automatically. Suggested internal links out: Standby generator failed test, healthcare water supply interruption, fire alarm impairment, Complys hospitals. Suggested links in: Hospitals commercial page, broad healthcare facilities checklist and generator failed test guide when a real outage follows a known defect. Cannibalisation boundary: This page owns an actual unplanned loss of primary or internal electrical supply and immediate patient and service response. Draft 64 owns degraded standby resilience discovered during a planned test while primary power may still be present. Draft 51 owns emergency-lighting fault response in an occupied workplace. Do not multiply by ward, city, device or power-source brand.