After the Switchboard Sleeps: Auditing Overnight Microbiology Decision-Making

The gap between the rota and the reality

Every UK microbiology laboratory has an out-of-hours arrangement. It may be a formal on-call rota, a shift system, a managed network across several sites, or a hybrid model in which a small number of staff remain on site while others provide remote advice. The rota answers a workforce question. It does not, by itself, answer the clinical question that wards actually ask: what can this service do tonight, how quickly, and with what level of confidence?

That distinction matters more as winter 2026/27 approaches. Respiratory demand, reduced elective activity in some trusts and the familiar pressure on beds can all sharpen expectations of a laboratory that is running with fewer people than it had during the working day. The risk is not that out-of-hours cover is absent. The risk is that it is assumed to be more comprehensive than it is.

An audit of out-of-hours microbiology should therefore begin with capability, not simply with presence. It should make the difference between available and safe to promise explicit, and it should give the trust a defensible account of how urgent results are generated, escalated and acted upon outside core hours.

Start with scope, not staffing

The first audit question is deceptively simple: which microbiology activities are genuinely offered overnight and at weekends? A laboratory may provide Gram stains and urgent culture reading, but not full identification or susceptibility testing. It may offer a telephone interpretation service but not a formal report. It may accept certain specimen types but refer others to a regional centre. Each of these boundaries is legitimate, provided it is known.

The audit should map the out-of-hours service against a small set of activity categories: urgent Gram stains; critical result notification; blood culture positivity and preliminary organism information; urgent molecular testing where commissioned; environmental or outbreak-related requests; and advice on antimicrobial therapy or infection control. For each category, the laboratory should record the stated offer, the actual cut-off time, the expected turnaround and the route by which a result reaches a decision-maker.

This map is not a performance league table. It is a boundary document. It allows the laboratory to say with confidence what the ward can expect, and it exposes the points at which a local assumption has outrun a local capability.

Four questions an audit should answer

A useful out-of-hours audit does not need to be elaborate. It needs to be repeatable. Four questions provide a workable structure.

First, what is the demand? Count calls, specimens and escalations by hour, day and specimen type over a defined period. A four-week sample is usually enough to reveal the shape of the service, provided it includes at least one bank holiday or period of unusual pressure.

Second, how quickly does the laboratory respond? Record the time from receipt of an urgent request to the first clinically useful communication. This is not the same as the time to a final report. A Gram stain result telephoned at 02:15 may be more valuable than a completed culture report at 09:00.

Third, who receives the message, and what happens next? A response time without a communication trail is only half a measure. The audit should sample calls and check that the recipient was identified, that the result was read back, and that the clinical action was documented in the patient record or the laboratory system.

Fourth, what does the morning handover carry? Overnight decisions often depend on context that is difficult to reconstruct at dawn. The audit should test whether unresolved issues, pending results and safety concerns are passed on in a structured way.

Measuring response without reducing it to a stopwatch

Response times are necessary but insufficient. A laboratory can meet a target for telephoning a critical result while still leaving the ward uncertain about what to do next. Conversely, a slightly slower call that includes clear interpretation and a named contact may be safer than a faster one that ends with a number.

The audit should therefore pair timeliness with quality markers. These might include whether the caller identified themselves and their role; whether the result was communicated in plain language; whether the receiving clinician was authorised to act; whether advice was escalated when the laboratory could not answer; and whether the interaction was recorded in a consistent place.

It is also worth distinguishing between critical result notification and urgent clinical advice. The former is a defined laboratory output. The latter is a conversation. Both matter out of hours, but they require different evidence. A trust that audits only notification times will learn little about the quality of overnight decision-making.

The communication trail is a clinical control

Out-of-hours microbiology is unusually dependent on the telephone. That makes the communication trail a clinical control, not an administrative nicety. The audit should sample the path from bench to ward and look for four things: a documented time, an identified recipient, a read-back or acknowledgement, and a record of any advice given.

Where electronic result acknowledgement exists, it should be checked for completeness rather than assumed to be reliable. Where it does not, the laboratory should be honest about the limits of the paper or telephone record. The aim is not to create a perfect archive. It is to ensure that a disputed decision can be reconstructed, and that a missed call can be recognised before it becomes a missed treatment.

This is particularly important for Gram-stain urgency. A Gram stain is often the first microbiological signal available outside core hours, and its value depends on reaching someone who can act on it. An audit that counts Gram stains but does not trace their communication is measuring activity, not safety.

Handover as a clinical control

Handover is where out-of-hours work either becomes continuity or disappears. A structured handover should capture pending critical results, unresolved queries, infection control concerns, equipment failures and any agreed follow-up. It should also record who owns each item.

The audit can test handover in two ways. First, review a sample of handover records against a standard template. Second, ask receiving staff whether the information was sufficient to continue the clinical conversation. The second question is harder to quantify, but it is closer to the purpose of handover than a tick-box review alone.

Staffing models and safe limits

Different staffing models are appropriate in different settings. A large teaching hospital may sustain a dedicated overnight shift; a smaller district general hospital may rely on an on-call scientist supported by a regional service. Neither model is inherently better. What matters is that the model is matched to the commissioned activity and that its limits are explicit.

The audit should record the model in use, the expected response time, the escalation route when the first contact is unavailable, and the arrangements for cover during leave or sickness. It should also capture the point at which the service becomes unsafe to extend — for example, when a single on-call scientist is simultaneously handling a blood culture alert, a Gram stain and a ward enquiry. That is not a staffing complaint. It is a capacity boundary that the trust should see.

Reporting without blame

The output of the audit should be a short, neutral report to the laboratory and the trust: what is offered, what is measured, what is working, and where the gap lies between ward expectation and laboratory capability. It should avoid turning individual calls into a blame exercise. The more useful question is whether the system makes the right action easy at 03:00.

Out-of-hours microbiology will always involve compromise. The purpose of auditing it is not to eliminate compromise, but to make it deliberate, documented and clinically defensible. A service that knows its own boundaries is easier to trust than one that leaves the ward to guess.