Unplanned by Design: Governing Add-On and Reflex Testing in UK Pathology

In UK pathology, a request rarely stays fixed. A clinician sees an initial result and asks for an add-on; a laboratory protocol triggers a reflex test when a result crosses a threshold. Both practices are clinically valuable and operationally efficient. Both are also poorly served by the idea that they are merely convenient extensions of the original order. The original request has already been separated from the patient, the sample and the clinical conversation. Any further testing should be governed as a deliberate next step in the diagnostic episode, not as an informal favour or an invisible rule.

Two different asks, one governance problem

An add-on is usually initiated by a clinician after the original request has been received. It may use stored material and may be communicated by phone, email, an electronic order comment or a conversation at the bench. A reflex test is different: it is a pre-agreed laboratory rule that fires when a defined result appears. Reflex testing can be automated, protocol-driven and largely invisible to the requesting clinician. Add-ons depend on communication and local knowledge.

The distinction matters because the risks are not identical. Add-ons are vulnerable to sample stability limits, ambiguous clinical questions and unclear ownership of the new result. Reflex rules are vulnerable to version drift, unexamined assumptions and uneven application across sites. A pathology network may believe it has one reflex policy while several laboratories are running different versions of it. Neither problem is solved by telling staff to be more careful. Both require visible rules, named owners and routine audit.

Why convenience is a weak safety case

A test that is easy to add is not necessarily safe to add. The first question is whether the stored sample is still fit for the new purpose. Stability windows vary by analyte, sample type, storage temperature and transport history. An add-on requested late on a Friday may be analytically possible but clinically misleading if the sample has exceeded its validated window. Laboratories need a simple, accessible way to confirm suitability before accepting the request, and a documented route for declining it when the sample no longer supports the question.

The second question is communication. An add-on creates a new result, and a new result creates an obligation. Who receives it? Who acts on it? Who tells the patient? If a general practice requests an add-on to a sample originally sent by a hospital clinic, the result may return to a different team from the one that ordered it. If the request arrives verbally, the laboratory may have no reliable record of the clinical question. The result can then be filed without anyone owning the next step.

The third question is equity. Add-on access often depends on knowing how the system works. Practices with experienced pathology links, stable staffing and good local relationships may find it easier to ask. Smaller practices, remote sites and patients with less continuity of care may not. Reflex rules can standardise access, but they can also encode assumptions that work less well for some populations. A rule that is never reviewed is not neutral; it is simply unexamined.

A decision gate before the test is added

UK laboratories should maintain a register of approved add-ons and reflex rules. The register need not be bureaucratic, but it should record the clinical question, the sample requirements, the stability limit, the result route, the rule owner, the authoriser and the review date. Before a new add-on is accepted, the laboratory should be able to answer six questions.

First, does the add-on answer a defined clinical question that cannot be answered from the existing result? Second, is the sample type, volume and storage history suitable? Third, is the result route and clinical responsibility explicit? Fourth, what is the impact on turnaround time and capacity? Fifth, is access to the add-on equitable across service users, or is there a documented reason for variation? Sixth, what will be audited, by whom, and when?

For reflex testing, the same discipline applies with additional emphasis on version control. Every rule should have a named clinical owner, a stated trigger, a defined action and a review date. When an assay, platform or clinical guideline changes, the rule should be re-approved rather than allowed to drift. The rule should also be visible in the laboratory information system and in order communications, so that requesting clinicians know when a reflex has occurred and why.

Who authorises what

Authority should be tiered. A routine add-on menu can be pre-approved by the laboratory clinical lead and service users, with clear limits on sample age and test type. Exceptional add-ons should require clinical lead review. Reflex rules should go through change control, with sign-off from the laboratory clinical lead and, where capacity or commissioning is affected, the relevant service commissioner. In an NHS pathology network, a rule that applies in one laboratory but not another should be explained rather than left to local habit.

Making the rules visible

Visibility is a safety control. Reflex rules should be published in the laboratory handbook and in electronic order sets. When a reflex is applied, the report should state the trigger and the rule version. Add-on menus should be available to all service users through a single route, not through informal contacts. Result comments should explain what was added and why. Dashboards should show add-on rates by requesting location, reflex trigger rates and turnaround times. Variation is not automatically a problem, but unexplained variation deserves attention.

Audit the second question

Audit should cover sample stability failures, add-on turnaround, duplicate testing, result acknowledgement and the proportion of add-ons that lead to a clinical action. For reflex rules, useful measures include trigger rate, positive yield, override rate and the number of results that required clarification. Incidents and near misses should be reviewed alongside complaints. If an add-on cannot meet its stated criteria, it should be retired or require a fresh sample. If a reflex rule produces low-value results, it should be revised rather than defended.

A minimum viable playbook

Start with a register. Classify each add-on and reflex rule by clinical risk and operational complexity. Assign an owner and an authoriser. Document sample stability, result routing and communication. Publish the rules where requesters can see them. Audit a small number of meaningful measures and review them on a fixed cycle. Finally, make it easy to say no when the sample or the clinical question does not support the test.

Add-on and reflex testing are not peripheral conveniences. They are ways of extending a diagnostic episode after the original request has been made. Governed well, they improve care and reduce unnecessary repeat sampling. Governed badly, they become invisible safety cases. The task for UK pathology is to keep the convenience while making the rules, responsibilities and limits visible.