GP practice governance guide

How to respond to a backlog of GP test results

When a practice finds test results that have not been reviewed or actioned, the task is to establish the real scope, place the queue under clinical control, arrange competent triage, contact patients, reconcile anything missing and prove the backlog has genuinely cleared. This guide covers GP practices in England, where services are regulated by CQC.

Establish what the backlog actually contains

A manager sees 80 unprocessed pathology messages. A partner is on leave and their personal inbox has not been covered. A locum ordered investigations but no one can find the results. A reception team has filed routine-looking messages without a documented clinical review. These are different failure modes. They share one immediate need: identify which patient results may not have received a timely, competent decision, and place the queue under clinical control.

CQC’s GP guidance on managing test results and clinical correspondence asks how a practice checks that ordered tests return, processes results during clinician absence and follows up abnormal findings. It also asks how the patient is informed and the outcome recorded. A queue counter cannot answer those questions. A result may be opened but not clinically acted on, or marked complete while the patient has not received the instruction that follows from it. CQC: GP mythbuster 46, managing test results and clinical correspondence.

The clinical lead should arrange a rapid assessment of the queue. Preserve its current state, including total items, oldest date, recipient inbox, message type and any items flagged urgent by the sending laboratory or service. Avoid a bulk “mark as read” action to make the count smaller. Do not let an administrator classify clinical significance from a subject line or colour alone. A qualified clinician should set the clinical triage approach and identify who can safely review each category. Escalate urgent findings according to local clinical pathways while the wider reconciliation continues.

A CQC inadequate-practice example describes a practice reporting daily action while inspectors found 57 unchecked pathology results from the previous two days, including 12 with abnormalities. The example illustrates why a verbal reassurance or dashboard total is not enough. It does not establish a universal safe backlog size or a clinical time limit for every test. The risk depends on the result, patient and intended follow-up. CQC: inadequate-practice example, effective clinical care.

This page owns a specific incident response when a test-results pathway has failed. It does not advise on diagnosis or prescribe a fixed triage algorithm. Those decisions belong with a qualified clinical team using the actual result and patient record.

Stop new results entering the same failure

While the old queue is being investigated, new results may continue to arrive. Assign a named duty clinician and operational coordinator for current incoming work. Check all routes, not only the inbox where the backlog was found. Electronic pathology, imaging reports, paper post, electronic discharge letters, shared NHSmail inboxes and direct clinician messages may each have their own worklist. A missed result problem can grow during a cleanup if staff focus only on the historic queue.

Map every active destination. For each, record the system, type of result, normal recipient, cover arrangement, review method, escalation route and daily confirmation that the queue was checked. Where a personal mailbox is involved, verify what happens during annual leave, sickness, a locum’s departure or a changed role. A forwarded message is not necessarily visible in the clinical record. A message that appears in the patient record may still lack a clinician’s decision. Establish how a result moves from receipt to action to patient communication.

NHS England’s NHSmail guidance recommends monitoring inboxes for clinical tasks, checking undelivered mail, managing leavers and temporary staff, and maintaining a business continuity plan if NHSmail is unavailable. It also says shared practice inboxes should not be used for urgent clinical advice. This is relevant where the backlog involves correspondence, but it does not mean email is the only route for pathology results. The local system map must reflect the practice’s actual integrations and contracts. NHS England: NHSmail guidance.

Do not solve the capacity problem by giving untrained people unrestricted authority to file results. CQC says staff who review results should have appropriate access, training, supervision and support. Non-clinical staff may perform certain tasks with safeguards, while clinicians remain responsible for acting on results that alter patient management. The practice should specify which tasks are administrative and which require a clinician. If the normal distinction has been blurred, pause the risky delegation until the clinical lead can assess it.

Triage the historic queue with a documented method

The clinical team should set a prioritisation method that accounts for known urgent flags, test type, age, patient context, prior request and whether a patient has already been contacted or treated. Do not assume the oldest item is always the most clinically urgent. Equally, do not clear new abnormal flags and leave older unreviewed items indefinitely. A qualified reviewer needs enough information from the patient record to decide what action is needed.

Record the backlog baseline and the review allocation. A useful operational worksheet lists result identifier in the clinical system, received date, requesting clinician, responsible reviewing clinician, review status, action status, communication status and unresolved dependency. Keep patient identifiers within the approved clinical system or a secure, approved clinical worklist. A general compliance spreadsheet should not become a shadow patient record. If a management action tracker needs evidence, reference a controlled incident and aggregate counts, not raw clinical details.

Set a handover process for results requiring a second opinion or follow-up. A clinician may review a laboratory value but still need to order a repeat test, contact a specialist, change treatment or speak to the patient. Marking the result reviewed should not close the action. The queue should distinguish reviewed with no further action, reviewed with action assigned, patient contact pending, clinical appointment pending, and completed. Each open state needs an owner and review date chosen on clinical grounds.

Check linked results. CQC warns that multiple tests ordered for one patient may return at different times, increasing the chance that one result is missed. The reviewer should see whether a partial set changes interpretation or whether an expected result has not arrived. A normal result in one component should not make an absent or abnormal component disappear. If the practice has received a report but cannot find the original request, establish who ordered it and which service owns follow-up. Do not assume that the external laboratory or hospital will contact the patient.

Where the backlog is large, plan reviewer capacity and quality checks. A target number of items per hour can encourage shallow filing. Clinicians should identify when complex results or uncertain ownership slow the process, and the coordinator should adjust staffing rather than press for a cosmetic queue reduction. Document the number of results reviewed, those still open, clinical escalations and any patient communication still due at each handover. A second clinician or quality lead may sample completed items to test whether the action was appropriate and recorded.

Reconcile missing and misrouted results

A visible backlog is only one part of the risk. The practice must ask whether ordered tests returned at all. CQC’s GP guidance explicitly includes tracking requests, following up non-attendance, checking when results return and acting when they do not. A cleared incoming queue cannot prove that every requested test was received. Build a reconciliation for the affected period by comparing order or request data with returned reports, where systems permit. Investigate missing or mismatched items with the sending organisation and the original requester.

The affected period may not start on the date someone noticed the backlog. Review staffing changes, inbox routing, system upgrades, interface failures, administrator permissions and prior audit findings. If a clinician left three months ago, their outstanding work may have been stranded. If a laboratory changed a message format, automatic routing may have stopped. If an inbox was hidden by a filter, staff may have been checking the wrong view. Establish when the control last worked and what evidence supports that boundary.

Do not close an item because a patient app shows a result. CQC says patients may be able to view results online, but the practice must still have a safe system for review, action and communication. Seeing a number without clinical context does not ensure that the patient understands what to do. Conversely, a patient may choose not to use online access. The practice needs an active communication policy that fits the result and the patient.

NHS England’s online record access guidance highlights that practices should consider the impact of new information, including test results, as it enters the record. A backlog response must not rely on withholding data from the app as a substitute for clinical contact, or on app release as proof that contact occurred. The clinical team should decide any appropriate communication safeguards in the approved record system. NHS England: online access to new GP health record information.

Communicate and follow up with patients

For each result needing action, a clinician should decide the urgency and message. The practice should use an agreed contact method, verify identity where needed and record what was said, sent or attempted. If a patient cannot be reached, the next step depends on clinical risk. A single unanswered text is not always a completed follow-up. CQC says practices should have a method to ensure the patient is aware of the result and expected actions, and should not rely only on patients contacting the surgery when results require action.

Reception staff need a clear script and escalation route. They should not be asked to explain a complex or sensitive finding from an ambiguous note such as “please inform patient.” A documented clinician instruction should state whether the patient needs an appointment, a repeat test, urgent assessment or other follow-up, without turning the receptionist into the clinical decision-maker. If the patient raises new symptoms or asks for interpretation, staff should know how to transfer the question to the clinician.

Review accessibility and confidentiality. A patient may need an interpreter, a different communication method, a carer involved with appropriate authority, or a private call rather than a text visible to others. A practice-wide message saying that some test results are delayed may be useful during disruption, but it does not discharge the need to act on a specific abnormal result. Document unsuccessful contact attempts and clinical decisions about further action. Avoid sending a generic reassurance before a qualified reviewer has assessed the individual result.

Where the delay may have caused harm or changed care, the clinical and governance leads should consider patient safety processes, candour and CQC notification on the facts. Not every delayed result reaches a statutory notification threshold. A late result is also not automatically harmless because the patient has not complained. The assessment should examine the actual clinical course, what action should have occurred, what did occur and whether a patient needs follow-up now. Seek appropriate advice rather than making a legal threshold decision from an operational queue count.

Restore a reliable service, then test it

Clearing the old queue is not proof that the system is fixed. The practice should identify the failure mode and set a control against recurrence. Examples include shared inbox coverage, a daily exception report, results-to-request reconciliation, locum handover, clear delegation limits, timed review of unprocessed items and escalation when reviewer capacity is exceeded. A control should have an owner, a working instruction, a backup and evidence that it ran. “Staff reminded” may be part of an action, but rarely resolves a broken routing rule or absent leave cover.

Test the new process with real examples. Select recent requests and follow them through ordering, receipt, clinical review, patient communication and final closure. Include a result that was normal, one that required action, a multi-part request and a case where the original clinician was unavailable. Check a result that never arrived and ask whether the practice would detect and follow it up. The sample should be chosen by the clinical governance lead and extended if it reveals new gaps.

CQC’s GP guidance says practices should audit the system regularly to check its effectiveness. Document the audit question, sample period, exceptions found, action owner and follow-up test. A one-off audit after an incident can show immediate repair, but the practice also needs ongoing monitoring. The interval and method should reflect local risk and system design rather than a universal number invented for an article. CQC: GP mythbuster 46, managing test results and clinical correspondence.

Review clinical correspondence alongside pathology. Discharge letters and referrals may carry medicine changes or follow-up instructions, and CQC says similar safeguards apply. If the backlog arose from shared staffing, mailbox or workflow controls, it may affect these other queues too. Investigate them on their own facts. Do not combine all items into a single letters-and-results count that hides the higher-risk types.

A worked example of a stranded locum queue

A six-week locum finishes on Friday. On Tuesday, the practice discovers that results addressed to the locum remain in a personal worklist. Thirty-one items are visible. The clinical lead assigns two regular clinicians to review the queue and asks the practice manager to preserve the starting count and dates. A coordinator checks other locum and former-employee worklists. New results are routed to the current duty clinician while the practice investigates the handover rule.

The reviewers find that eight items need patient contact, three need repeat tests and one needs urgent clinical review. They identify two requested tests whose reports are absent. Reception follows clinician-approved instructions for contact, with a route back to the reviewing clinician for questions and failed contact. The missing results are chased with the sending laboratory and the patient records are checked for later reports. The practice does not mark all thirty-one as closed when the first review is complete because several patient and laboratory actions remain open.

The root cause is a leaver process that disabled the locum’s access without transferring outstanding clinical tasks. The practice changes the leaver checklist so a named clinician accepts open work before access ends. It also tests the shared cover process during a planned absence. A later audit samples recent locum requests and confirms each result returned and was acted on. This example shows how a visible queue can lead to an owner, a patient action and a control change. It is not a triage algorithm for actual patients.

FAQs

Is there a safe backlog size or a universal clearance time?

No. CQC guidance does not prescribe one workflow or a universal clearance time, and the inspection examples are illustrative rather than a threshold. The risk depends on the specific result, patient and intended follow-up, so a qualified clinical team sets the triage approach using the actual record.

Can a bulk 'mark as read' clear the queue?

No. A queue counter cannot show whether a result was clinically acted on or whether the patient received the instruction that follows from it. Preserve the current state, avoid a bulk mark-as-read that only shrinks the count, and have a clinician review each item.

Does a result showing in the patient app mean the item is closed?

No. Patients may view results online, but the practice must still have a safe system for clinical review, action and communication. Seeing a number without clinical context does not tell a patient what to do, and a patient may not use online access at all.

What does Complys do in a test-results incident?

Subject to product verification, Complys can hold a non-clinical incident reference, named process owners, the updated standard operating procedure, training evidence, audit dates and proof that corrective actions were tested, along with access controls. It is not a GP clinical system, pathology inbox, patient record or result triage engine, and clinical results do not go into it. Clinical review, prioritisation and patient contact stay in the approved GP system with the clinical team.

Where Complys fits

Subject to product verification, Complys can hold a non-clinical incident reference, named process owners, the updated standard operating procedure, training evidence, audit dates, access controls and proof that corrective actions were tested. It should not be described as a GP clinical system, pathology inbox, patient record or result triage engine, and clinical results do not go into it. Patient-specific results, clinical instructions and contact notes stay in the approved clinical environment, and clinical review, prioritisation, patient contact and any notification decision remain with the clinical and governance leads.

Clinic compliance software →Incident reporting →