There are five numbers NHS England will now ask every practice in England to report against. Most practice managers can name two of them on demand. Almost none have seen all five mapped against what their own telephony and appointments data actually show. That gap is the reason the next partner meeting on access will go in circles.

The contract letter for 2026/27 set the framework. Practice-level data collection against five access metrics, sitting alongside the binding same-day urgent access requirement, with the data explicitly intended to evidence future interventions. The NHS Confederation’s own analysis is direct about it: what gets measured this year almost certainly informs what gets incentivised or mandated next year. Treat the five metrics as if they matter, because they will.

This post does three things. It names each of the five metrics in operational terms. It gives you a realistic baseline range for a typical list size so you know whether your numbers are normal, exposed, or genuinely strong. And it flags where each metric breaks down in practice, because at least two of them are not as clean as the contract letter implies.

The five metrics, named

NHS England will collect practice-level data on:

  1. Call waiting time between 8am and 10am
  2. Call waiting time during core hours
  3. Percentage of clinically urgent patients seen on the same day
  4. Percentage of non-clinically urgent patients seen within one week
  5. Percentage of non-clinically urgent patients seen within two weeks

That is the formal list. Three are appointment outcomes. Two are telephony measures. They are not weighted, they are not banded, and the contract does not yet set thresholds. What it does is create a public, comparable record of how every practice in England is performing on the front door of the NHS. The DES specification was published on 26 March 2026, and cloud-based telephony regulations are now the mechanism for capturing the call data.

Before the metric-by-metric breakdown, one piece of structural context. Clinical urgency under the 2026/27 contract is determined by the practice, not the patient and not a national algorithm. NHS England has been explicit that there is no mandated triage definition. That means metrics 3, 4, and 5 are being measured against a denominator each practice defines for itself. This sounds like flexibility. In reporting terms, it is a calibration problem the data collection cannot solve for you.

Metric 1: Call waiting time between 8am and 10am

This is the headline number. It is the one ICBs will look at first, the one that will appear in benchmarking reports, and the one that practices with no AI or queue-routing layer will struggle with most.

The 8am to 10am window is where the structural mismatch lives. A 10,000-patient practice typically sees 400 to 500 inbound calls in the first two hours of the day. Six reception staff, even fully resourced and on the phone non-stop, can handle a fraction of that simultaneously. The waiting time metric measures, in effect, how long the gap is between a patient picking up the phone and a human or system responding.

What a realistic baseline looks like:

  • Practices on legacy telephony with no call queuing or call-back functionality: average wait times in this window commonly sit between 8 and 20 minutes, with peaks of 30 minutes or more on Mondays and post-bank-holiday days.
  • Practices on cloud telephony with call queuing and call-back enabled: average wait times in this window typically run 3 to 8 minutes, though peak waits during the 8am to 8.30am sub-window can still exceed 15 minutes.
  • Practices using an AI voice layer that handles inbound calls in parallel: average wait times in this window approach zero, because there is no queue to wait in. Each call is answered the moment it arrives.

The first thing to do is pull your existing data. Cloud telephony providers report average wait time by hour. If you are on a legacy system that does not, that is itself a data point worth raising with your PCN. The reporting requirement does not come with a budget to upgrade telephony. Practices that already have hourly data have a head start, and the gap will widen.

Before I get to the next metric, one note on what the call waiting time data does not capture. The patients who abandon the call before being answered do not appear in the wait time average. They appear, if at all, in the abandonment rate, which is a separate measure most cloud telephony providers track but NHS England has not specified for collection. A low average wait time can sit on top of a 25% abandonment rate. Both numbers are worth pulling for your own baseline, even though only one is being reported.

Metric 2: Call waiting time during core hours

Core hours under the GP contract run from 8am to 6.30pm, Monday to Friday. This metric measures the average wait time across the whole day, not just the morning spike.

The reason it exists separately from metric 1 is to capture practices that have engineered the 8am wait down at the cost of everything else. Some practices route the morning spike aggressively, then run thin coverage from 11am onwards. The core hours metric flattens that out. It also catches practices that handle the morning well but have a secondary peak around lunchtime or after school pickup that gets ignored in the operational review.

Baseline ranges for a typical 10,000-patient practice:

  • Average core-hours wait, legacy telephony, no queue management: 5 to 12 minutes.
  • Average core-hours wait, cloud telephony with queueing: 2 to 5 minutes.
  • Average core-hours wait, AI voice layer handling parallel calls: under 30 seconds, often well under, because the answer is structural rather than operational.

The honest read on this metric is that for most practices the core-hours number will look much better than the 8am to 10am number. The gap between the two metrics is itself diagnostic. A practice with a 14-minute 8am-to-10am wait and a 4-minute core-hours wait has a peak-load problem, not a coverage problem. A practice where both numbers are high has a coverage problem.

Metric 3: Percentage of clinically urgent patients seen on the same day

This is the metric that ties directly to the binding contractual requirement introduced in the 2026/27 contract: requests identified as clinically urgent, as determined by the practice, must be dealt with on the same day. Patients cannot be asked to call back another day for urgent needs.

In reporting terms, this looks like a clean percentage. In operational terms, it has two complications.

First, the denominator is set by the practice. If your triage process classifies 15% of inbound contacts as clinically urgent and another practice classifies 35% of theirs as urgent using a different threshold, your numerators and denominators are not comparing the same thing. Both can hit 100% on the same-day metric and represent very different workloads.

Second, the data has to be captured at the point of triage, not retrospectively. Most clinical systems will not auto-classify an urgency level on the basis of the appointment slot used. Someone has to record the urgency decision into the system in a way that is reportable. For practices using a structured online triage tool, this is straightforward. For practices running care navigation through reception, it requires a workflow change and probably a coding agreement at practice level.

Baseline expectations for this metric vary widely depending on how the urgent denominator is defined. Practices that classify 10% to 20% of contacts as urgent and have functioning same-day capacity typically report 95% to 100% on this metric. Practices that classify 30%+ of contacts as urgent often dip to 80% to 90%, not because urgent patients are being missed but because the bar is set higher.

If you have not yet agreed an internal definition of clinical urgency with your clinical team, that is the first action this metric forces. The reporting will start from data you are already capturing, badly or well, and the quality of the numerator depends entirely on whether the urgency flag is being recorded consistently at the front end.

Metric 4: Percentage of non-clinically urgent patients seen within one week

This metric captures the practices that are managing access on routine demand, not just emergency demand. It is the one that gets harder, not easier, when same-day urgent access is the priority.

The mechanical issue is that same-day urgent slots are typically protected. If 20% of your daily appointment capacity is held for same-day urgent and that capacity is fully used most days, the remaining 80% is what serves non-urgent demand. As list size grows or staff sessions reduce, the routine wait extends first, because it has the least protection.

Realistic baselines for a 10,000-patient practice:

  • Practices with stable workforce and effective triage routing: 60% to 75% of non-urgent patients seen within one week.
  • Practices with workforce gaps, locum cover, or seasonal pressure (winter, summer holidays): 40% to 60% within one week.
  • Practices that have absorbed list growth without proportionate session growth: 30% to 50% within one week.

The metric does not distinguish between an appointment with a GP and an appointment with another clinician on the team. A pharmacist consultation for a non-urgent medication review counts. A nurse appointment for a routine review counts. This is worth knowing because the operational lever for this metric is not always more GP sessions. It is sometimes a better routing of non-urgent demand to the appropriate role.

Metric 5: Percentage of non-clinically urgent patients seen within two weeks

This is metric 4 with the window stretched. The reason it exists separately is that it gives NHS England a second data point for practices where one-week access is not feasible.

For most practices, the two-week figure is 15 to 25 percentage points higher than the one-week figure. A practice at 55% within one week typically lands at 75% to 80% within two weeks. A practice at 70% within one week typically lands at 90% or higher within two weeks. If those gaps do not appear in your own data, the data quality is probably the issue, not the appointment availability.

The combined read of metrics 4 and 5 tells you something the individual numbers do not. If both are low and close together, you have a structural capacity shortfall. If the one-week number is low but the two-week number is high, you have a peak-load problem that resolves over time. If the one-week number is high and the two-week number is only slightly higher, you have strong routine access throughout.

I will come to the operational implications in a moment. First, the metric this list does not include.

What the five metrics do not measure

Three operational realities sit outside this framework. They matter for any honest baseline.

Call abandonment is not measured. A patient who hangs up after eight minutes on hold does not contribute to the wait time average. They show up in cloud telephony reports as an abandoned call, but the data collection NHS England is running does not require that figure. A high abandonment rate is the strongest signal that the 8am wait time number is masking a worse underlying picture.

Repeat contact is not measured. A patient who called yesterday, was told to call back today at 8am, called at 8am, waited 12 minutes, was told to try again on Friday, and finally got through on Friday at 8.07am contributes one wait-time data point: the seven minutes from the successful call. The four hours of effort across the week is invisible to the metric framework.

Channel substitution is not measured. The contract requires online consultation systems to be open throughout core hours and not capped. If you have shifted demand into online routes successfully, your phone metrics improve. If patients without digital access have therefore had to wait longer for a phone slot, the metric framework does not surface that distribution.

The five metrics give NHS England a comparable view of the front door of the NHS. They do not give a complete one. The gap between what is measured and what is happening is where the operational work actually lives.

How to baseline your practice in 30 days

You cannot wait for the reporting period to know what your numbers look like. A baseline this quarter is the difference between walking into the next partner meeting with a position and walking in with a question.

Week one: Pull call waiting time data for the last quarter from your telephony provider. Split it into the 8am to 10am window and the full core-hours average. If your provider does not produce these splits, ask. Most cloud telephony platforms have this in standard reporting; legacy systems will not.

Week two: Run your appointment data for the same quarter through your clinical system. You need three percentages: clinically urgent contacts seen same-day, non-urgent seen within one week, non-urgent seen within two weeks. If your system cannot produce these splits cleanly, this is the action point. The reporting framework assumes the data is recoverable.

Week three: Cross-reference. If your same-day metric is 100% but your urgent classification rate is unusually low, the metric is meeting target on a denominator that may not survive an external audit. If your one-week routine access is 40% and your two-week is 45%, your routine demand is stalled, not just delayed.

Week four: Take the four numbers and the gap analysis to the partners. The conversation that matters is not ‘are we hitting the metric.’ It is ‘what does each number tell us about where the access constraint actually sits.’

Most practices completing this baseline find one of three patterns. A telephony problem (high wait times, urgent and routine access both functional once people get through). A capacity problem (wait times reasonable, but routine appointments slipping past one week as standard). Or a triage definition problem (numbers look fine on paper but the urgent classification is so narrow or so broad that the metric is not telling you what it claims to).

Each of those patterns has a different operational response. Knowing which one you are dealing with is the value of the baseline.

Where this leaves practices in 2026/27

The five metrics are the start of a reporting framework, not a complete one. NHS England has signalled that this data is being collected to evidence future interventions, which is professional language for ‘we will use this to set thresholds in future contracts.’ Practices that build clean baselines this year will be in the position to argue from data when those thresholds are set. Practices that wait will be benchmarked against thresholds calibrated on data they did not contribute to.

The telephony metrics in particular favour practices that can answer calls in parallel rather than in series. That is the structural shift. A reception team in series is one or two people at a time on the phone. A queue management system is people in series with a holding mechanism. An AI voice layer is calls answered in parallel, which is the only architecture that resolves the metric structurally rather than managerially.

Auxilis is one of a small number of vendors building specifically for this. If you want to see what the front-door numbers look like in a practice running an AI voice layer alongside reception, the Park Street Surgery data is worth reviewing: 81% of inbound calls absorbed, 91.1% completion rate, 51 hours of staff time recovered in 8 weeks. The 20-minute Jackie demo walks through how that data is produced and what the operational setup looks like on existing telephony.

The point of building the baseline now is not to hit a target. It is to know your own numbers well enough to make every conversation about access, about staffing, about technology investment, and about contract reporting a conversation that starts from your data, not from someone else’s framing of it.

Frequently Asked Questions

What are the five access metrics NHS England is collecting from 2026?

The five metrics are: call waiting time between 8am and 10am, call waiting time during core hours, percentage of clinically urgent patients seen on the same day, percentage of non-clinically urgent patients seen within one week, and percentage of non-clinically urgent patients seen within two weeks. The data is collected at practice level under the 2026/27 GP contract.

Does my practice have to report against the five access metrics?

Yes. The data collection is set out in the 2026/27 GP contract and applies to all GP practices in England. Cloud-based telephony regulations are the mechanism for capturing the call wait time data, and clinical system data is used for the appointment metrics.

Who decides which patients are clinically urgent?

The 2026/27 contract is explicit that clinical urgency is determined by the GP practice. There is no national triage algorithm or mandated threshold. Each practice sets its own definition, which means the same-day urgent metric is measured against a denominator the practice defines.

Are there penalties for missing the access metric targets?

Not in 2026/27. The contract letter describes the data as being collected to support understanding of demand and to encourage practices to use this data to improve services. The NHS Confederation’s analysis warns that this data will be used to evidence future interventions, so thresholds are likely in future contracts even though none are set now.

How can my practice baseline against these metrics today?

Pull call waiting time data from your telephony provider, split into the 8am to 10am window and the full core-hours average. Run your appointment data to produce three percentages: same-day urgent, one-week non-urgent, two-week non-urgent. Cross-reference the figures to identify whether you have a telephony, capacity, or triage definition issue.

What is the relationship between the access metrics and the new practice-level GP reimbursement scheme?

The £292 million practice-level reimbursement scheme is tied to the same-day urgent access requirement, which is one of the five metrics. The contract letter does not make the payment conditional on hitting metric thresholds, but the scheme and the metrics sit alongside each other as the structural change to access in 2026/27.

Does the call waiting time metric capture abandoned calls?

No. The wait time metric measures the time between a patient calling and being answered. Patients who abandon the call before being answered do not contribute to the average. Abandonment rate is tracked by most cloud telephony providers but is not part of the five metrics NHS England is collecting.