# How CQC scores a GP practice, worked through Published 6 October 2026 in Single Assessment Framework by CQC AI Auditor team. Source: https://www.cqcaudit.ai/blog/cqc-scores-gp-practice-explained **In short:** CQC scores each quality statement it assesses from 1 to 4, adds the scores up for each of the five key questions, turns the total into a percentage, and converts that percentage into a rating using fixed thresholds. Two limiting rules stop a single weak statement from being hidden by good ones. This post walks through each step with CQC's own GP practice example, shows how a first assessment uses your existing rating for the statements CQC does not look at, and explains what CQC has said about removing scoring in future. Most practice managers know their rating. How CQC arrives at it is less well known, partly because the CQC scoring method is spread across several guidance pages. The arithmetic is simple once it is laid out, and it explains things that otherwise look odd: why one statement can change a key question rating, why a practice assessed on three statements still gets a score for all eight, and why two practices rated good can be in very different places. Everything quoted below is CQC's wording, from pages read on 6 October 2026. The worked examples that are ours are marked as illustrative. ## Step 1: Each quality statement gets a score from 1 to 4 CQC's [How we reach a rating](https://www.cqc.org.uk/guidance-regulation/providers/assessment/assessing-quality-and-performance/reach-rating) page sets out the scale. Each quality statement CQC assesses gets one of four scores: - **1:** "Evidence shows significant shortfalls" - **2:** "Evidence shows some shortfalls" - **3:** "Evidence shows a good standard" - **4:** "Evidence shows an exceptional standard" The score is a judgement, not a tally. CQC says inspectors "use professional judgement to assign the score, based on the evidence categories they have looked at for each quality statement." In its GP practice example for infection prevention and control, the evidence includes patient surveys and complaints, a visit to look at the care environment, a call with staff, and the practice's "cleaning, hygiene, and infection control policies and records". CQC then says: "We consider all this evidence to produce a balanced judgement. Based on that judgement, we apply the appropriate score to the quality statement." One change worth knowing: since 2 December 2024 CQC has stopped scoring the individual evidence categories. In its words, "we now only score at quality statement level." > If you want the 34 statements themselves, with the evidence CQC usually looks at for each, our [guide to the CQC quality statements for GP practices](https://www.cqcaudit.ai/blog/cqc-quality-statements-gp-practices-evidence-examples) lists all of them in CQC's wording. ## Step 2: The scores become a percentage for each key question CQC adds up the statement scores under a key question and divides by the maximum possible. The maximum is "the number of quality statements under the key question multiplied by the highest score for each statement, which is 4." Safe has eight quality statements, so the maximum for Safe is 32. CQC's own worked example for a GP practice looks like this: - **Learning culture:** 2 - **Safe systems, pathways and transitions:** 3 - **Safeguarding:** 3 - **Involving people to manage risks: **2 - **Safe environments:** 3 - **Infection prevention and control:** 3 - **Safe and effective staffing:** 2 - **Medicines optimisation**: 3 The total is 21. CQC's calculation: "a percentage score for the key question of 65.6% (this is 21 divided by 32)." The other key questions work the same way with their own maximums: Effective has six statements (24), Caring five (20), Responsive seven (28) and Well-led eight (32). ## Step 3: The percentage becomes a rating At key question level, CQC turns the percentage into a rating "rather than a score", using these thresholds: - 25 to 38% = inadequate - 39 to 62% = requires improvement - 63 to 87% = good - 88% and above = outstanding So in CQC's example, 65.6% puts Safe in the good range, and CQC concludes: "Therefore, the rating for the safe key question in this case is good." The scale starts at 25% because that is the lowest possible result: every statement scoring 1. The percentage also tells CQC where within a rating a practice sits. CQC says the score "will indicate a more detailed position within the rating scale", so a good rating can be "in the upper threshold, nearing outstanding" or "in the lower threshold, nearer to requires improvement." Two practices with the same rating can be in quite different places, and the score shows it. ## Step 4: The two limiting rules The percentage is not the whole story. CQC applies two rules, in its words to "make sure any areas of poor quality are not hidden": - "If the key question score is within the good range, but one or more of the quality statement scores is 1, the rating is limited to requires improvement." - "If the key question score is within the outstanding range, but one or more of the quality statement scores is 1 or 2, the rating is limited to good." Two illustrative examples, with our arithmetic: - **Good range, one statement at 1.** Seven Safe statements score 3 and learning culture scores 1. The total is 22 of 32, which is 68.75% and inside the good range. Because one statement scored 1, Safe is limited to requires improvement. - **Outstanding range, one statement at 2.** Seven Safe statements score 4 and one scores 2. The total is 30 of 32, which is 93.75%. Because one statement scored 2, Safe is limited to good. > This is the part of the method that matters most day to day. A key question rating is not an average that strong areas can carry. One statement where the evidence shows significant shortfalls is enough to hold a good key question at requires improvement. ## Step 5: The five key questions become an overall rating The overall rating is built from the five key question ratings, not from the percentages. CQC's principles: - "The 5 key questions are all equally important and are weighted equally when aggregating." - "At least 2 of the 5 key questions would normally need to be rated as outstanding and 3 key questions rated as good before an aggregated rating of outstanding can be awarded." - "The overall rating will normally be good if there are no key question ratings of inadequate and no more than one key question rating of requires improvement." - "If 2 or more of the key questions are rated as requires improvement, then the overall rating will normally be requires improvement." - "If 2 or more of the key questions are rated as inadequate, then the overall rating will normally be inadequate." Note the word "normally" in each. These are principles CQC applies, not a formula. ## Your first assessment: what happens to the statements CQC does not look at This is the part that is easiest to miss. Under the current approach CQC does not have to assess all 34 statements at once. For a practice with an existing rating, CQC's page on [calculating the first scores](https://www.cqc.org.uk/guidance-regulation/providers/assessment/assessing-quality-and-performance/how-we-reach-rating/how-we-will-calculate-first-scores) says it "will select which quality statements to look at", based on "national priorities, set by the type of service, as well as a consideration of the information we hold about your service." The statements it assesses are scored on the new evidence. The rest are scored from your current published rating for that key question: - 4 for each statement where the key question is rated outstanding - 3 for each statement where it is rated good - 2 for each statement where it is rated requires improvement - 1 for each statement where it is rated inadequate There are two exceptions that apply to GP practices. Workforce wellbeing and enablement, now under Caring, takes its starting score from your Well-led rating because the topic moved key question. Environmental sustainability gets no starting score because it is new to the framework. An illustrative example, with our arithmetic. A practice is currently rated good for Safe, so the five Safe statements CQC does not assess each carry a 3, which is 15. CQC assesses the other three and scores them 2, 3 and 2. The total is 22 of 32, 68.75%, and Safe stays good. If one of those three had scored 1 instead (2, 3 and 1), the total would be 21 of 32, 65.6%. That is still in the good range, but the limiting rule applies and Safe is limited to requires improvement. > So the statements CQC chooses to look at carry a lot of weight. Practices that have never been inspected are treated differently: there are no starting scores, and CQC says it "will normally collect evidence for all the quality statements within the first year." ## What is changing CQC has said it will move away from this method. In its [initial response](https://www.cqc.org.uk/about-us/how-we-involve-you/consultations/initial-response-public-consultation-better-regulation-better-care) to the Better regulation, better care consultation: "We will remove scoring from our assessment approach. Future rating judgements will be made holistically using the professional judgement of our inspection teams, informed by evidence and with reference to the rating characteristics and supporting guidance." It has also said it "will no longer make formal judgements below the level of our 5 key questions". The new sector frameworks are still being tested. CQC's [June 2026 update](https://www.cqc.org.uk/about-us/improving-how-we-work/0626-update) says the pilots run "between June and October 2026, with final evaluation in November 2026", and that pilot assessments run alongside existing inspections, not instead of them. CQC wants "to get our new assessment approach right before it goes live across all services." Its [August update](https://www.cqc.org.uk/about-us/improving-how-we-work/0826-update) says testing and piloting continued "throughout the summer". So today, the method above is the one in use. When it changes, the arithmetic will change, but the evidence an inspector reads and the judgement they make from it will still be what a rating rests on. ## What this means for your practice - **Know where each statement stands, not just each key question.** A good overall rating can sit on a statement that would score 1 if it were assessed tomorrow. - **Know which statements were assessed last time.** If CQC looked at only a few, the rest are carrying your previous rating, and a focused assessment can move them. - **Look for the weakest statement first.** Because of the limiting rules, one significant shortfall can matter more to a key question rating than several small improvements elsewhere. - **Think in evidence, statement by statement.** Each score is a judgement on what the evidence for that statement shows. ## How CQC AI Auditor uses this method CQC AI Auditor scores each quality statement 1 to 4 using CQC's descriptions, rolls the scores up to a percentage and a rating for each key question with CQC's thresholds, and applies both limiting rules, so you can see which statement is holding a key question back. Our [How we score](https://www.cqcaudit.ai/how-we-score) page sets out the method, and [How it works](https://www.cqcaudit.ai/how-it-works) explains the two levels: the self check scores the answers you give, and the AI assisted audit scores the evidence you hold. It is our assessment of your evidence, not a prediction of your CQC rating, and when CQC changes its method we will change with it. You can [run your first self check free](https://www.cqcaudit.ai/). ## Frequently asked questions **How does CQC score a GP practice?** Each quality statement CQC assesses is scored from 1 (evidence shows significant shortfalls) to 4 (evidence shows an exceptional standard). The scores for each key question are added up, divided by the maximum possible and turned into a percentage, which CQC converts into a rating using fixed thresholds. **What percentage do you need for a good rating from CQC?** At key question level, 63 to 87% is good and 88% and above is outstanding, unless a limiting rule applies. Any statement scoring 1 limits a key question in the good range to requires improvement, and any statement scoring 1 or 2 limits one in the outstanding range to good. **Does CQC assess all 34 quality statements at every assessment?** Not necessarily. For a practice with an existing rating, CQC selects which statements to look at. The ones it does not assess are scored from your current rating for that key question: 3 each if it is rated good, for example. **Is CQC getting rid of scoring?** CQC has said it will remove scoring and make rating judgements holistically at key question level. The new frameworks are being piloted, with evaluation in November 2026, and are not live yet. The current scoring method is the one in use today. ## Ask us We do not run comments on Insights. If you have a question about CQC scoring, ask on [our LinkedIn page](https://www.linkedin.com/showcase/cqc-ai-auditor/) or reply to the newsletter. ## Sources - CQC, [How we reach a rating](https://www.cqc.org.uk/guidance-regulation/providers/assessment/assessing-quality-and-performance/reach-rating) (page last updated 13 May 2025). Read 6 October 2026. - CQC, [Example: how we reach a rating](https://www.cqc.org.uk/guidance-regulation/providers/assessment/assessing-quality-and-performance/reach-rating/example-gp-practice) (11 December 2024). Read 6 October 2026. - CQC, [Calculating the first scores using our new approach](https://www.cqc.org.uk/guidance-regulation/providers/assessment/assessing-quality-and-performance/how-we-reach-rating/how-we-will-calculate-first-scores) (11 December 2024). Read 6 October 2026. - CQC, [Update on our current scoring approach](https://www.cqc.org.uk/about-us/improving-how-we-work/update-our-current-scoring-approach) (19 May 2025). Read 6 October 2026. - CQC, [Initial response to the Better regulation, better care consultation](https://www.cqc.org.uk/about-us/how-we-involve-you/consultations/initial-response-public-consultation-better-regulation-better-care) (page last updated 25 March 2026). Read 6 October 2026. - CQC, [Piloting, testing and evaluation of new assessment method](https://www.cqc.org.uk/about-us/improving-how-we-work/0626-update) (4 June 2026) and [Rebuilding our regulatory approach and digital services](https://www.cqc.org.uk/about-us/improving-how-we-work/0826-update) (27 August 2026). Read 6 October 2026.