Key takeaways
- Decide what you want to change before you write a question; feedback with no decision attached rarely leads anywhere.
- Even the national GP Patient Survey gets a 22.1% response rate, and most replies now come online[1].
- Online-only surveys miss the patients who struggle most to get through; council and social renters report more difficulty contacting their practice[2].
- Keep surveys to three to five questions, sent within 48 hours of the experience you are asking about.
- If you publish reviews, you must take steps to stop fake or hidden-incentive reviews appearing[3].
Why most practice feedback goes nowhere
Most practices collect plenty of feedback: Friends and Family Test responses, comment cards, online reviews, complaints, patient participation group minutes and the annual GP Patient Survey. Far fewer can point to a change that came from it. The usual reasons are that the questions were too general, the answers arrived too late to act on, nobody owned the results, or the people who replied were not the people with the problem.
National results show how mixed the picture can be. In 2026, 76.7% of patients rated their own GP practice good overall[1], while only 35% of the public were satisfied with GP services in 2025[4]. A single satisfaction score tells you very little about what to do on Monday morning. Useful feedback is narrower: it asks about one part of the patient journey, at the right moment, from the right people.
Start with the decision, not the question
Before writing a survey, write down the decision it should inform. “Should we move the phone line opening from 8am to 7.30am?” leads to precise questions about when people call and why. “How are we doing?” leads to a vague score.
Name the decision
One sentence: what might change as a result of this feedback, and who will decide.
Pick the moment
Choose the point in the journey you want to understand: booking, the phone call, the wait, the consultation, the follow-up.
Choose who to ask
Everyone who had that experience in a set period, or a sample that includes groups you know struggle.
Write three to five questions
One rating, one or two specific questions, one free-text question asking what would have made it better.
Set a date to act
Agree when the results will be reviewed and by whom, before you send anything.
Timing and sampling
Memories fade and blur quickly. A text sent within a day or two of an appointment produces answers about that appointment; a survey sent a month later produces answers about the practice in general. For phone access, ask callers shortly after the call, not at the next visit.
| Method | Best timing | Strengths | Watch out for |
|---|---|---|---|
| Text link after an appointment | Within 24 to 48 hours | Quick, cheap, high volume | Misses patients without mobiles or who avoid links |
| Paper card at reception | On the day | Reaches patients who do not use digital routes | Low numbers, often only very happy or very unhappy |
| Phone follow-up of a sample | Within a week | Rich answers, reaches less-heard groups | Staff time; needs a short script |
| Short survey after a phone call | Same day | Measures the phone experience directly | Needs a telephony or text system that can trigger it |
| Patient participation group session | Quarterly | Discussion and ideas, not just ratings | Members may not represent the whole list |
| Online reviews | Ongoing | Public and visible to prospective patients | Self-selected; must not be manipulated |
Combine methods so that one group's views do not drown out another's. A practice that only uses text links will hear mostly from younger, digitally confident patients, and those are not always the people having the hardest time.
Who replies, and who does not
The national GP Patient Survey received 654,714 responses in 2026, a 22.1% response rate, and 566,443 of those replies were completed online[1]. If a large, well-funded national survey hears from about one patient in five, a practice survey will usually hear from fewer, and the people who reply will be different from those who do not.
- Online 566,443 87%
- Other modes 88,271 13%
Source: Ipsos (GP Patient Survey provider for NHS England), 2026[1]
Show the data
| Item | Value |
|---|---|
| Online | 566,443 |
| Other modes | 88,271 |
The patients you most need to hear from are often the least likely to reply. ONS analysis found that adults renting from a council were 46% more likely, and those renting from another social landlord 51% more likely, to report difficulty contacting their GP practice than outright homeowners[2]. A text survey after an online booking will not reach someone who gave up on the phone queue.
- Add a paper option at reception and in the pharmacy.
- Phone a small sample of patients who have not booked online recently.
- Ask carers and patients with recorded communication needs in the format they use.
- Translate the survey into the main languages spoken on your list.
- Record which method each answer came from, so you can compare.
Writing questions people can answer
Good questions ask about one thing, in plain words, with answer options that match how people think. Avoid double questions such as “Was the receptionist friendly and helpful?” and leading ones such as “How much did you like our new phone system?”
Example (fictional) three-question survey after a phone call: 1. How easy was it to get through to us today? (Very easy to very difficult.) 2. Did you know what would happen next by the end of the call? (Yes, no, not sure.) 3. What one thing would have made today's call better? (Free text.)
The second question is the useful one. It measures something the practice controls and can improve within a week, and it matches the national expectation that patients should know on the day how their request will be handled[5].
Making sense of free text quickly
Free-text answers hold most of the value and most of the work. A simple coding routine turns a pile of comments into a short list of themes in an hour.
- Read the first 30 comments and write down the themes you see, in a few words each.
- Go through all comments and tag each with one or two themes.
- Count the tags and pick the top three themes.
- For each theme, choose two or three comments that describe it well, with identifying details removed.
- Note any comment that raises a safety or safeguarding concern and pass it to a clinician the same day.
Do the coding with two people if you can, such as the practice manager and a receptionist. They will notice different things, and the receptionist often knows exactly which part of the process a comment is about.
Using what you already have
Before running a new survey, mine the feedback already sitting in the practice. Complaints and compliments, informal concerns logged at reception, call recordings, online reviews and patient participation group notes all describe real experiences, and none of them needs a new questionnaire. A quarterly review that pulls these sources together often answers the question you were about to survey for.
Staff are a source too. Receptionists hear the same frustrations dozens of times a week and can usually name the top three in a minute. Ask them at a team meeting, write the answers down and compare them with what patients say. Where the two match, you have a strong signal; where they differ, you have a useful question to ask patients directly.
Keep a simple feedback log with the date, source, theme and any action taken. Over a year it becomes the evidence base for your CQC conversations and your annual report, and it stops good ideas from being lost when a manager changes.
Online reviews: ask fairly
Online reviews are feedback too, and prospective patients read them. You can invite reviews, but you cannot manipulate them. Under the Digital Markets, Competition and Consumers Act 2024, businesses that publish consumer reviews must take steps to prevent fake reviews and reviews with hidden incentives from appearing[3]. In practice: ask every patient the same way rather than only the happy ones, never offer anything in return without saying so, and never write or commission reviews.
Respond to reviews briefly and without confirming that the reviewer is a patient or discussing their care. Offer a route to talk privately.
Example: from feedback to a change in a month
Example (fictional): Brook Lane Surgery wants to know why patients still say the phones are hard after a new system went in. It sends a three-question text to callers for two weeks and phones 20 patients who rarely book online. The top theme is that callers do not know whether they are in the queue for an appointment or for a call-back. The practice changes its queue message and its reception script, then repeats the survey a month later. The share saying they knew what would happen next rises, and the practice publishes the change on its waiting room board.
The next step is to tell patients what you did. See “you said, we did”: closing the feedback loop, and for measures to track over time, see patient satisfaction metrics. Feedback from carers and patients with dementia needs a gentler approach; the guides to supporting carers and dementia-friendly practice cover it.
Frequently asked questions
How many questions should a practice survey have?
Three to five is usually enough: one rating, one or two specific questions and one free-text question. Longer surveys get fewer and less thoughtful replies.
What response rate should we expect?
Lower than you would like. The national GP Patient Survey achieved 22.1% in 2026[1]. Focus on hearing from the right mix of patients rather than the highest number.
Can we offer a prize draw for completing a survey?
For an internal survey, a small prize draw is common, but do not link incentives to public reviews unless they are clearly disclosed. Hidden incentives for reviews are banned[3].
How do we reach patients who do not use digital services?
Use paper cards at reception and pharmacy, phone a small sample, and ask through carers. Patients in social housing report more difficulty contacting practices[2], so make sure they are in the sample.
References
Ipsos (GP Patient Survey provider for NHS England) (2026). 2026 GP Patient Survey results released (opens in a new tab) ↑
Office for National Statistics (2026). Experiences of NHS healthcare services in England: September 2026 (opens in a new tab) ↑
Competition and Markets Authority (2025). Fake reviews: guidance on the banned practice (CMA208) (opens in a new tab) ↑
Nuffield Trust and The King's Fund (2025). Public satisfaction with the NHS and social care in 2025: results from the British Social Attitudes survey (opens in a new tab) ↑
NHS England and DHSC (2023). Delivery plan for recovering access to primary care (opens in a new tab) ↑